Content uploaded by Kim Mcgrail
Author content
All content in this area was uploaded by Kim Mcgrail on Jun 30, 2015
Content may be subject to copyright.
Statistique
Canada
Statistics
Canada
Catalogue no. 82-003-X
Health
Volume 21, Number 2
Reports
How to obtain more information
Specifi c inquiries about this product and related statistics or services should be directed to: Health Information and Research
Division, Statistics Canada, Ottawa, Ontario, K1A 0T6 (telephone: 613-951-1765).
For information about this product or the wide range of services and data available from Statistics Canada, visit our website at
www.statcan.gc.ca, e-mail us at infostats@statcan.gc.ca, or telephone us, Monday to Friday from 8:30 a.m. to 4:30 p.m., at the
following numbers:
Statistics Canada’s National Contact Centre
Toll-free telephone (Canada and United States):
Inquiries line 1-800-263-1136
National telecommunications device for the hearing impaired 1-800-363-7629
Fax line 1-877-287-4369
Local or international calls:
Inquiries line 1-613-951-8116
Fax line 1-613-951-0581
Depository Services Program
Inquiries line 1-800-635-7943
Fax line 1-800-565-7757
To access and order this product
This product, Catalogue no. 82-003-X, is available free in electronic format. To obtain a single issue, visit our website at
www.statcan.gc.ca and select “Publications.”
This product, Catalogue no. 82-003-X, is also available as a standard printed publication at a price of CAN$24.00 per issue and
CAN$68.00 for a one-year subscription.
The following additional shipping charges apply for delivery outside Canada:
Single issue Annual subscription
United States CAN$6.00 CAN$24.00
Other countries CAN$10.00 CAN$40.00
All prices exclude sales taxes.
The printed version of this publication can be ordered as follows:
• Telephone (Canada and United States) 1-800-267-6677
• Fax (Canada and United States) 1-877-287-4369
• E-mail infostats@statcan.gc.ca
• Mail Statistics Canada
Finance
R.H. Coats Bldg., 6th Floor
150 Tunney's Pasture Driveway
Ottawa, Ontario K1A 0T6
• In person from authorized agents and bookstores.
When notifying us of a change in your address, please provide both old and new addresses.
Standards of service to the public
Statistics Canada is committed to serving its clients in a prompt, reliable and courteous manner. To this end, Statistics Canada has
developed standards of service that its employees observe. To obtain a copy of these service standards, please contact Statistics
Canada toll-free at 1-800-263-1136. The service standards are also published on www.statcan.gc.ca under “About us” > “Providing
services to Canadians.”
Published by authority of the Minister responsible for Statistics Canada
© Minister of Industry, 2010
All rights reserved. The content of this electronic publication may be reproduced,
in whole or in part, and by any means, without further permission from Statistics
Canada, subject to the following conditions: that it be done solely for the purposes
of private study, research, criticism, review or newspaper summary, and/or for non-
commercial purposes; and that Statistics Canada be fully acknowledged as follows:
Source (or “Adapted from”, if appropriate): Statistics Canada, year of publication,
name of product, catalogue number, volume and issue numbers, reference period
and page(s). Otherwise, no part of this publication may be reproduced, stored in a
retrieval system or transmitted in any form, by any means-electronic, mechanical
or photocopy-or for any purposes without prior written permission of Licensing
Services, Client Services Division, Statistics Canada, Ottawa, Ontario, Canada
K1A 0T6.
June 2010
Catalogue no. 82-003-XPE, Vol. 21, No. 2
ISSN 0840-6529
Catalogue no. 82-003-XIE, Vol. 21, No. 2
ISSN 1209-1367
Frequency: Quarterly
Ottawa
Note of Appreciation
Canada owes the success of its statistical system to a long-standing partnership
between Statistics Canada, the citizens of Canada, its businesses, governments
and other institutions. Accurate and timely statistical information could not be
produced without their continued cooperation and goodwill.
Health
Health
Reports
Reports
Catalogue no. 82-003-XPE • Volume 21 Number 2
A Canadian peer-reviewed journal of
population health and health services research
2
Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Symbols
The following standard symbols are used in Statistics Canada publications:
. not available for any reference period
.. not available for specifi c reference period
... not applicable
p preliminary
r revised
x suppressed to meet the confi dentiality requirements of the Statistics Act
E use with caution
F too unreliable to be published
The paper used in this publication meets the minimum requirements of American
National Standard for Information Sciences – Permanence of Paper for Printed
Library Materials, ANSI Z39.48 – 1984.
Editor-in-Chief
Christine Wright
Senior Editor
Mary Sue Devereaux
Assistant Editor
Anne Marie Baxter
Production Manager
Robert Pellarin
Creative Services
Rasha Bradic
Administration
Amber Doy-Yat
Associate Editors
David Buckeridge
McGill University
Elizabeth Lin
The Clarke Institute of Psychiatry
Doug Manuel
Ottawa Health Research Institute
and Statistics Canada
Nazeem Muhajarine
University of Saskatchewan
Georgia Roberts
Statistics Canada
Nancy Ross
McGill University and Statistics Canada
Geoff Rowe
Statistics Canada
Michelle Simard
Statistics Canada
Electronic version: Health Reports is available free in PDF
or HTML format. The current issue may be obtained at
www.statcan.gc.ca/healthreports. For previous issues, select
“Other issues in the series” from the left sidebar of the Health
Reports website.
Author information: We seek submissions from researchers
based in government or academia. Submissions can come in
the form of a traditional research article, a shorter descriptive
piece that we call “Health Matters,” or a contribution that
addresses technical issues related to the analysis of complex
health surveys or administrative databases—“Methodological
Insights.” For detailed author guidelines, please visit the jour-
nal’s website at: www.statcan.gc.ca/healthreports.
Aussi disponible en français : Rapports sur la santé, no 82-003-X au catalogue
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010 3
About
About
Health Reports
Health Reports
Health Reports publishes original research on diverse
topics related to the health of populations and the
delivery of health care. The journal archives, for the
research and policy communities and for the general public,
discoveries from analyses of national/provincial surveys and
administrative databases, as well as results of international
comparative health research. Health Reports is also a forum
for sharing methodological information by those using
health surveys or administrative databases. Health Reports
is produced by the Health Analysis Division at Statistics
Canada. Articles appear monthly in electronic format and
quarterly in print, and are indexed in Index Medicus and
MEDLINE.
For more information about Health Reports, contact
the Editor-in-Chief, Health Analysis Division, Statistics
Canada, 24th Floor, R.H. Coats Building, Ottawa,
Ontario, Canada K1A 0T6. Telephone: (613) 951-1765;
fax: (613) 951-3959; email: Christine.Wright@statcan.gc.ca
Editorial Board
David L. Streiner, Scientifi c Editor
University of Toronto
Bill Avison
University of Western Ontario
Adam Baxter-Jones
University of Saskatchewan
Lise Dubois
University of Ottawa
James Dunn
University of Toronto and Centre for
Research on Inner City Health
Bob Evans
University of British Columbia
David Feeny
Kaiser Permanente
Rick Glazier
Institute for Clinical Evaluative Sciences and
University of Toronto
Judy Guernsey
Dalhousie University
Glenn Irwin
Health Canada
Howard Morrison
Public Health Agency of Canada
Cameron Mustard
Institute for Work and Health, University of
Toronto
Tom Noseworthy
University of Calgary
Patricia O’Campo
University of Toronto and Centre for
Research on Inner City Health
Jennifer O’Loughlin
University of Montreal
Indra Pulcins
Canadian Institute for Health Information
Nancy Ross
McGill University and Statistics Canada
Paul Veugelers
University of Alberta
Michael Wolfson
Statistics Canada
4
Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Research articles
Waiting time for medical specialist consultations ..... 7
by Gisèle Carrière and Claudia Sanmartin
Factors beyond medical need are associated with how long patients
wait to see a specialist.
Blood pressure in Canadian children
and adolescents ........................................................... 15
by Gilles Paradis, Mark S. Tremblay, Ian Janssen, Arnaud
Chiolero and Tracey Bushnik
A small percentage of Canadians aged 6 to 19 years have borderline or
elevated blood pressure.
Using population data to measure outcomes of care:
The case of hip and knee replacements .................... 23
by Claudia Sanmartin, Kimberlyn McGrail, Mike Dunbar and
Eric Bohm
Outcomes of joint replacement differ by the initial diagnosis or reason
for the surgery.
In this issue
In this issue
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010 5
Health matters
Weight gain during pregnancy: Adherence to
Health Canada’s guidelines ....................................... 31
by Hélène Lowell and Doris C. Miller
Relatively high percentages of women who are young, primiparous,
less educated or Aboriginal gain more weight than recommended while
they are pregnant.
Methodological insights
The Manitoba Human Papillomavirus vaccine
surveillance and evaluation system .............................. 37
by Erich V. Kliewer, Alain A. Demers, Marc Brisson, Alberto
Severini, Robert Lotocki, Brenda Elias, Gregory Hammond,
George Wurtak and the Manitoba HPV Research Group
The availability of extensive linkable databases in Manitoba allows for
the development of a comprehensive Human Papillomavirus vaccine
surveillance and evaluation system.
Evaluating the Hyperactivity/Inattention Subscale of
the National Longitudinal Survey of Children
and Youth ....................................................................45
by Alice Charach, Elizabeth Lin and Teresa To
High scores on the Hypeactivity/Inattention Subscale of the National
Longitudinal Survey of Children and Youth are associated with current
methylphenidate use and diagnosed emotional disorder.
7
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Waiting time for medical specialist consultations in Canada, 2007 • Research article
Waiting time for medical specialist
consultations in Canada, 2007
by Gisèle Carrière and Claudia Sanmartin
Abstract
Background
Waiting for specialist consultations can represent a
substantial component of overall waiting time in the
continuum of care. However, relatively little is known
about the factors associated with how long patients
wait for an initial specialist consultation.
Data and methods
The analysis is based on a subsample of 5,515
respondents aged 15 or older to the 2007 Canadian
Community Health Survey who had consulted a
specialist about a new condition in the previous 12
months and reported a waiting time. Multivariate
logistic regression models were used to identify
patient- and provider-related factors associated with
waiting time.
Results
Female patients were less likely than male patients
to see a specialist within a month. The nature of
the new condition and the source of referral were
signifi cantly associated with waiting time. Compared
with those referred by a family physician, patients
referred by another specialist or a health care provider
other than a physician, or who did not require a
referral, were more likely to have a shorter waiting
time. For men, but not women, household income
and immigrant status were associated with waiting
time.
Interpretation
This analysis suggests that factors beyond medical
need are associated with how long patients wait to
see a specialist. More research could usefully explore
decision-making and communication processes
between primary care physicians and specialists to
better understand how urgency is assessed, how
patients are triaged for specialist consultations, and
how these patterns differ among various groups of
patients.
Keywords
access to care, specialists, immigrant, socio-
economic
Authors
Gisèle Carrière (1-604-666-5907; Gisele.Carriere@
statcan.gc.ca) is with the Health Analysis Division
at Statistics Canada; she is based in Vancouver,
British Columbia. Claudia Sanmartin (1-613-951-
6059; Claudia.Sanmartin@statcan.gc.ca) is with the
Health Analysis Division at Statistics Canada, Ottawa,
Ontario, K1A 0T6.
ccessibility is fundamental to the quality of
health care. In Canada, waiting time has been
identifi ed as a key measure of access and the major
barrier among those who experienced diffi culties
obtaining care.1,2 In 2005, approximately 20% of
Canadians reported adverse effects as a result of
waiting for health care, including worry and stress
and pain.2
A
Recently, numerous initiatives across
Canada have endeavored to reduce
waiting time for specialized health
services,1,3-5 particularly for non-
emergency procedures in fi ve priority
areas identifi ed in the 2004 Health
Accord.6 While waiting times for surgery
and other procedures can be a signifi cant
barrier to care, they represent only one
of the waiting periods experienced
across the continuum of care.7 Interest
is now shifting “upstream” toward
waits that occur earlier in the delivery
of health care, including waiting for
specialist consultations, which can
account for a signifi cant component of
overall waiting time. For example, in
2005, among Canadians who had had
a joint replacement, waits for an initial
orthopedic specialist consultation made
up nearly 30% of total waiting time.5
Despite growing interest in access to
specialists, little is known about patient-
and provider-related factors associated
with shorter versus longer waiting times
for initial consultations. Access to
specialists, like other types of health care
services, may be associated with a range
of factors.8 Patients’ socio-economic
characteristics have been related to the
use of specialist services,9-13 but it is
not known if these characteristics are
also associated with waiting time for
specialist consultations. Provider-related
variables,13,14 including physicians’
decision processes in assessing
urgency,15 have also been related to who
gets referred to specialists. But again, it
is unclear if these factors are associated
with how long patients wait.
Based on information from the 2007
Canadian Community Health Survey,
this study examines associations between
patient- and provider-related factors
and the length of time patients wait to
consult a specialist about a new illness or
condition.
8Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Waiting time for medical specialist consultations in Canada, 2007 • Research article
Methods
Data source
The data are from a subsample of
respondents aged 15 or older in the
10 provinces, to whom the “Access to
Care” and “Waiting Times” modules of
the 2007 Canadian Community Health
Survey were administered. These
modules, formerly the Health Services
Access Survey, were incorporated into
the Canadian Community Health Survey
in 2003.
The survey response rate was 75.7%.
Residents of institutions, the three
territories, Indian reserves, Crown
lands and certain remote regions and
full-time members of the Canadian
Forces were excluded from the survey.
Proxy responses were not permitted.
Since respondents in this analysis
are a subsample, the multiple sample
frames of the parent survey apply.
More information about the Canadian
Community Health Survey is available
in other reports16,17 and on Statistics
Canada’s website: http://www.statcan.
gc.ca/cgi-bin/imdb/p2SV.pl?Function=g
etSurvey&SDDS=3226&lang=en&db=i
mdb&adm=8&dis=2).
The study pertains to 5,515
respondents who reported that they had
consulted a specialist about a new illness
or health condition in the previous 12
months and who reported a waiting time.
Analytical techniques
Factors associated with waiting times for
specialist consultations were determined
with multivariate logistic regression
analyses in total cohort and sex-specifi c
models. The outcome of interest was
waiting time for the initial specialist
consultation, expressed as a dichotomous
variable, indicating whether patients
waited: 1) less than one month, or 2)
longer. This cut-off was chosen based
on the median waiting time (4.3 weeks).
To account for the complex survey
design, standard errors, coeffi cients of
variation and 95% confi dence intervals
were estimated using the bootstrap
technique.18,19 Differences between
estimates were tested for statistical
signifi cance, established at the level of
p<0.05.
The patient-related factors
hypothesized to be associated with
waiting time for a specialist consultation
were sex, age, education, household
income, immigrant status and rural/urban
residence.
Immigrants were defi ned as
respondents who were born outside of
Canada and were not Canadian citizens
by birth. They were categorized
according to their duration of residence
in Canada: less than 10 years, or 10 or
more years before the survey date.
Based on a national distribution of
total household income (adjusted for
household size), respondents were
classifi ed into household income
quintiles.
Education is the highest level of
personal educational attainment.
Waiting time for an initial specialist
consultation has been shown to be
related to the nature of the underlying
health condition.5 Therefore, adjustment
was made for the type of new condition
reported and the presence of chronic
conditions. The chronic conditions
were asthma, arthritis, cancer, diabetes,
chronic obstructive pulmonary disease,
heart disease, and mood disorders
(depression, bipolar disorder, mania and
dysthymia).
People with chronic conditions,
particularly those with multiple
comorbidities, often experience poorer
health. This may affect the severity of
the condition for which they seek care,
and in turn, waiting time. To partially
adjust for this possibility, respondents
were classifi ed according to the number
of selected chronic conditions they
reported: none, one, or two or more.
In addition, respondents were identifi ed
as having (or not having) high blood
pressure.
Provider-related factors were
represented by two variables: having
a regular doctor and the source of the
specialist referral (family doctor, another
specialist, another health care provider,
or did not require a referral).
The multivariate models initially
included province of residence, but
the results did not differ from those
not adjusted for province. Therefore,
because of the limited sample size,
province was removed from the fi nal
models to preserve statistical power.
All independent variables in the
models were tested for multicollinearity.
Results
Characteristics of patients
consulting specialists
In 2007, an estimated 3 million patients
aged 15 or older reported having consulted
a specialist about a new condition in the
previous year (Table 1). Almost 60% of
these patients were female. More than
half of the patients were aged 45 or older.
Men consulting specialists were slightly
older than women, averaging 50 years
versus 47 years (data not shown). The
educational attainment and household
income of patients tended to be slightly
higher than those of the population
overall (data not shown). Approximately
20% were immigrants, just under three-
quarters of whom had been in Canada
more than a decade.
The top three conditions about
which specialists were consulted were
gynecological conditions (12%), heart/
stroke (9%), and cancer (7%), though
of course, this varied by sex. Fully 21%
of the women had consulted a specialist
about a new gynecological condition
(data not shown). Men were more likely
than women to have consulted a specialist
because of a new heart condition/stroke
(13% versus 7%).
Slightly fewer than half the patients
also had at least one chronic condition,
and 17% reported multiple comorbidities.
Most of the patients (91%) who had
seen a specialist had a regular doctor.
Over two-thirds (68%) of the patients
had been referred to the specialist by their
family doctor, 11% by another specialist,
12% by another health care provider, and
9% reported that they had not needed a
referral. The most common sources of
referral varied by province, especially in
Quebec, where almost 20% of patients
9
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Waiting time for medical specialist consultations in Canada, 2007 • Research article
reported that they had been referred by
a health care provider who was not their
family doctor or another specialist, and
17% had not required a referral (Table 2).
Distribution of waiting times
Nearly half (46%) of the patients had
waited less than a month for their
initial specialist consultation (Table 2).
An additional 40% waited one to three
months, and 14% waited more than three
months. The percentage who saw the
specialist within a month varied from
37% in Newfoundland and Labrador
and Manitoba to 51% in Quebec. Just
under half (49%) of those who required
a consultation for a new mental health
condition waited less than a month.
The length of the wait depended on
the nature of the new condition. Not
surprisingly, patients with potentially
life-threatening illnesses were the most
likely to have seen a specialist within a
month. Almost 60% of those with a heart
condition/stroke or cancer waited less
than a month for their initial consultation,
compared with 29% of those with
arthritis/rheumatism (Table 3). Just
under half (49%) of those who required
a consultation for a new mental health
condition waited less than a month.
Overall, 51% of male patients with a
new condition waited less than a month
for their initial consultation. However,
63% of men with a new heart condition/
stroke saw a specialist within a month, as
did 56% of those with cancer, 55% with
eye conditions, and 52% with mental
disorders.
Compared with men, a lower
percentage (42%) of female patients had
their fi rst specialist consultation within a
month. Again, the likelihood of a short
wait varied with the condition. More
than half of those with cancer (57%) or
a heart condition/stroke (55%) had their
fi rst consultation within a month. On the
other hand, relatively small percentages
with gynecological conditions (39%),
skin conditions (39%) or arthritis/
rheumatism (25%) waited less than a
month.
Table 1
Characteristics of patients who consulted specialist about new condition,
household population aged 15 or older, Canada excluding territories, 2007
Sample
count
Weighted
estimate
(’000)
Column
(%)
Total aged 15 or older 5,515 3,043 100.0
Sex
Male 2,035 1,226 40.3
Female 3,480 1,816 59.7
Age group
15 to 34 1,138 746 24.5
35 to 44 877 588 19.3
45 to 64 2,099 1,148 37.7
65 or older 1,401 560 18.4
Education
Less than secondary graduation 1,015 469 15.5
Secondary graduation 801 418 13.8
Some postsecondary 423 248 8.2
Postsecondary graduation 3,250 1,894 62.5
Household income quintile
1 (lowest) 1,004 532 17.5
2 962 478 15.7
3 983 605 19.9
4 1,000 537 17.7
5 (highest) 984 580 19.1
Missing 582 311 10.2
Immigrant status
Immigrant (0 to 10 years in Canada) 148 173 5.7
Immigrant (more than 10 years in Canada) 659 446 14.7
Canadian-born 4,688 2,412 79.6
Residence
Urban core 3,290 2,179 71.6
Urban fringe 149 79 2.6
Urban area outside Census Metropolitan Area/
Census Agglomeration
427 130 4.3
Secondary urban core 60 46E1.5E
Mix of urban and rural 815 254 8.3
Rural 774 355 11.7
New condition
Gynecological condition 598 372 12.2
Heart condition/Stroke 530 279 9.2
Cancer 373 212 7.0
Skin condition 323 180 5.9
Cataract or other eye condition 320 164 5.4
Arthritis/Rheumatism 196 93 3.1
Mental health disorder 183 103 3.4
Asthma or other breathing condition 126 66 2.2
Other 2,852 1,567 51.5
High blood pressure†
No 4,046 2,381 78.4
Yes 1,459 657 21.7
Number of selected chronic conditions‡
None 2,603 1,602 53.2
One 1,775 903 30.0
2 or more 1,076 504 16.8
Has regular family doctor
Yes 5,132 2,766 90.9
No 382 275 9.1
Source of specialist referral
Family doctor 4,012 2,061 67.9
Other specialist 570 339 11.1
Other health care provider 571 372 12.2
Did not require referral 352 263 8.7
† excluded from count of chronic conditions
‡ selected chronic conditions were asthma, arthritis, cancer, chronic obstructive pulmonary disease, diabetes, heart disease, mood
disorders (depression, bipolar disorder, mania, dysthymia)
E use with caution (coeffi cient of variation 16.6% to 33.3%)
Notes: Estimates are based on population who completed initial consultation with medical specialist in previous 12 months and
provided information about waiting time. Except for total household income, analyses exclude non-response (“don’t know,”
“not stated,” “refusal”).
Source: 2007 Canadian Community Health Survey.
10 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Waiting time for medical specialist consultations in Canada, 2007 • Research article
Unadjusted bivariate results
Unadjusted analyses for the entire
subsample suggest that, in addition
to gender and the nature of the new
condition, several other factors were
associated with waiting time for specialist
consultations (Table 4).
People with high blood pressure were
less likely than those not affl icted to be
seen within a month. Comparatively
high percentages of patients without a
regular medical doctor and those who
were referred by someone other than
a family doctor or who did not need a
referral waited less than a month.
Unadjusted sex-specifi c analyses
show that among female patients, waiting
less than a month was signifi cantly
associated with only two variables: the
nature of the new condition and referral
source. However, among male patients,
in addition to the nature of the condition
and referral source, not having high
blood pressure or a regular family doctor
and being an immigrant were associated
with waiting less than a month.
Multivariate logistic regression
results
Results from the full model indicate that
even when the infl uence of the other
variables was controlled, female patients
were signifi cantly less likely than male
patients to see a specialist in less than
a month. And as expected, for both
sexes, the nature of the new condition
was signifi cantly associated with waiting
time: compared with those who had a
new heart condition/stroke, the odds of
seeing a specialist within a month were
Table 3
Unadjusted percentage distribution of waiting times to consult specialist about
new condition, by nature of new condition, household population aged 15 or
older, Canada excluding territories, 2007
New condition
Waiting time Less than 1 month
Less
than 1
month
1 to 3
months
3 month
or longer Males Females
Total 45.6 40.5 13.9 51.0 42.1
Heart condition/Stroke†59.3 33.3 7.4E63.3 54.5
Cancer 56.8 30.5 12.7E56.3 57.2
Mental health disorder 48.7 36.6 14.7E51.5 47.2E
Cataract or other eye condition 46.4* 39.6 14.1E* 54.6 41.8
Asthma or other breathing condition 45.0* 45.6 9.5Ex 43.7E
Other 44.4* 41.7* 13.9* 48.7* 40.6*
Skin condition 42.2* 42.6 15.2E* 45.4* 39.0*
Gynecological condition 39.1* 46.3* 14.6* ... 39.1*
Arthritis/Rheumatism 28.9* 40.1 31.0* x 24.6E*
† reference group
* signifi cantly different from estimate for reference group (p < 0.05)
E use with caution (coeffi cient of variation 16.6% to 33.3%)
x suppressed to meet confi dentiality requirements of Statistics Act
... not applicable
Notes: Estimates are based on population who completed initial consultation with medical specialist in previous 12 months and
provided information about waiting time. Analyses exclude non-response (“don’t know,” “not stated,” and “refusal”).
Source: 2007 Canadian Community Health Survey.
Table 2
Unadjusted percentage distribution of waiting times and of referral sources to consult specialist about new condition, by
province, household population aged 15 or older, Canada excluding territories, 2007
Province
Sample
count
Weighted
estimate
Waiting time
Waited
longer
than
median
Referral source
Less
than 1
month
1 to 3
months
3 month
or longer
Family
doctor
Other
specialist
Other
health
care
provider
Did not
require
referral
Number ’000 %
Total 5,515 3,043 45.6 40.5 13.9 39.1 67.9 11.2 12.2 8.7
Newfoundland and Labrador 217 49 37.0* 41.7 21.4* 49.3* 69.0 12.1E11.0EF
Prince Edward Island 173 13 44.6 44.3 11.1E41.7 77.8* 8.8E9.1EF
Nova Scotia 322 99 47.6 37.6 14.8E39.7 78.7* 10.2E4.5E* 6.6E
New Brunswick 284 76 44.3 37.9 17.8E45.2 73.6 8.8E9.5E8.1E
Quebec 520 711 51.0* 36.9 12.2 33.3* 46.2* 17.5* 19.8* 16.5*
Ontario 2,391 1,195 44.7 40.7 14.5 40.2 72.4* 10.2 11.0 6.4*
Manitoba 365 108 37.2* 48.8* 14.0E47.2* 76.7* 6.8E* 9.2E7.4E
Saskatchewan 363 81 46.3 38.8 15.0E43.3 78.1* 10.8E7.9E*F
Alberta 403 314 41.8 45.9 12.4E40.8 75.4* 8.1E9.1E7.4E
British Columbia 477 396 44.4 41.1 14.6 39.5 78.5* 7.1E* 9.5E4.9E*
* signifi cantly different from estimate for “rest of Canada” which represents all respondents not in province indicated (p < 0.05)
E use with caution (coeffi cient of variation 16.6% to 33.3%)
F too unreliable to be published (coeffi cient of variation greater than 33.3%)
Notes: Estimates are based on population who completed initial consultation with medical specialist in previous 12 months and provided information about waiting time. Analyses exclude non-response
(“don’t know,” “not stated,” and “refusal”).
Source: 2007 Canadian Community Health Survey.
11
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Waiting time for medical specialist consultations in Canada, 2007 • Research article
Table 4
Unadjusted prevalence and adjusted odds ratios for waiting time less than 1 month to consult specialist, by selected
characteristics and sex, household population aged 15 or older, Canada excluding territories, 2007
Total Males Females
Un-
adjusted
preva-
lence
95%
confidence
interval Adjusted
odds
ratio
95%
confidence
interval
Un-
adjusted
preva-
lence
95%
confidence
interval Adjusted
odds
ratio
95%
confidence
interval
Un-
adjusted
preva-
lence
95%
confidence
interval Adjusted
odds
ratio
95%
confidence
interval
from to from to from to from to from to from to
Total aged 15 or older 45.6 43.5 47.7 ... ... ... 50.9 47.3 54.6 ... ... ... 42.0 39.2 44.7 ... ... ...
Sex
Male†51.0 47.3 54.6 1.0 ... ... ... ... ... ... ... ... ... ... ... ... ... ...
Female 42.0*39.2 44.7 0.8*0.6 1.0 ... ... ... ... ... ... ... ... ... ... ... ...
Age group
15 to 34 47.3 42.7 51.9 ... ... ... 55.8 46.9 64.7 ... ... ... 43.1 38.0 48.2 ... ... ...
35 to 44 45.5 40.0 50.9 ... ... ... 49.9 40.8 59.0 ... ... ... 42.4 36.2 48.6 ... ... ...
45 to 64 43.4 40.1 46.7 ... ... ... 48.7 43.2 54.1 ... ... ... 40.0 35.4 44.1 ... ... ...
65 or older†47.9 43.6 52.2 ... ... ... 51.4 44.9 58.0 ... ... ... 44.6 39.1 50.1 ... ... ...
Age modelled as continuous term ... ... ... 1.0 1.0 1.0 ... ... ... 1.0 0.9 1.0 ... ... ... 1.0 1.0 1.0
Age modelled as quadratic term ... ... ... 1.0 1.0 1.0 ... ... ... 1.0 1.0 1.0 ... ... ... 1.0 1.0 1.0
Education
Less than secondary graduation 48.4 43.5 53.3 1.2 0.9 1.6 52.5 44.8 60.2 1.4 0.9 2.2 45.7 39.2 52.1 1.2 0.8 1.6
Secondary graduation 45.6 40.0 51.3 1.2 0.9 1.5 50.0 40.1 60.0 1.1 0.7 1.7 43.5 36.8 50.1 1.2 0.9 1.6
Some postsecondary 45.3 37.5 53.1 1.0 0.7 1.3 53.9 41.9 65.9 1.1 0.6 1.9 40.0 30.0 50.0 0.9 0.6 1.4
Postsecondary graduation†44.8 42.0 47.5 1.0 ... ... 50.2 45.4 55.0 1.0 ... ... 40.8 37.3 44.3 1.0 ... ...
Household income quintile
1 (lowest) 48.7 43.7 53.7 0.9 0.7 1.3 51.5 42.4 60.7 0.5*0.3 0.9 47.5 41.2 53.7 1.4 0.9 2.1
2 47.6 42.8 52.5 0.9 0.7 1.3 54.5 45.9 63.0 0.7 0.4 1.2 43.9 37.6 50.3 1.2 0.8 1.8
3 46.7 41.7 51.7 0.9 0.7 1.3 51.1 43.0 59.3 0.7 0.4 1.1 43.2 36.7 49.7 1.3 0.9 1.9
4 41.3 36.8 45.8 0.8 0.6 1.1 44.9 38.2 51.5 0.6*0.4 0.9 38.6 32.6 44.6 1.1 0.7 1.6
5 (highest)†47.0 41.7 52.4 1.0 ... ... 54.8 46.8 62.8 1.0 ... ... 38.4 31.4 45.5 1.0 ... ...
Missing 39.7 33.2 46.3 0.8 0.5 1.1 45.8 35.2 56.3 0.6 0.3 1.1 36.8 29.1 44.6 0.9 0.6 1.5
Immigrant status
Immigrant (0 to 10 years in Canada) 53.5 42.9 64.1 1.4 0.9 2.4 66.3*51.1 81.4 2.1 0.9 4.6 44.0E29.6 58.5 1.2 0.6 2.4
Immigrant (more than 10 years in Canada) 49.6 43.7 55.5 1.4*1.0 1.8 60.5*51.7 69.3 2.0*1.3 3.0 42.1 34.8 49.3 1.1 0.8 1.5
Canadian-born†44.4 42.0 46.7 1.0 ... ... 48.1 44.1 52.0 1.0 ... ... 41.9 38.8 44.9 1.0 ... ...
Residence
Urban core†45.7 43.1 48.4 1.0 ... ... 52.1 47.9 56.4 1.0 ... ... 41.4 37.9 44.9 1.0 ... ...
Urban fringe 34.7 24.2 45.3 0.6 0.4 1.0 x x x 0.6 0.3 1.5 33.4E20.3 46.5 0.6 0.3 1.1
Urban area outside Census Metropolitan Area/
Census Agglomeration
49.3 40.2 58.3 1.3 0.9 1.9 51.9 39.4 64.4 1.4 0.7 2.5 47.7 36.0 59.4 1.4 0.9 2.2
Secondary urban core 52.7E28.0 77.5 1.5 0.5 4.3 x x x 1.4 0.2 9.1 x x x 1.6 0.6 3.9
Mix of urban and rural 44.0 38.6 49.3 1.1 0.8 1.4 47.8 39.0 56.6 1.1 0.7 1.7 41.8 35.3 48.2 1.1 0.8 1.5
Rural 45.9 40.0 51.7 1.1 0.8 1.4 47.9 38.1 57.7 1.0 0.6 1.4 44.3 36.9 51.7 1.2 0.9 1.7
New condition
Heart condition/Stroke†59.3 52.3 66.4 1.0 ... ... 63.3 54.4 72.3 1.0 ... ... 54.5 43.4 65.5 1.0 ... ...
Cancer 56.8 48.5 65.0 0.9 0.5 1.4 56.3 43.6 69.0 0.7 0.4 1.4 57.2 46.7 67.6 1.1 0.6 2.1
Skin condition 42.2*33.4 51.0 0.5*0.3 0.8 45.4*32.0 58.7 0.4*0.2 0.9 39.0*28.3 49.7 0.5*0.3 1.0
Cataract or other eye condition 46.4*36.8 56.0 0.4*0.3 0.7 54.6 40.0 69.3 0.5*0.2 1.0 41.8 29.9 53.7 0.4*0.2 0.8
Arthritis/Rheumatism 28.9*20.3 37.5 0.3*0.2 0.5 x x x 0.4*0.2 0.9 24.6E*13.9 35.2 0.3*0.1 0.5
Mental health disorder 48.7 37.0 60.3 0.7 0.4 1.1 51.5 36.4 66.7 0.6 0.2 1.4 47.2E31.5 63.0 0.7 0.3 1.4
Asthma or other breathing condition 45.0*33.9 56.1 0.5*0.3 0.8 x x x 0.4*0.2 0.8 43.7E28.1 59.4 0.6 0.3 1.4
Gynecological condition 39.1*32.7 45.5 0.4*0.3 0.6 ... ... ... ... ... ... 39.1*32.7 45.5 0.4*0.3 0.8
Other 44.4*41.5 47.3 0.5*0.3 0.6 48.7*44.1 53.4 0.5*0.3 0.7 40.6*36.9 44.3 0.5*0.3 0.8
High blood pressure
No†46.8 44.3 49.3 1.0 ... ... 54.0 49.9 58.2 1.0 ... ... 42.1 38.9 45.2 1.0 ... ...
Yes 41.4*37.4 45.5 0.7*0.5 0.9 41.4*35.0 47.8 0.6*0.4 0.8 41.4 36.3 46.6 0.8 0.6 1.1
Number of selected chronic conditions‡
None†46.9 43.9 50.0 1.0 ... ... 54.2 49.1 59.3 1.0 ... ... 41.7 37.9 45.5 1.0 ... ...
One 43.8 40.0 47.6 0.9 0.7 1.1 48.4 42.4 54.4 0.8 0.6 1.1 40.9 36.2 45.6 0.9 0.7 1.2
2 or more 45.3 40.6 50.0 0.9 0.7 1.1 44.5 36.2 52.8 0.7 0.5 1.2 45.8 39.9 51.7 1.0 0.7 1.4
Has regular family doctor
Yes†44.3 42.1 46.5 1.0 ... ... 48.5 44.8 52.3 1.0 ... ... 41.5 38.7 44.4 1.0 ... ...
No 58.9*51.0 66.8 1.2 0.8 1.7 69.2*58.4 80.0 1.6 1.0 2.7 47.3 36.6 58.1 0.9 0.5 1.4
Source of specialist referral
Family doctor†38.9 36.5 41.3 1.0 ... ... 43.5 39.4 47.6 1.0 ... ... 35.9 32.9 38.9 1.0 ... ...
Other specialist 52.3*45.1 59.5 1.7*1.3 2.4 58.3*45.7 70.9 1.7*1.0 2.8 48.2*40.1 56.3 1.6*1.1 2.3
Other health care provider 58.7*52.4 65.0 2.2*1.7 3.0 62.4*53.0 71.8 2.0*1.3 3.1 55.7*47.0 64.4 2.5*1.7 3.7
Did not require referral 71.6*63.9 79.3 4.1*2.8 6.0 79.7*69.6 89.9 5.4*2.9 10.0 65.7*55.1 76.4 3.8*2.4 6.1
† reference group
‡ selected chronic conditions were asthma, arthritis, cancer, chronic obstructive pulmonary disease, diabetes, heart disease and mood disorders (depression, bipolar disorder, mania, dysthymia)
* signifi cantly different from estimate for reference category, denoted by † (p < 0.05)
E use with caution (coeffi cient of variation 16.6% to 33.3%)
x suppressed to meet confi dentiality requirements of Statistics Act
... not applicable
Notes: Estimates are based on population who completed initial consultation with specialist in previous 12 months and provided information about waiting time. Except for total household income, analyses exclude
non-response (“don’t know,” “not stated,” and “refusal”). Modelled odds ratio results for total cohort model are based on information from 5,397 respondents (1,982 male model; 3,415 female model).
Source: 2007 Canadian Community Health Survey.
12 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Waiting time for medical specialist consultations in Canada, 2007 • Research article
signifi cantly lower for men and women
with skin and eye conditions and arthritis/
rheumatism. As well, men with asthma
or other breathing conditions and women
with gynecological conditions had lower
odds of consulting a specialist within a
month.
The sex-specifi c multivariate results
also confi rm the importance of referral
source in waiting times for both male
and female patients. Compared with
patients referred by their family doctor,
those referred by another specialist or
another health care provider had about
twice the odds of seeing a specialist
within a month. And among patients
who indicated no referral was required,
the odds of seeing a specialist within a
month were fi ve times higher for men
and almost four times higher for women,
compared with those referred by their
family doctor.
For men, but not women, several other
factors were signifi cantly associated
with waiting time to see a specialist.
Among male patients, the odds of seeing
a specialist within a month were twice
as high for those who had immigrated
more than 10 years earlier than for those
who were Canadian-born. As well, male
patients reporting high blood pressure
had signifi cantly low odds of seeing a
specialist within a month, compared
with those without high blood pressure.
Household income was also signifi cant
for male patients. Compared with men
in the top income quintile, those in the
lowest were less likely to see a specialist
within a month. Yet this was also true for
men in the second-highest quintile.
Discussion
This national study identifi es patient-
and provider-related factors associated
with waiting time for an initial specialist
consultation about a new condition. Not
surprisingly, waiting time varied with the
nature of that condition, with generally
shorter waits for those that were
potentially life-threatening. But even
when the infl uence of this variable was
taken into account, the results highlight
signifi cant differences in waiting times
by sex, source of referral, and for
male patients, household income and
immigration status.
Women were signifi cantly less
likely than men to see a specialist
within a month. This could result from
systemic gender biases in access to
health care services, evidence of which
has previously been demonstrated. For
example, gender differences in access
to primary care for heart disease have
been reported , 20-22 including physicians’
diagnostic and management practices.22
Differential access to specialized
cardiovascular care based on non-clinical
patient attributes, such as social status,
has also been reported.23
However, the disparity between male
and female patients in waiting time may
refl ect differences in the severity of the
condition that prompted the specialist
consultation. Because information
about the patients’ health status before
the visit is limited, and no measure of
the severity of the new condition is
available, it was not possible to fully
adjust for health status. It may be that
men’s shorter waiting time for specialist
consultations was attributable to more
advanced conditions. Men are less likely
than women to use physician services or
to have a regular family doctor,12,24 and
consequently, may have less continuity
of primary care. As a result, men may
present at more advanced stages of
disease and require expedited specialist
consultations.
Differences in specialist waiting
time by immigration status among male
patients could also refl ect greater severity
of the emergent health condition. This is
consistent with well-known associations
between immigration status and changes
to health over time, as well as differences
in the use of and access to care among the
immigrant population. Immigrants tend
to have lower health literacy,,25 which
may contribute to less use of preventive
care. For example, signifi cantly lower
rates of cancer screening have been
found among visible minorities, a large
proportion of whom are immigrants. 26
And although recent immigrants tend to
be in better health than the Canadian-
born population, over time they are more
likely to report health deterioration.27
Therefore, the differences among men in
waiting time for specialist consultations
by immigrant status may be attributable
to medical need.
Finally, the results of this study
demonstrate the importance of the
referral source in specialist waiting time.
There may be several explanations.
First, the referral source may indicate the
point at which a patient is located in the
pathway of diagnosis and treatment. The
question on the Canadian Community
Health Survey pertained to waiting
time for a specialist consultation in the
previous 12 months, but the specifi c visit
What is already
known on this
subject?
■Waiting time for an initial consultation
with a specialist can constitute a
substantial part of the continuum of
care.
■Little known about factors associated
with waiting time for specialist
consultations.
What does this study
add?
■This study identifies factors
associated with shorter versus
longer waiting times for specialist
consultations.
■As might be expected, the nature
of the health condition prompting
the consultation was significantly
associated with how long patients
waited.
■Women tended to wait longer than
men.
■For both sexes, waiting time varied
significantly depending on the source
of referral.
■For men only, household income
and immigrant status were significant
factors in waiting time.
13
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Waiting time for medical specialist consultations in Canada, 2007 • Research article
about which the respondent answered
might have resulted from prior visits to
other specialists to confi rm a diagnosis.
This process may have yielded greater
diagnostic certainty, and perhaps,
infl uenced assessed urgency. General
practitioners often use referrals to obtain
assistance with diagnoses or advice
about therapy. 28 Therefore, diagnostic
uncertainty may be involved when a
family physician is the referral source,
and may affect assessed urgency. As well,
suboptimal communication between
general practitioners and specialists has
been cited as a diffi culty in the referral
process.29 Recently, referral tools
have been developed to improve and
standardize communications between
general practitioners and specialists.15
Respondents who reported not
needing a referral had much higher odds
of seeing the specialist within a month,
compared with those referred by a family
doctor. To some extent, this may be
attributable to provincial variations in
how services are organized, especially
in Quebec. Relatively high percentages
of Quebec residents reported referral
sources other than family doctors or
self-referrals. According to recent
data,24 Quebec residents are less likely
than people in other provinces to have a
regular family doctor, and so may rely on
nurses30 or other health care professionals
working in primary health care teams in
community health centres. This, in turn,
may facilitate referral to specialists.
Limitations
The distribution of waiting times
reported to the Canadian Community
Health Survey was skewed. To attempt
to preserve the continuous nature of the
data, a logarithmically transformed,
continuous dependent variable using
linear regression was employed. But
because respondents could report waiting
times in days, weeks or months, and
because they tended to round responses,
especially for longer waits, these models
were diffi cult to fi t. Consequently, a
dichotomous outcome was derived, and
logistic regression analyses were used.
The study is based on self-reported
data that were not clinically validated
and may be subject to recall bias.
As well, if respondents interpreted
the question, “In the past 12 months,
did you require a visit to the specialist
for a diagnosis or consultation for a
new illness or condition?” to mean that
both the occurrence of the new illness
and the consultation arose within the
past 12 months, some respondents with
long waiting times may not have been
captured.
Clinical need is a crucial indicator
of waiting time for health care services,
so it was expected that those in greater
need would wait less time for a specialist
consultation. However, it was not
possible to fully adjust for the need for
services because information about the
severity of the new or existing conditions
is not available from these data.
Conclusion
Waiting time for specialist services
represents a key indicator of access to
health care in Canada. Data from the
Canadian Community Health Survey
provide a unique opportunity to explore
factors associated with how long patients
wait for specialist care.
The results of this study suggest that,
in addition to the nature of the new
condition, gender and referral source are
associated with obtaining a consultation
within a month. And for males,
household income and immigration
status are also signifi cant.
This is only a preliminary examination
of factors related to waiting time for
specialist consultations; more research is
obviously required. In particular, given
the apparent importance of the source
of referral, future analyses might focus
on decision-making and communication
processes to determine how urgency is
assessed and how patients are triaged for
specialist consultations. The fi ndings
from this and subsequent research may
be relevant to a better understanding of
the role of different health care providers
in accessing specialists and how these
processes vary across patient groups.■
14 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Waiting time for medical specialist consultations in Canada, 2007 • Research article
References
1. Sanmartin C, Houle C, Berthelot J-M,
et al. Access to Health Care Services in
Canada, 2001 (Statistics Canada, Catalogue
82-575-XIE) Ottawa: Statistics Canada, 2002.
2. Sanmartin C, Berthelot J-M. Access to
Health Care Services in Canada. January to
December 2005 (Statistics Canada, Catalogue
82-575-XIE) Ottawa: Statistics Canada, 2006.
3. Health Council of Canada. Wading through
Waiting Times – What Do Meaningful
Reductions and Guarantees Mean? An Update
on Waiting Times for Health Care, June 2007.
Available at: http: //www.healthcouncilcanada.
ca/docs/rpts/2007/Waiting_times/
hcc_Waiting-times-update_200706_
Final%20ENGLISH.pdf. Accessed February
26, 2008.
4. Canadian Institute for Health Information.
Waiting Time Tables - A Comparison by
Province, 2008. Available at: http: //www.
secure.cihi.ca/cihiweb/products/Final_
Waiting_Times_AIB_EN_080229.pdf.
Accessed March 16, 2008.
5. Canadian Institute for Health Information.
Waiting for Health Care in Canada: What We
Know and What We Don’t Know. Available at:
http: //www.secure.cihi.ca/cihiweb/products/
WaitingTimesReport_06_e.pdf. Accessed
March 16, 2008.
6. Canadian Intergovernmental Conference
Secretariat. A 10-Year Plan to Strengthen
Health Care. Available at: http://www.scics.
gc.ca/cinfo04/800042005_e.pdf. Accessed
June 06, 2008.
7. Sanmartin C. Toward standard definitions
of waiting times. Healthcare Management
Forum 2006; web release.
8. Andersen R, Newman J. Societal and
individual determinants of medical care
utilization in the United States. Milbank
Memorial Fund Quarterly 1973; 51: 95-124.
9. Alter DA, Iron K, Austin PC, et al.
Socioeconomic status, service patterns, and
perceptions of care among survivors of acute
myocardial infarction in Canada. Journal of
the American Medical Association 2004; 291:
1100-7.
10. Dunlop S, Coyte PC, McIssac W.
Socio-economic status and the utilization of
physicians’ services: results from the Canadian
National Population Health Survey. Social
Science and Medicine 2000; 51(1): 123-33.
11. Veugelers PJ, Yip AM. Socioeconomic
disparities in health care use: Does universal
coverage reduce inequalities in health? Journal
of Epidemiology and Community Health 2003;
57(6): 424-8.
12. Nabalamba A, Millar WJ. Going to the doctor.
Health Reports (Statistics Canada, Catalogue
82-003) 2007; 18(1): 23-33.
13. Chan B, Austin PC. Patient, physician,
and community factors affecting referrals
to specialists in Ontario, Canada: a
population-based, multi-level modelling
approach. Medical Care 2003; 41(4): 500-11.
14. Suter LG, Fraenkel L, Holmboe ES. What
factors account for referral delays for patients
with suspected rheumatoid arthritis? Arthritis
and Rheumatism 2006; 55(2): 300-5.
15. De Coster C, Fitzgerald A, Cepoiu M.
Priority-setting tools for rheumatology
disease referrals: a review of the literature.
The Journal of Clinical Rheumatology 2008;
27(11): 1411-6.
16. Béland Y. Canadian Community Health
Survey – Methodological overview. Health
Reports (Statistics Canada, Catalogue 82-003)
2002; 13(3): 1-14.
17. Canadian Community Health Survey (CCHS)
2007 Microdata Files User Guide (June
2008). Available at: http://www.statcan.gc.ca/
cgi-bin/af-fdr.cgi?l=eng&loc=http://www.
statcan.gc.ca/imdb-bmdi/document/3226_
D7_T9_V4-eng.pdf&teng=User%20
Guide&tfra=Guide%20de%20l’utilisateur.
18. Rao JNK, Wu CFJ, Yue K. Some recent work
on resampling methods for complex surveys.
Survey Methodology (Statistics Canada,
Catalogue 12-001) 1992; 18(2): 209-17.
19. Rust KF, Rao JNK. Variance estimation for
complex surveys using replication techniques.
Statistical Methods in Medical Research 1996;
5: 281-310.
20. Adams A, Buckingham CD, Lindenmeyer
A, et al. The influence of patient and doctor
gender on diagnosing coronary heart disease.
Sociology of Health & Illness 2008; 30 (1):
1-18.
21. Arber S, McKinlay J, Adams A, et al. Patient
characteristics and inequalities in doctors’
diagnostic and management strategies relating
to CHD: A video-simulation experiment.
Social Science and Medicine 2006; 62(1):
103-15.
22. Richards H, McConnachie A, Morrison
C, et al. Social and gender variation in
the prevalence, presentation and general
practitioner provisional diagnosis of chest
pain. Journal of Epidemiology and Community
Health 2000; 54: 714-8.
23. Alter, DA, Basinski ASH, Naylor CD. A
survey of provider experiences and perceptions
of preferential access to cardiovascular care in
Ontario, Canada. Annals of Internal Medicine
1998; 129: 567-72.
24. Statistics Canada. Health Indicators 2008.
Available at: http://www.statcan.gc.ca/
daily-quotidien/080618/dq080618a-eng.htm.
Accessed June 18, 2008.
25. Rootman I, Gordon-El-Bihbety D. A Vision
for a Health Literate Canada. Report of the
Expert Panel on Health Literacy. Ottawa:
Canadian Public Health Association, 2008.
26. Quan H, Fong A, De Coster C, et al. Variation
in health services utilization among ethnic
populations. Canadian Medical Association
Journal 2006; 174 (6): 787-90.
27. Ng E, Wilkins R, Gendron F, et al. Dynamics
of Immigrants’ Health in Canada: Evidence
from the National Population Health
Survey (Statistics Canada, Catalogue
82-618-MWE2005002) Ottawa: Statistics
Canada, 2005.
28. Donohoe MT, Kravitz RL,Wheeler DB,
et al. Reasons for outpatient referrals from
generalists to specialists. Journal of General
Internal Medicine 1999; 14(5): 281-6.
29. Gandhi TK, Sittig DF, Franklin M, et al.
Communication breakdown in the outpatient
referral process. Journal of General Internal
Medicine 2000; 15(6): 626-31.
30. Carrière G. Consultations with doctors and
nurses. Health Reports (Statistics Canada,
Catalogue 82-003) 2005; 16(4): 45-8.
15
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Blood pressure in Canadian children and adolescents • Research article
Blood pressure in Canadian children and
adolescents
by Gilles Paradis, Mark S. Tremblay, Ian Janssen, Arnaud Chiolero and Tracey Bushnik
Abstract
Background
Because blood pressure (BP) tracks from childhood
to adulthood, assessing levels in youth is relevant.
There are no recent BP data for Canadian children
and adolescents, and past studies have used a
variety of design and measurement devices.
Data and methods
With a clinically validated oscillometric device, resting
BP was measured in 2,079 respondents aged 6 to 19
years from the Canadian Health Measures Survey.
The average of the last fi ve of six BP measures
taken one minute apart at a single visit was used in
this report. Borderline or elevated BP was defi ned
as greater than or equal to the 90th percentile of US
reference values for participants aged 6 to 17 years.
Borderline or elevated BP for 18- to 19-year-olds was
defi ned as equal to or greater than 120 systolic BP or
equal to or greater than 80 diastolic BP. Participants
of any age who reported taking antihypertensive
medication in the past month were also defi ned as
having elevated BP.
Results
At ages 6 to 11 years, mean (standard error) systolic/
diastolic blood pressure was 93(0)/61(1) in boys and
93(0)/60(0) mmHg in girls, and at ages 12 to 19 years,
101(1)/63(1) and 98(1)/63(1) mmHg, respectively.
An estimated 2.1% (95% confi dence interval 1.3% to
3.0%) of Canadian children and youth had borderline
levels; 0.8% (0.4% to 1.4%) had elevated BP.
Interpretation
Despite the prevalence of obesity among young
people, BP levels were lower than reported in
provincial samples, which may, in part, refl ect
differences in methodologies and measurement
instruments.
Keywords
diastolic pressure, hypertension, obesity, overweight,
survey, systolic pressure
Authors
Gilles Paradis (1-514-398-1418; gilles.paradis@
mcgill.ca) is with the Department of Epidemiology,
Biostatistics and Occupational Health, McGill
University, Health Centre Research Institute and the
Institut national de santé publique du Québec; Mark
S. Tremblay is with the Children’s Hospital of Eastern
Ontario Research Institute; Ian Janssen is with
Queen’s University; Arnaud Chiolero is with McGill
University; and Tracey Bushnik is with Statistics
Canada.
o nationally representative blood pressure (BP)
data for Canadian children and adolescents
have been collected since the 1978 Canada Health
Survey.1 With the results of the 2007-2009 Canadian
Health Measures Survey (CHMS), launched by
Statistics Canada in partnership with Health Canada
and the Public Health Agency of Canada, it is possible
to address this data gap.2-5 The CHMS is the most
comprehensive direct health measures survey ever
conducted in Canada. In addition to a detailed health
interview, the survey involves direct measurement
of indicators and of risk factors for chronic diseases,
infectious diseases, environmental exposures,
nutritional status, physical activity and physical
fi tness.2-5
N
Elevated BP is one of the most important
causes of death and disability worldwide,6
accounting for 7.6 million premature
deaths and 92 million disability-adjusted
life years annually. In adolescence,
hypertension is associated with
increased left ventricular mass, diastolic
dysfunction,7 fatty streaks and fi brous
plaques in the coronary arteries and the
aorta,8 and arterial wall thickening.9
BP levels track from childhood to
adulthood,10,11 indicating that elevated
BP at young ages is a risk factor for
the development of hypertension in
adulthood. The strength of BP tracking
increases with body mass index (BMI),
such that tracking is strongest in
overweight and obese youth.12,13
Population information about BP
levels in children and adolescents can be
useful from a public health and clinical
perspective to guide prevention planning,
help establish norms, and monitor trends
over time. However, Quebec is the only
Canadian province to have relatively
recent measures for youth: in 1999,
12% to 23% of youth aged 9, 13 and
16 years had high-normal or elevated
16 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Blood pressure in Canadian children and adolescents • Research article
BP.9 A 2004 study of BP levels among
American youth found that from 1988-
1994 to 1999-2000, mean systolic blood
pressure (SBP) increased 1.4 mmHg, and
mean diastolic blood pressure (DBP)
increased 3.3 mmHg.15 A longer-term
review of trends in American youth from
1963 to 2002 also demonstrated a slight
upturn in the prevalence of elevated BP
in the last decade.16 But such fi ndings
have not been consistent. For example, a
study of 15-year-old Russian adolescents
between 1995 and 2004 found a
signifi cant decrease in DBP among boys,
and a signifi cant decrease in SBP among
both sexes.17 As well, comparisons of
results from past studies are complicated
by different survey methods, including
different measurement devices.
Based on data from the 2007-
2009 CHMS, this study presents BP
distributions and estimates of elevated
BP for a representative sample of
Canadian children and adolescents aged
6 to 19 years.
Methods
Data source
Data are from cycle 1 of the Canadian
Health Measures Survey (CHMS), which
collected information at 15 sites from
March 2007 through February 2009.
The CHMS covered the population aged
6 to 79 years living in private households.
Residents of Indian Reserves or Crown
lands, institutions and certain remote
regions and full-time members of the
regular Canadian Forces were excluded.
Approximately 96.3% of Canadians
were represented.18
Health Canada’s Research Ethics
Board gave ethics approval to conduct
the survey. Informed written consent
was obtained from respondents aged 14
years or older. For younger children,
a parent or legal guardian provided
written consent, in addition to written
assent from the child. Participation was
voluntary; respondents could opt out of
any part of the survey at any time.
The response rate for households
selected for inclusion in the CHMS
was 69.6%—meaning that in 69.6% of
selected households, the sex and date
of birth of all household members were
provided by a household resident. In
each responding household, one or two
members were chosen to participate;
88.5% of selected 6- to 19-year-olds
completed the household questionnaire,
and 86.9% of those who completed
the questionnaire participated in
the subsequent examination centre
component. The fi nal response rate for
6- to 19-year-olds, after adjusting for
the sampling strategy, was 53.5%. This
article is based on 2,079 examination
centre respondents aged 6 to 19 years
(after removing 8 with missing BP data)
(Appendix Table A).
Measures
At the respondent’s home, an interviewer
administered a questionnaire covering
socio-demographic characteristics,
medical history, current health status
and lifestyle behaviours (Table 1). In
the chronic conditions component of
the questionnaire, respondents aged 12
years or older were asked if they had high
BP (diagnosed by a health professional
and expected to last or had already lasted
six months or more) and if they had taken
“medicine for high blood pressure” in the
past month.
One day to six weeks after the
home interview, the respondent visited
a mobile examination centre for a
battery of physical measurements,
including anthropometry, BP, heart rate,
spirometry, physical fi tness, oral health
and biospecimen collection.4 BMI
was calculated as weight in kilograms
divided by height in meters squared (kg/
m2), and respondents were classifi ed as
overweight, obese, or neither.19,20 BP
was measured after urine collection,
but before blood collection and fi tness
testing.4
BP and heart rate were measured with
the BpTRU™ BP-300 (BpTRU Medical
Devices Ltd., Coquitlam, British
Columbia). The BpTRU™, an automated
electronic monitor, automatically infl ates
and defl ates the upper-arm cuff and uses
an oscillometric technique to calculate
SBP and DBP. It has passed international
validation protocols for accuracy.21,22
An advantage of an automated device
is that it enables BP to be measured
in the absence of another person,
thereby eliminating observer errors
such as digit bias, zero preference and
incorrect defl ation rates, and reducing
“white-coat hypertension” (a rise in
BP associated with the presence of
the health care professional and the
measurement procedures).23 For more
detailed information on the procedures
and protocol, including staff training,
equipment calibration, and quality
assurance and control, see Resting blood
pressure and heart rate measurement in
the Canadian Health Measures Survey
cycle 1.24
Defi nitions
Measures of SBP and DBP were
calculated as the average of the
fi rst set (last fi ve of six measures
taken one minute apart) of valid BP
measurements.24 For those aged 6 to
17 years, based on age and sex, each
respondent’s height and average SBP
and DBP were converted to z-scores,
which were used to calculate individual
BP percentiles as per the equations in
Appendix B of the fourth report of the
National High Blood Pressure Education
Program Working Group on High Blood
Pressure in Children and Adolescents
(NHBPEP4).25 With these calculated
percentiles, children and youth in this
age group were classifi ed into BP
categories. As well, respondents who
reported taking medicine for high BP
in the past month were classifi ed as
having “elevated” BP, regardless of
their BP percentile value (fewer than 10
respondents). The seventh report of the
Joint National Committee on Prevention,
Detection, Evaluation and Treatment of
High Blood Pressure (JNC7) was used to
classify youth aged 18 or 19 years.26 The
NHBPEP4 classifi cation parallels that of
the JNC7.
Normal BP for respondents aged 6
to 17 years was defi ned as a calculated
SBP percentile and DBP percentile less
than the 90th percentile. For respondents
17
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Blood pressure in Canadian children and adolescents • Research article
Table 1
Selected characteristics of sample (weighted), by age group and sex, household population aged 6 to 19 years, Canada,
March 2007 to February 2009
Age group (years)
6 to 11 12 to 19
Boys
95%
confidence
interval
Girls
95%
confidence
interval
Boys
95%
confidence
interval
Girls
95%
confidence
interval
from to from to from to from to
Mean age (years) 8.6 8.2 9.0 8.7 8.5 8.9 15.2 15.0 15.5 15.7 15.5 16.0
Measured body mass index (kg/m2) 17.9 17.5 18.2 17.7 17.2 18.1 22.6 21.6 23.6 22.4 21.7 23.0
Overweight† (%) 17.3 13.5 21.9 16.3 11.6 22.4 18.0 14.6 21.9 17.6 13.5 22.7
Obese† (%) 7.1 5.1 9.8 5.8E3.2 10.1 12.3E6.6 21.7 8.3 5.8 11.6
Smoke daily (%) … … … … … … 8.0E4.0 15.2 6.8E3.3 13.4
Physically active†† (%) 84.5 80.8 87.6 82.5 79.0 85.6 77.4 69.3 83.9 65.0 59.5 70.2
Immediate family history of high blood pressure (%) 12.6 8.8 17.7 15.9 12.0 20.8 25.8 20.3 32.2 22.1 16.3 29.3
Household education more than secondary graduation (%) 88.5 84.6 91.5 85.3 79.1 89.9 86.6 83.4 89.2 83.0 74.2 89.2
Household type - couple with children (%) 82.6 77.1 87.1 79.4 72.9 84.6 72.5 65.7 78.4 76.3 68.5 82.7
Low household income††† (%) 7.7E4.4 12.9 6.5E3.8 11.1 5.4E3.0 9.6 11.2E7.1 17.2
Born in Canada (%) 92.4 81.8 97.0 92.0 81.8 96.7 90.5 77.5 96.4 88.3 80.2 93.3
† 18- to 19-year-olds classifi ed as overweight (BMI 25 to 29.9 kg/m2) or obese (BMI 30 kg/m2 or more)(Source: Health Canada. Canadian Guidelines for Body Weight Classifi cation in Adults (Catalogue
H49-179) Ottawa: Health Canada, 2003); 6- to 17-year-olds classifi ed as overweight or obese based on defi nitions proposed by International Obesity Task Force (Source: Cole TJ, Bellizzi MC, Flegal KM,
et al. Establishing a standard defi nition for child overweight and obesity worldwide: international survey. British Medical Journal 2000; 320(7244): 1240-3).
†† for ages 6 to 11, physically active for at least 60 minutes 4 or more days in typical week; for ages 12 to 19, categorized as “active”or “moderately active according to Physical Activity Index
†††
based on household size and income range; denominator is 1,920 respondents with valid household income value
E interpret with caution (coeffi cient of variation 16.6% to 33.3%)
... not applicable
Source: 2007 to 2009 Canadian Health Measures Survey.
aged 18 or 19 years, it was defi ned as
a measured mean SBP less than 120
mmHg and a measured mean DBP less
than 80 mmHg. This corresponds to
the “normal” category proposed by the
NHBPEP4 and JNC7.
Borderline BP for respondents aged 6
to 17 years was defi ned as a calculated
SBP percentile or DBP percentile greater
than or equal to the 90th percentile, but
less than the 95th percentile, or a measured
SBP/DBP greater than 120/80 mmHg,
even if less than the 90th percentile. For
respondents aged 18 or 19 years, it was
defi ned as a measured mean SBP of 120
to 139 mmHg and a measured mean DBP
of 80 to 89 mmHg; or SBP of 120 to 139
mmHg and DBP lower than 80 mmHg;
or SBP lower than 120 mmHg and DBP
80 to 89 mmHg. This corresponds to the
“prehypertension” category proposed by
the NHBPEP4 and JNC7.
Elevated BP for respondents aged 6
to 17 years was defi ned as a calculated
SBP percentile or DBP percentile greater
than or equal to the 95th percentile, or
the respondent’s report of using BP
medication in the past month. For
respondents aged 18 or 19 years, elevated
BP was defi ned as a measured mean
SBP/DBP of 140/90 mm Hg or higher, or
the respondent’s report of BP medication
use in the past month. This corresponds
to the “Stage 1 or Stage 2 hypertension”
category proposed by the NHBPEP4 and
JNC7.
Analytical techniques
Weighted data were analyzed
separately by sex and age. Estimates of
proportions, means, standard errors, and
percentiles were produced. Standard
errors, coeffi cients of variation and 95%
confi dence intervals (CI) were estimated
using bootstrap weights to account for the
complex survey design of the CHMS.27,28
Gender differences in SBP and DBP
were tested using t-tests. Analyses were
conducted with SUDAAN.
Results
Mean SBP (standard error) rose with age
from 91(1) mmHg among boys aged 6
to 7 years to 104(1) mmHg at 18 to 19
years; for girls, the increase was from
92(1) to 99(1) mmHg (Table 2). Mean
SBP was similar in boys and girls from
ages 6 to 7 through 10 to 11 years, and
also at 14 to 15 years. However, at 12
to 13 years and 16 through 19 years,
mean SBP was higher in boys (p<0.01).
Median SBP was very close to the mean
in all age/sex categories.
The sample size was too small to
obtain percentile values by single-year-
of-age or 95th percentile values for most
two-year age groups. At ages 6 to 11
years, the 95th percentile (95% CI) for
SBP was 105 (102 to 107) mmHg among
boys and 106 (104 to 108) mmHg among
girls; at ages 12 to 19 years, the 95th
percentile for SBP was 116 (113 to 119)
mmHg among boys and 111 (108 to 114)
mmHg among girls.
Mean DBP also rose with age, but
not as much as SBP (Table 3). From
ages 6 to 7 to 18 to 19 years, mean DBP
increased from 59(1) to 65(1) mmHg
among boys and from 60(1) to 64(1)
among girls. Mean DBP was similar in
both sexes. Median DBP was very close
to the mean in all age/sex groups.
In 2007-2009, few Canadian children
and adolescents had borderline or
elevated BP: 3.7% (2.3% to 6.0%) at
ages 6 to 11 years and 2.2% (1.2% to
4.0%) at 12 to 19 years (Table 4).
Mean SBP was higher among children
and adolescents who were overweight or
18 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Blood pressure in Canadian children and adolescents • Research article
Table 3
Percentile distribution of measured diastolic blood pressure (DBP) (mmHg) values, by sex and two-year age group,
household population aged 6 to 19 years, Canada, March 2007 to February 2009
Sex and
two-year
age group
(years)
Sample
size Mean
Standard
error
25th percentile 50th percentile 75th percentile 90th percentile 95th percentile
Value
95%
confidence
interval
Value
95%
confidence
interval
Value
95%
confidence
interval
Value
95%
confidence
interval
Value
95%
confidence
interval
from to from to from to from to from to
Boys 1,051 62 1 57 55 58 61 60 63 66 65 68 71 69 72 74 72 76
6 to 11 538 61 1 56 55 57 60 58 61 64 62 65 68 66 71 72 68 75
6-7 164 59 1 53 48 58 58 56 60 63 60 65 69 64 74 F … …
8-9 172 61 1 57 55 58 60 58 62 63 61 66 69 65 73 F … …
10-11 202 61 0 57 55 58 61 60 62 65 63 66 68 66 69 71 67 74
12 to 19 513 63 1 57 56 59 63 61 65 67 65 69 72 70 74 76 73 79
12-13 160 62 1 57 56 58 61 59 63 65 63 66 69 66 71 F … …
14-15 119 62 1 55 51 58 62 59 66 68 64 71 73 69 78 F … …
16-17 139 64 1 58 56 60 63 60 66 68 66 71 73 70 76 F … …
18-19 95 65 1 60 56 64 64 62 67 69 66 73 F … … F … …
Girls 1,028 62 0 56 55 57 61 60 62 66 65 68 71 70 72 74 73 75
6 to 11 529 60 0 55 53 56 60 59 61 65 63 66 70 69 70 72 70 74
6-7 159 60 1 54 51 57 59 56 62 64 61 66 69 65 74 F … …
8-9 157 61 1 55 51 59 61 60 62 65 62 68 70 68 72 F … …
10-11 213 60 1 55 52 57 60 58 62 65 63 66 68 67 70 70 68 72
12 to 19 499 63 1 57 55 58 62 60 64 67 65 69 72 70 73 74 73 75
12-13 132 60 1 54 53 56 59 57 62 65 63 67 67 66 68 F … …
14-15 126 63 1 57 56 59 62 60 64 68 64 72 74 70 79 F … …
16-17 127 64 1 59 57 61 63 59 66 69 66 71 71 70 73 F … …
18-19 114 64 1 57 53 60 64 61 67 68 67 70 73 72 75 F … …
F too unreliable to be reported (coeffi cient of variation greater than 33% or small sample size)
... not applicable
Source: 2007 to 2009 Canadian Health Measures Survey.
Table 2
Percentile distribution of measured systolic blood pressure (SBP) (mmHg) values, by sex and two-year age group,
household population aged 6 to 19 years, Canada, March 2007 to February 2009
Sex and
two-year
age group
(years)
Sample
size Mean
Standard
error
25th percentile 50th percentile 75th percentile 90th percentile 95th percentile
Value
95%
confidence
interval
Value
95%
confidence
interval
Value
95%
confidence
interval
Value
95%
confidence
interval
Value
95%
confidence
interval
from to from to from to from to from to
Boys 1,051 98 1 91 90 92 96 95 98 103 101 106 110 107 113 114 112 116
6 to 11 538 93 0 88 87 89 92 91 93 97 95 99 101 99 103 105 102 107
6-7 164 91 1 85 84 87 90 89 92 94 92 97 99 95 103 F … …
8-9 172 93 1 88 87 90 91 90 93 96 95 98 101 98 104 F … …
10-11 202 95 1 90 88 92 94 93 96 98 97 100 103 99 106 105 103 108
12 to 19 513 101 1 94 92 96 100 97 103 106 103 109 113 110 116 116 113 119
12-13 160 97 1 90 88 91 96 92 100 103 100 107 109 105 114 F … …
14-15 119 100 1 93 91 96 98 94 103 105 99 112 112 107 118 F … …
16-17 139 104 2 97 94 100 102 99 106 107 102 111 116 109 123 F … …
18-19 95 104 1 97 95 100 103 99 107 109 105 114 F … … F … …
Girls 1,028 96 0 90 89 91 95 94 96 100 99 101 106 104 108 109 108 111
6 to 11 529 93 0 87 86 88 92 91 94 98 97 99 103 102 104 106 104 108
6-7 159 92 1 86 84 88 90 87 93 96 92 100 102 98 106 F … …
8-9 157 94 1 88 85 91 93 91 95 98 97 100 104 100 108 F … …
10-11 213 94 1 88 87 89 94 91 96 99 97 101 103 100 105 105 103 107
12 to 19 499 98 1 92 90 94 96 95 98 102 100 103 108 106 110 111 108 114
12-13 132 94 1 90 88 91 94 91 96 97 94 100 101 98 104 F … …
14-15 126 99 1 94 92 96 97 94 100 103 98 107 109 103 114 F … …
16-17 127 98 1 93 91 96 97 95 99 101 99 103 109 105 113 F … …
18-19 114 99 1 93 89 97 98 95 102 104 101 106 109 106 113 F … …
F too unreliable to be reported (coeffi cient of variation greater than 33% or small sample size)
... not applicable
Source: 2007 to 2009 Canadian Health Measures Survey.
19
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Blood pressure in Canadian children and adolescents • Research article
obese (Table 5). The SBP differences
between BMI categories reached
statistical signifi cance among boys aged
12 to 19 years, girls aged 6 to 11 years,
and in both age groups when the genders
were combined. Differences in DBP
by BMI category were less apparent,
reaching statistical signifi cance only
among obese boys aged 12 to 19 years.
Discussion
The main fi nding of this analysis is
the remarkably low overall prevalence
of borderline or elevated BP among
Canadian children and adolescents.
However, echoing the results of other
studies,14,15 mean SBP was signifi cantly
higher among boys aged 12 to 19 years
and girls aged 6 to 11 years who were
overweight or obese. Excess weight
is believed to infl uence BP through
increased sympathetic nervous system
activation, which is associated with SBP.
The association of weight with DBP was
much less pronounced.
The generally low levels of BP
obtained from the CHMS appear
inconsistent with the rise of childhood
and adolescent obesity in Canada.29,30
And despite a trend toward excess weight
among youth in other countries, BP levels
have not shown consistent increases.31
Table 4
Percentage distribution of measured
blood pressure status, by sex and age
group, household population aged 6
to 19 years, Canada, March 2007 to
February 2009
Sample
size %
95%
confidence
interval
from to
Total
Normal 2,019 97.2 96.1 98.0
Borderline 47 2.1E1.3 3.1
Elevated 13 0.8E0.4 1.4
Boys
Normal 1,019 96.9 95.7 97.7
Borderline or
elevated
32 3.1 2.3 4.3
Girls
Normal 1,000 97.6 95.9 98.6
Borderline or
elevated
28 2.4E1.4 4.1
6 to 11 years
Normal 1,029 96.3 94.0 97.7
Borderline or
elevated
38 3.7E2.3 6.0
12 to 19 years
Normal 990 97.8 96.0 98.8
Borderline or
elevated
22 2.2E1.2 4.0
E interpret with caution (coeffi cient of variation 16.6% to
33.3%)
Notes: For respondents aged 6 to 17 years, blood pressure
status was derived using the methodology outlined in
Appendix B of The Fourth Report on the Diagnosis,
Evaluation, and Treatment of High Blood Pressure
in Children and Adolescents, Pediatrics 2004; for
respondents aged 18 to 19, the classifi cation in the
seventh report of the Joint National Committee on
Prevention, Detection, Evaluation and Treatment of
High Blood Pressure was used.
Source: 2007 to 2009 Canadian Health Measures Survey.
Hence, population-level increases in BP
may not necessarily be a consequence of
rising weight. More research is required
to explain this apparent paradox.
For each age and sex category, mean
child and adolescent SBP in Canada
was about 10 mmHg lower than the
most recent United States National
Health and Nutrition Examination
Survey (NHANES) data.32 The only
other recent BP data from a large,
representative sample of youth in Canada
were collected in 1999 by the Quebec
Child and Adolescent Health and Social
Survey (QCAHS) from respondents aged
9, 13 and 16 years.14 Compared with the
results of the QCAHS, mean SBP at
these ages in the CHMS was 9, 16 and 20
mmHg lower in boys, and 8, 17 and 16
mmHg lower in girls.
CHMS values for DBP generally
exceeded the NHANES results, with
a mean difference of 5 mmHg higher
in boys and 2 mmHg higher in girls.
And compared with the QCAHS, the
CHMS values were 8, 7 and 7 mmHg
higher in boys aged 9, 13 and 16 years,
respectively, and 9, 5 and 7 mmHg higher
among girls of the same ages.14
Differences in measurement
instruments and procedures may, in
part, explain the disparities in BP levels
in the three surveys. The CHMS used
Table 5
Mean measured value of systolic (SBP) (mm/Hg) and diastolic blood pressure (DBP) (mm/Hg), by age group, sex and
body mass index (BMI) category, household population aged 6 to 19 years, Canada, March 2007 to February 2009
Systolic blood pressure Diastolic blood pressure
6 to 11 years 12 to 19 years 6 to 11 years 12 to 19 years
Sample
size Mean
95%
confidence
interval Sample
size Mean
95%
confidence
interval Sample
size Mean
95%
confidence
interval Sample
size Mean
95%
confidence
interval
from to from to from to from to
Total
Neither overweight nor obese†836 92 91 93 751 98 96 99 836 60 59 61 751 63 61 64
Overweight 159 97* 93 100 180 101* 99 104 159 62 58 65 180 63 61 65
Obese 71 97* 94 101 77 106* 103 109 71 62 59 65 77 65 62 68
Boys
Neither overweight nor obese†413 92 91 93 378 99 97 100 413 60 59 61 378 62 61 63
Overweight 86 97 92 102 94 104* 100 107 86 62 57 68 94 64 61 67
Obese 38 97 91 102 40 108* 104 112 38 63 58 68 40 66* 63 70
Girls
Neither overweight nor obese†423 92 91 93 373 97 95 98 423 60 59 61 373 63 61 65
Overweight 73 97* 94 99 86 99 96 102 73 61 58 63 86 63 60 65
Obese 33 98* 95 101 37 103 98 107 33 61 58 64 37 64 60 68
† reference category
* signifi cantly different from reference category p<.025 (Bonferroni corrected)
Notes: BMI categories for ages 6 to 17 are based on the Cole cut-points. BMI categories for ages 18 and 19 years are based on the World Health Organization cut-points.
Source: 2007-2009 Canadian Health Measures Survey.
20 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Blood pressure in Canadian children and adolescents • Research article
the BpTRU™ device; QCAHS used
the DINAMAP (Critikon Co, FL)
device; and NHANES used mercury
sphygmomanometers. The last has been
the gold standard for BP assessment for
many years, but its use in children is
decreasing because mercury-containing
instruments are being removed from
pediatric environments, and because
auscultatory methods are subject
to various biases (digit preference,
rounding, white coat hypertension, etc.).
The substantial differences between
the CHMS and QCAHS may be due to
opposing systematic differences between
BP measured by mercury manometers
and by the DINAMAP and the BpTRU.
The DINAMAP has been reported to
overestimate SBP by about 10 mmHg
and slightly underestimate DBP, whereas
the BpTRU may slightly underestimate
DBP (by 2.1 mmHg), compared with
the mercury manometer.33 Most cases of
borderline or elevated BP among CHMS
participants had diastolic rather than
systolic elevation, whereas clinically,
most reported cases of pediatric
hypertension are the result of an increase
in SBP, thought to refl ect, at least in part,
hyperactivity of the sympathetic nervous
system. Counterintuitively, children
aged 6 to 11 years were somewhat more
likely to have borderline or elevated BP
than were adolescents aged 12 to 19
years.
The CHMS procedures may also
have contributed to lower mean SBP.
Measurement in a quiet room and in
the absence of staff may have been
conducive to maximal subject relaxation,
which could decrease sympathetic
activation and lower SBP. By contrast,
the QCAHS measurements took place
in school settings, usually a room where
other survey-related measures were
going on and in the presence of a staff
member recording BP readings.14,34
Limitations
The overall CHMS response rate was
slightly above 50%. Although survey
weights were adjusted to the socio-
demographic characteristics of the
Canadian population, it was not possible
to adjust for many factors that could be
associated with BP levels. Selection bias
would be present if the BP levels of non-
participants differed systematically from
those of participants. In addition, the
logistical and cost constraints associated
with the use of mobile examination
centres restricted the number of collection
sites to 15.18 Whether this sampling
strategy affected the results is unknown.
Conclusion
A small percentage of Canadians aged 6
to 19 years have borderline or elevated
BP. More research is required to improve
our understanding of BP levels and their
determinants in order to help maintain
healthy levels over the life-course. ■
Funding
Gilles Paradis holds a Canadian Institutes
of Health Research (CIHR) Applied
Research Chair in Public Health. Arnaud
Chiolero holds a Ph.D. fellowship
from CIHR. Ian Janssen holds New
Investigator Awards from the CIHR
and Ontario Ministry of Research and
Innovation.
Acknowledgements
We thank Liane Fransblow and Nathalie
Theoret for their contributions.
What is already
known on this
subject?
■Elevated blood pressure (BP) is an
important cause of disability and
death worldwide.
■Elevated BP at young ages is a
risk factor for the development of
hypertension in adulthood.
■The strength of BP tracking
increases with body mass index.
■No nationally representative BP
data for Canadian children and
adolescents have been collected
since the 1978 Canada Health
Survey.
What does this study
add?
■Based on data from the 2007-2009
Canadian Health Measures Survey,
an estimated 0.8% of Canadians
aged 6 to 19 had elevated BP, and
2.1% had borderline levels.
■The differences in mean systolic BP
between BMI categories reached
statistical significance among boys
aged 12 to 19 years, girls aged 6 to
11 years, and in both age groups
when the genders were combined.
■Differences in mean diastolic BP
by BMI category reached statistical
significance only among obese boys
aged 12 to 19 years.
21
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Blood pressure in Canadian children and adolescents • Research article
References
1. Health and Welfare, Statistics Canada. The
Health of Canadians: Report of the Canada
Health Survey (Catalogue 82-538) Ottawa:
Health and Welfare Canada/Statistics Canada,
1981.
2. Tremblay MS, Connor Gorber, S. Canadian
Health Measures Survey: brief overview.
Canadian Journal of Public Health 2007;
98: 453-6.
3. Tremblay MS, Wolfson M, Connor Gorber
S. Canadian Health Measures Survey:
Background, rationale and overview. Health
Reports (Statistics Canada, Catalogue 82-003)
2007; 18(Suppl.): 7-20.
4. Bryan S, St-Denis M, Wojtas D. Canadian
Health Measures Survey: Clinic operations
and logistics. Health Reports (Statistics
Canada, Catalogue 82-003) 2007; 18(Suppl):
53-70.
5. Day B, Langlois R, Tremblay M, et al.
Canadian Health Measures Survey: Ethical,
legal and social issues. Health Reports
(Statistics Canada, Catalogue 82-003).2007;
18(Suppl): 37-52.
6. Lawes CM, Vander Hoorn S, Rodgers A.
Global burden of blood-pressure-related
disease, 2001. Lancet 2008; 371(9623):
1513-8.
7. Daniels S, Meyer R, Loggie J. Determinants
of cardiac involvement in children and
adolescents with essential hypertension.
Circulation 1990; 82(4): 1243-8.
8. Berenson GS, Srinivasan SR, Bao W, et
al. The Bogalusa Heart Study. Association
between multiple cardiovascular risk factors
and atherosclerosis in children and young
adults. New England Journal of Medicine
1998; 338(23): 1650-6.
9. Ayer JG, Harmer JA, Nakhla S, et al.
HDL-cholesterol, blood pressure, and
asymmetric dimethylarginine are significantly
associated with arterial wall thickness in
children. Arteriosclerosis, Thrombosis, and
Vascular Biology 2009; 29(6): 943-9.
10. Cruickshank JK, Mzayek F, Liu L, et al.
Origins of the “black/white” difference
in blood pressure: roles of birth weight,
postnatal growth, early blood pressure, and
adolescent body size: the Bogalusa heart study.
Circulation 2005; 111(15): 1932-7.
11. Chen X, Wang Y. Tracking of blood pressure
from childhood to adulthood: A systematic
review and meta-regression analysis.
Circulation 2008; 117(25): 3171-80.
12. Lauer RM, Mahoney LT, Clarke WR. Tracking
of blood pressure during childhood: the
Muscatine Study. Clinical and Experimental
Hypertension 1986; 8(4-5): 515-37.
13. Burke V, Beilin LJ, Dunbar D, Kevan M.
Associations between blood pressure and
overweight defined by new standards for
body mass index in childhood. Preventive
Medicine 2004; 38(5): 558-64.
14. Paradis G, Lambert M, O’Loughlin J, et al.
Blood pressure and adiposity in children
and adolescents. Circulation 2004; 110(13):
1832-8.
15. Muntner P, He J, Cutler JA, et al. Trends
in blood pressure among children and
adolescents. Journal of the American Medical
Association 2004; 291(17): 2107-13.
16. Din-Dzietham R, Liu Y, Bielo M-V, Shamsa F.
High blood pressure trends in children and
adolescents in national surveys, 1963 to 2002.
Circulation 2007; 116(13): 1488-96.
17. Rogacheva A, Laatikainen T, Tossavainen K,
et al. Changes in cardiovascular risk factors
among adolescents from 1995 to 2004 in the
Republic of Karelia, Russia. The European
Journal of Public Health 2007; 17(3): 257.
18. Giroux S. Canadian Health Measures Survey.
Sampling strategy overview. Health Reports
(Statistics Canada, Catalogue 82-003) 2007;
18(Suppl): 31-6.
19. Health Canada. Canadian Guidelines for Body
Weight Classification in Adults (Catalogue
H49-179) Ottawa: Health Canada, 2003.
20. Cole TJ, Bellizzi MC, Flegal KM, et al.
Establishing a standard definition for
child overweight and obesity worldwide:
international survey. British Medical Journal
2000; 320(7244): 1240-3.
21. Mattu GS, Perry TL, Wright JM. Comparison
of the oscillometric blood pressure monitor
(BPM-100) with the ausculatory mercury
sphygmomanometer. Blood Pressure Monitor
2001; 6: 153-9.
22. Wright JM, Mattu GS, Perry TL, et al.
Validation of a new algorithm for the
BPM-100 electronic oscillometric office blood
pressure monitor. Blood Pressure Monitor
2001; 6: 161-5.
23. Myers MG, Valdivieso MA. Use of an
automated blood pressure recording device,
the BpTRU, to reduce the “white coat effect”
in routine practice. American Journal of
Hypertension 2003; 16: 494-7.
24. Bryan S, St-Pierre Larose M, Campbell N,
et al. Resting blood pressure and heart rate
measurement in the Canadian Health Measures
Survey, cycle 1. Health Reports (Statistics
Canada, Catalogue 82-003) 2010; 21(1): 71-8.
25. National High Blood Pressure Education
Program Working Group on High Blood
Pressure in Children and Adolescents. The
fourth report on the diagnosis, evaluation, and
treatment of high blood pressure in children
and adolescents. Pediatrics 2004; 114(2 suppl
4th report): 555-76.
26. Chobanian AV, Bakris GL, Black HR, et
al. Seventh Report of the Joint National
Committee on Prevention, Detection,
Evaluation, and Treatment of High Blood
Pressure. Hypertension 2003; 42: 1206-52.
27. Rao JNK, Wu CFJ, Yue K. Some recent work
on resampling methods for complex surveys.
Survey Methodology (Statistics Canada,
Catalogue 12-001) 1992; 18(2): 209-17.
28. Rust KF, Rao JNK. Variance estimation for
complex surveys using replication techniques.
Statistical Methods in Medical Research 1996;
5: 281-310.
29. Shields M. Overweight and obesity among
children and youth. Health Reports (Statistics
Canada, Catalogue 82-003) 2006; 17(3):
27-42.
30. Tremblay MS, Willms JS. Secular trends
in body mass index of Canadian children.
Canadian Medical Association Journal 2000;
163: 1429-33; 2001; 164(7): 970.
31. Chiolero A, Bovet P, Paradis G, Paccaud F.
Has blood pressure increased in children in
response to the obesity epidemic? Pediatrics
2007; 119(3): 544-53.
32. Ostchega Y, Carroll M, Prineas RJ, et al.
Trends of elevated blood pressure among
children and adolescents: data from the
National Health and Nutrition Examination
Survey 1 988-2006. American Journal of
Hypertension 2009; 22(1): 59-67.
33. Mattu GS, Heran BS, Wright JM. Overall
accuracy of the BpTRU―an automated
electronic blood pressure device. Blood
Pressure Monitor 2004; 9(1): 47-52.
34. Paradis G, Lambert M, O’Loughlin J, et al.
The Quebec Child and Adolescent Health
and Social Survey: design and methods of a
cardiovascular risk factor survey for youth.
Canadian Journal of Cardiology 2003; 19(5):
523-31.
22 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Blood pressure in Canadian children and adolescents • Research article
Table A
Sample sizes for selected characteristics, by sex and age group, household
population aged 6 to 19 years, Canada, March 2007 to February 2009
Age group (years)
6 to 11 12 to 19
Boys Girls Boys Girls
Blood pressure status 538 529 513 499
Measured body mass index (kg/m2) 537 529 512 496
Current smoking ... ... 507 497
Physical activity 538 528 507 497
Immediate family history of high blood pressure 530 518 481 470
Household education 524 518 498 483
Household type 538 529 513 499
Household income 524 513 457 426
Country of birth 538 529 513 499
... not applicable
Source: 2007-2009 Canadian Health Measures Survey.
Appendix
23
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Using population data to measure outcomes of care: The case of hip and knee replacements • Research article
Using population data to measure
outcomes of care: The case of hip and
knee replacements
by Claudia Sanmartin, Kimberlyn McGrail, Mike Dunbar and Eric Bohm
oint (hip and knee) replacement can provide
substantial relief to people suffering from pain and
limited mobility. In Canada, approximately 23,000
hip replacements and 38,400 knee replacements
were conducted in 2006/2007.1 The rate at which
these procedures were performed more than
doubled between 1995/1996 and 2005/2006, with
even sharper increases between 2004/2005 and
2006/2007.2 The rising rate is partially a refl ection
of an aging population; the recent acceleration is
likely related to the identifi cation of joint replacement
among the fi ve priority areas selected for meaningful
reductions in waiting times.3
J
Accumulating evidence points to the
health benefi ts of joint replacement for
osteoarthritis, including reduced pain
and greater mobility, which improve
health-related quality of life.4-9 But
despite generally positive results, some
patients do not appear to benefi t from
these procedures.10 Recent reviews by
Jones et al. indicated that 15% to 30%
of arthroplasty patients reported little or
no improvement in health-related quality
of life after surgery.11,12 However, the
generalizability of most outcome studies
is limited, as they were based on selected
samples representing specifi c geographic
regions, institutions, clinical sites, and/or
providers.
The evidence is less clear about the
effectiveness of hip replacement for
hip fracture patients. Considerable
disagreement remains about the best
course of treatment, depending on
factors such as age, type of fracture
and condition of the hip.13-15 While
surgery is almost always indicated for
such patients, the indications for type of
surgery are less clear for some subtypes
of hip fracture.15 Some studies report
higher rates of infection and mortality
after hip replacement, compared with
alternative procedures such as internal
fi xation.16-18 Other studies report lower
rates of re-operation and comparable hip
function and health-related quality of life
in the long term.18,19
Abstract
Background
Accumulating evidence points to overall
improvements in health-related quality of life
after joint replacement for osteoarthritis. Some
patients, however, do not appear to benefi t from
joint replacement. This study investigates health
outcomes of patients who underwent hip or knee
replacement surgery.
Methods
Linked survey and administrative data were used to
compare the health-related quality of life of individuals
who underwent surgery (surgical group) with that
of their contemporaries who did not (comparison
group), adjusting for other determinants of health.
Weighted multivariate linear regression analyses were
conducted.
Results
When the results were adjusted for other covariates
known to be associated with health, the surgical group
reported lower functional health (post-operative) than
did the comparison group. Differences ranged from
6% lower functional health among hip replacement
patients diagnosed with osteoarthritis to 21% lower
functional health for those with hip fractures. Among
surgical patients with osteoarthritis, co-morbid
conditions and being underweight were associated
with lower post-operative functional health.
Interpretation
This study is a unique application of linked data to
the study of health outcomes of joint replacement at
the population level. Outcomes of joint replacement
differed by the initial diagnosis or reason for the
surgery. For patients with osteoarthritis, poorer
post-operative health outcomes were associated with
co-morbidites and with being underweight.
Keywords
arthroplasty, databases, data collection, health status,
hip fractures, hospital records, osteoarthritis
Authors
Claudia Sanmartin (613-951-6059; Claudia.
Sanmartin@statcan.ca) is with the Health Analysis
Division at Statistics Canada in Ottawa, Ontario K1A
0T6; Kimberlyn McGrail is with the Health Analysis
Division and the University of British Columbia; Mike
Dunbar is with Dalhousie University; and Eric Bohm
is with the University of Manitoba Joint Replacement
Group.
24 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Using population data to measure outcomes of care: The case of hip and knee replacements • Research article
A population approach to health
outcomes yields information about
the results of care obtained in various
settings, representing a wide range
of patients, providers and health care
institutions. Most previous research in
this area has relied either on survey data,
which offer only limited information
about health care services received, or
on administrative data, which often lack
information about health outcomes and
about patient characteristics that may
explain why some fare better than others.
This study takes an innovative,
population-based approach to the analysis
of health outcomes using linked survey
and administrative data. Responses to
the 2000/2001 Canadian Community
Health Survey (CCHS) were linked with
administrative data from the Hospital
Morbidity Database (HMDB) on the use
of inpatient acute-care services. Linkage
of these two datasets makes it possible to
take advantage of the strengths of each.
The primary objective is to study
patients’ health outcomes after hip and
knee replacement: specifi cally, whether
those who have these procedures
(surgical group) return to the average
health status of their peers (comparison
group). Combining patient-based
information from the CCHS and from
the HMDB allows for an investigation
of a wide range of factors hypothesized
to be associated with outcomes of care,
as identifi ed in the Health Outcome
Framework developed by Statistics
Canada and the Canadian Institute for
Health Information.20
The second, more data-driven,
objective is to examine the potential
of linked data for the analysis of
health outcomes of specifi c surgical
interventions. This will provide some
policy perspective on gains to be made
in future data investments, for example,
surveys of patients who have undergone
surgical interventions.
Methods
Data source
The data are from the Canadian
Community Health Survey (CCHS)
and the Hospital Morbidity Database
(HMDB). The CCHS is a nationally
representative cross-sectional survey that
collects information about Canadians’
health status and use of health care.
Cycle 1.1 was conducted in 2000/2001
with a sample size of 131,535.21 The
survey covers approximately 98% of
the population aged 15 or older living in
private dwellings.
The HMDB is a national administrative
database representing all inpatient
acute hospital admissions. It contains
information on dates of admission and
separation, up to sixteen ICD-9 diagnoses
identifying the reason(s) for the stay, and
up to ten procedure codes (based on ICD-
9/-10 codes22) indicating interventions
during the stay.
Study sample
To identify the “surgical group” (those
who had joint replacement surgery),
data from cycle 1.1 of the CCHS were
linked to HMDB data covering the fi ve
years before the survey (1995/1996 to
2000/2001) using probabilistic data
linkage techniques based on health
insurance number, sex, date of birth and
postal code.23,24 The analyses included
only respondents who agreed to have
their survey information linked to
administrative data. The Statistics Canada
Policy Committee approved the linkage.
To address potential bias introduced by
non-linkers, new survey weights were
derived. The analyses excluded CCHS
respondents from Quebec, because
data provided to Statistics Canada by
Quebec for the HMDB have scrambled
health insurance numbers, which make it
impossible to link administrative records
and survey responses.
Hospital stays were included in the
analysis only if they were coded with a
fi rst surgical intervention indicating hip
or knee replacement (Table 1). Some
individuals had more than one acute
inpatient admission with the relevant
procedure codes. In these cases, the
hospital event closest to the survey date
was retained for analysis; subsequent
admissions were dropped. No attempt
was made to differentiate revisions
from primary replacements; individuals
(n=16) who stayed in hospital for these
surgeries both before and after their
CCHS interview were excluded. As well,
hospital stays that occurred within the
six months before the CCHS interview
were excluded, because in these cases,
answers to the survey questions about
heath status would refl ect the post-
operative recovery/rehabilitation period
rather than full recovery. The sample
was limited to CCHS respondents aged
40 or older because joint replacement at
younger ages is rare and generally has
different precursors and causes.
The “comparison group” consisted
of CCHS respondents aged 40 years or
older who had not had joint replacement
in the fi ve years before their interview
(n=58,667).
Analytical techniques
Univariate analyses and weighted
multivariate linear regression were
used to compare the health status of
individuals who had joint replacement
surgery (“surgical group”) with those
who did not (“comparison group”),
controlling for factors associated with
post-operative health status. The same
variables were then modelled to identify
factors associated with health status
among surgical patients diagnosed
with osteoarthritis. Small sample sizes
prevented similar analyses for the
other diagnostic groups. Analyses were
conducted with Stata software using the
Table 1
Procedure and diagnosis codes used
to identify surgery groups
Surgery
group
Procedure
code
Diagnosis
code
Hip replacement
with osteoarthritis 935, 936 715
Hip replacement
with fracture 935, 936 820, 821
Knee replacement
with osteoarthritis 934 (ICD-9) 715
Complications of
surgery (hip/knee) 934, 935, 936 996, 997,
998, 999
Other diagnoses
(hip/knee) 934, 935, 936 All other
diagnoses
25
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Using population data to measure outcomes of care: The case of hip and knee replacements • Research article
xi: regression procedure. Special linkage
weights were developed by Statistics
Canada to adjust the linked data for
those who did not consent to link and
those who could not be linked because
the information required for linkage was
insuffi cient.
Variables
Health outcome measure
The primary health outcome measure
is the health utility index (HUI), a
multidimensional preference-based
measure of health status25,26 that has been
used in studies of population health27-29
and in clinical settings,30 including
among joint replacement patients. The
HUI has a theoretical range between -0.3
(living in a state worse than death) and 1
(perfect health). It is intended to capture
an individual’s functional health status
across eight dimensions: vision, hearing,
speech, dexterity, cognition, emotion,
mobility and pain. The two latter
dimensions are particularly relevant for
individuals undergoing hip and knee
replacement surgery. A difference of
0.03 in the HUI is considered clinically
signifi cant.29
Independent variables
The CCHS includes demographic
information (age, sex, marital status,
province of residence), socio-economic
variables (household income, education),
and risk factors that are hypothesized to
be related to health status (presence of
chronic conditions, body mass index,
smoking). Education refers to the highest
level attained by the respondent: less
than secondary graduation; secondary
graduation or some postsecondary, and
postsecondary graduation. Household
income, adjusted for household size, was
measured in quintiles.
The CCHS collects information about
chronic conditions including arthritis,
diabetes, chronic obstructive pulmonary
disease, asthma, hypertension, stroke,
heart conditions, chronic pain, cancer
and depression. Individuals were
classifi ed by the number chronic
conditions they reported as diagnosed
by a health professional and lasting
more than six months. Body mass index
(BMI) was based on self-reported height
and weight (weight in kilograms/height
in metres squared). Smoking status was
categorized as never smoked, former
smoker, or current smoker based on self-
reported smoking habits.
The surgical cohort was divided into
diagnostic groups according to the reason
for joint replacement as indicated by the
most responsible diagnosis code on the
hospital separation record for the surgical
procedure: osteoarthritis, fracture (hip
replacements), complications (specifi c
ICD codes indicating complications of
a surgical intervention), or other (for
example, cancer, rheumatoid arthritis).
This classifi cation refl ects the hypothesis
that post-operative recovery differs
depending on the reason for the surgery.
Individuals undergoing joint replacement
due to fractures, for example, experience
a different trajectory of care and
outcomes, given that the surgery is in
response to an acute event.31
Results
Descriptive
A total of 598 individuals had a hip or
knee replacement sometime between
six months and fi ve years before their
CCHS cycle 1.1 interview (Table 2).
Osteoarthritis was the most common
diagnosis among both hip and knee
replacement patients: 29.5% and 40.0%,
respectively. Hip fractures accounted
for 8.7% of the cohort. Almost equal
percentages had a joint replaced with or
resulting from complications (10.5%), or
with other diagnoses such as cancer or
rheumatoid arthritis (11.2%).
The surgical group was, on average,
older than the comparison group (47.3%
versus 10.3% were aged 75 or older) and
more likely to be female (63.4% versus
51.6%) and to have co-morbidities
(89.7% versus 52.4%) (Table 3).
Average (unadjusted) health status,
measured by the HUI, was 0.615 for
Table 2
Distribution of surgery groups, by
surgical procedure and diagnosis,
respondents aged 40 or older to
2000/2001 Canadian Community
Health Survey, Canada excluding
Quebec
Surgical procedure
and diagnosis Number %
Total 598 100.0
Hip replacement
Osteoarthritis 177 29.5
Fracture 52 8.7
Knee replacement (osteoarthritis) 239 40.0
Hip or knee replacement
with/resulting from complications 63 10.5
Hip or knee replacement
with other diagnoses
(for example, cancer, arthritis) 67 11.2
Sources: 2000/2001 Canadian Community Health Survey;
Hospital Morbidity Database.
What is already
known on this
subject?
■The rate at which hip and knee
replacements are performed has
increased sharply since 1995/1996.
■Despite generally positive results,
some patients report little or no
improvement in health-related quality
of life after joint replacement.
What does this study
add?
■This study is the first population-
based analysis of the health
outcomes of joint replacement using
linked survey and administrative data
at the national level in Canada.
■People aged 40 to 79 who underwent
joint replacement reported lower
post-operative functional health than
did the comparison group.
■Among surgical patients with
osteoarthritis, co-morbid conditions
and being underweight were
associated with lower post-operative
functional health.
■Linked survey and administrative
data show promise for assessing
outcomes of health care
interventions.
26 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Using population data to measure outcomes of care: The case of hip and knee replacements • Research article
the surgical group and 0.844 for the
comparison group (Table 3). The
differences were mostly in the dimensions
of mobility and pain, as shown, for
example, on a radar plot for those age 65
to 74 (Figure 1). The pattern was similar
for the other age groups and when the
fracture group was removed from the
analysis (data not shown).
Multivariate regression analysis
Overall, the surgical group reported
lower functional health than did the
comparison group, when the results
were adjusted for other covariates
hypothesized to be associated with
health (Table 4). The results, however,
varied by diagnosis. Joint replacement
patients with a primary diagnosis of
osteoarthritis “regained” more health,
reporting 6% (hip replacement) and
9% (knee replacement) less functional
health compared with the control group,
whereas the hip facture group reported
21% less functional health.
Among joint replacement patients
with osteoarthritis, several other factors
were signifi cantly associated with post-
operative health status (Table 5). Their
functional health decreased with each
Table 3
Selected characteristics of surgery and comparison groups, respondents aged 40 or
older to 2000/2001 Canadian Community Health Survey, Canada excluding Quebec
Characteristic
Surgery group
Comparison
group
Number % Number %
Total 598 100.0 57,493 100.0
Demographic
Age group
40 to 64 116 19.4 42,881 74.6
65 to 74 199 33.3 8,699 15.1
75 or older 283 47.3 5,912 10.3
Sex
Men 219 36.6 27,820 48.4
Women 379 63.4 29,673 51.6
Marital status
Married/Common-law 377 63.0 42,448 73.8
Widowed 163 27.2 5,221 9.1
Separated/Divorced 28 4.6 5,608 9.8
Never married 31 5.2 4,155 7.2
Region
Newfoundland and Labrador, Prince Edward Island,
Nova Scotia, New Brunswick
60 10.0 6,104 10.6
Ontario 307 51.3 29,063 50.6
Manitoba, Saskatchewan 55 9.2 5,127 8.9
Alberta, British Columbia 176 29.5 17,200 29.9
Socio-economic
Education
Less than secondary graduation 280 46.9 14,717 25.6
Secondary graduation/Some postsecondary 121 20.3 15,351 26.7
Postsecondary graduation 190 31.7 26,791 46.6
Household income quintile
Lowest 20 3.4 1,797 3.1
Lower-middle 52 8.7 3,577 6.2
Middle 190 31.8 10,933 19.0
Upper-middle 180 30.2 17,659 30.7
Highest 78 13.1 17,118 29.8
Health/Lifestyle
Number of chronic conditions
None 62 10.4 27,369 47.6
One 155 26.0 15,837 27.5
Two 172 28.8 8,831 15.4
Three 124 20.8 3,794 6.6
Four 50 8.4 1,319 2.3
Five 31 5.2 276 0.5
Six or more F F 53 0.1
Body mass index (BMI)
Underweight 33 5.6 3,268 5.7
Normal 188 31.4 22,305 38.8
Overweight 219 36.7 20,920 36.4
Obese 145 24.2 9,708 16.9
Smoking
Never 236 39.4 18,420 32.0
Former 309 51.7 26,100 45.4
Current 52 8.7 12,828 22.3
Mean Health Utility Index 0.615 … 0.844 …
... not applicable
F too unreliable to be published (coeffi cient of variation 16.6% to 33.3%)
Source: 2000/2001 Canadian Community Health Survey; Hospital Morbidity Database.
Table 4
Adjusted† difference in Health
Utility Index between surgical and
comparison groups, by surgical
procedure and diagnosis, respondents
aged 40 or older to 2000/2001
Canadian Community Health Survey,
Canada excluding Quebec
Surgical
procedure
(diagnosis) Coeffi cient
95%
confidence
interval
from to
Hip (osteoarthritis) -0.056* -0.086 -0.025
Hip (fracture) -0.209* -0.265 -0.153
Knee (osteoarthritis) -0.089* -0.115 -0.063
Hip or knee
(complications) -0.075* -0.126 -0.024
Hip or knee (other) -0.164* -0.214 -0.115
No surgery ... ... ...
† adjusted for demographic, socio-economic and health/life-
style characteristics
* signifi cantly different from “no surgery” (p<0.01)
... not applicable
Sources: 2000/2001 Canadian Community Health Survey;
Hospital Morbidity Database.
27
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Using population data to measure outcomes of care: The case of hip and knee replacements • Research article
consequences of hip fractures may
adversely affect patients’ health-related
quality of life. It is likely that the fracture
itself has a negative impact on their health
trajectory; for example, the hospital stay
itself can result in changes in functional
status.34-36 Fractures among the elderly
are as much a cause as a consequence of
frailty, representing a closer to terminal
event in the process of health decline.37,38
The linked database made it possible to
explore a range of factors associated with
health outcomes of joint replacement
among a nationally representative
population. The results indicate that,
among people with osteoarthritis who
underwent joint replacement, being
underweight and having co-morbid
conditions were associated with poorer
post-operative health. Although sex,
age and marital status also seemed to be
associated with poorer health, the results
did not attain statistical signifi cance,
likely because of the small sample
size. These results are consistent with
other fi ndings that point to a variety of
factors associated with outcomes of joint
replacement,39,40 including co-morbid
conditions41 and lack of social support.42
These associations may indicate
the expected effectiveness of joint
replacement, in terms of health status, for
individuals with osteoarthritis.
The better health of former smokers,
compared with those who never smoked,
was unanticipated. However, former
smokers include both recent and long-
time quitters, the latter of whom often
achieve health status and adopt health
care practices similar to those of non-
smokers.43,44 In fact, some evidence
suggests that long-time quitters are more
likely than non-smokers to believe in the
effi cacy of modifying other risk factors.45
It is possible, then, that former smokers
(at least, long-time quitters) have adopted
other healthy lifestyles, such as greater
physical activity, that improve their
overall health.
Limitations
This study has several limitations. First,
the sample size is small—the analysis
pertains only to joint replacement patients
Figure 1
Mean (unadjusted) Health Utility Index scores, by attribute, for surgery and
comparison groups, aged 65 to 74, to 2000/2001 Canadian Community Health
Survey, Canada excluding Quebec
Source: 2000/2001 Canadian Community Health Survey; Hospital Morbidity Database.
0.00
0.25
0.50
0.75
1.00
Mobility
Pain
Vision
Hearing
Speech
Dexterity
Emotion
Cognition
Surgery group
Comparison group
On average, individuals who
underwent joint replacement surgery
were not restored to a level of functional
level compared with a similar population
group. As expected, the results varied
by type of diagnosis, from 6% (hip
replacement) and 9% (knee replacement)
lower functional health among those with
a diagnosis of osteoarthritis to 21% lower
functional health among the hip fracture
group. After surgery, patients with
fractures do not “regain” health to the
same degree as the osteoarthritis group.
This fi nding supports evidence about the
outcomes of treatment for hip fractures.
Hip fracture has been associated with
excess mortality, compared with the
general population32 and compared with
other hip replacement recipients.33 As
previously observed, the evidence about
the effectiveness of joint replacement
for hip fracture patients is mixed. Other
additional chronic condition (13% less).
Those who were underweight reported
24% less functional health, compared
with “normal” weight individuals.
Former smokers reported more functional
health (7%), compared with those who
never smoked.
Discussion
This is the fi rst population-based
analysis of health outcomes of joint
replacement using linked survey and
administrative data at the national
level in Canada. Unlike studies based
solely on administrative health data, the
availability of health-related quality of
life information (HUI) in the survey data
allowed a more direct assessment of
health outcomes on a range of patients, in
a variety of care settings and providers.
28 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Using population data to measure outcomes of care: The case of hip and knee replacements • Research article
Table 5
Linear regression coeffi cients relating selected characteristics to Health Utility
Index, joint replacement patients with osteoarthritis, respondents aged 40 or older
to 2000/2001 Canadian Community Health Survey, Canada excluding Quebec
Characteristic Coeffi cient
95% confidence
interval
from to
Demographic
Age group
40 to 64†... ... ...
65 to 74 0.064 -0.016 0.143
75 or older -0.078 -0.163 0.007
Sex
Men†... ... ...
Women -0.055 -0.117 0.007
Marital status
Married/Common-law†... ... ...
Widowed 0.055 -0.015 0.125
Separated/Divorced 0.074 -0.069 0.217
Never married -0.136 -0.282 0.010
Region
Newfoundland and Labrador, Prince Edward Island,
Nova Scotia, New Brunswick 0.094 -0.007 0.195
Ontario -0.007 -0.071 0.058
Manitoba, Saskatchewan 0.041 -0.068 0.150
Alberta, British Columbia†... ... ...
Socio-economic
Education
Less than secondary graduation 0.000 -0.070 0.070
Secondary graduation/Some postsecondary 0.034 -0.054 0.122
Postsecondary graduation†... ... ...
Household income
Lowest -0.105 -0.325 0.115
Lower-middle -0.129 -0.260 0.002
Middle -0.042 -0.137 0.053
Upper-middle -0.072 -0.165 0.021
Highest†... ... ...
Health/Lifestyle
Number of chronic conditions -0.134* -0.157 -0.112
Body Mass Index (BMI)
Underweight -0.243* -0.429 -0.057
Normal†... ... ...
Overweight 0.047 -0.028 0.122
Obese 0.024 -0.056 0.105
Smoking
Never†... ... ...
Former 0.074* 0.015 0.133
Current 0.011 -0.119 0.141
† reference category
* signifi cantly different from “no surgery” (p<0.05)
... not applicable
Sources: 2000/2001 Canadian Community Health Survey; Hospital Morbidity Database.
who were respondents to the 2000/2001
CCHS. Subsequent studies may benefi t
from ongoing efforts at Statistics Canada
to link several waves of the CCHS
to hospital administrative data. This
limitation, however, is counterbalanced
by gains in generalizability—the data
represent the Canadian population, not a
single hospital or a single health insurance
provider or even a single province.
Second, because the sample is
restricted to the household population,
it does not represent outcomes of
joint replacement among residents
of institutions such as long-term care
facilities.
Finally, the study does not directly
measure the change in health status before
and after surgery. Rather, it compares the
post-operative health status of surgical
patients to a population comparison
group. This approach assumes that the
surgery was intended to restore patients
to a level of health similar to that of their
contemporaries. However, a negative
fi nding does not necessarily signal the
absence of a gain in health-related quality
of life as a result of the surgery.
Conclusion
This study is a unique application of
linked data to the study of health outcomes
after a health care intervention, namely,
joint replacement. The data allow for a
population approach to the assessment
of health outcomes, taking into account
a range of factors. The outcomes of joint
replacement differ depending on the
initial diagnosis or reason for the surgery.
In particular, patients with osteoarthritis
who are underweight or have co-morbid
conditions may be susceptible to poorer
outcomes. Linked data show promise
for studying outcomes of health care
interventions, especially interventions
that are common and are well-
documented in administrative records.■
29
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Using population data to measure outcomes of care: The case of hip and knee replacements • Research article
References
1. Canadian Institute for Health Information.
Surgical Volume Trends, 2008 - Within and
Beyond Wait Time Priority Areas. Ottawa:
Canadian Institute for Health Information,
2008.
2. Canadian Institute for Health Information.
Hip and Knee Replacements in Canada 2007,
Annual Report. Ottawa: Canadian Institute
for Health Information, 2008.
3. Canadian Intergovernmental Conference
Secretariat. A 10-Year Plan to Strengthen
Health Care. Available at: <http://www.scics.
gc.ca/cinfo04/800042005_e.pdf>. Accessed
June 6, 2008.
4. Hawker G, Wright J, Coyte P, et al.
Health-related quality of life after knee
replacement. Journal of Bone and Joint
Surgery 1998; 80(2): 163-73.
5. Jones CA, Voaklander DC, Johnston
DW, Suarez-Almazor ME. Health related
quality of life outcomes after total hip and
knee arthroplasties in a community based
population. Journal of Rheumatology 2000;
27(7): 1745-52.
6. Laupacis A, Bourne R, Rorabeck C, et al.
The effect of elective total hip replacement on
health-related quality of life. Journal of Bone
and Joint Surgery 1993; 75(11): 1619-26.
7. Fitzgerald JD, Orav EJ, Lee TH, et al. Patient
quality of life during the 12 months following
joint replacement surgery. Arthritis Care and
Research 2007; 51: 100-9.
8. Ethgen O, Bruyere O, Richy F, et al.
Health-related quality of life in total hip and
total knee arthroplasty. A qualitative and
systematic review of the literature. Journal
of Bone and Joint Surgery 2004; 86-A(5):
963-74.
9. Montin L, Leino-Kilpi H, Suominen T,
Lepisto J. A systematic review of empirical
studies between 1966 and 2005 of patient
outcomes of total hip arthroplasty and related
factors. Journal of Clinical Nursing 2008;
17(1): 40-5.
10. Wylde V, Dieppe P, Hewlett S, Learmonth ID.
Total knee replacement: Is it really an effective
procedure for all? Knee 2007; 14(6): 417-23.
11. Jones CA, Beaupre LA, Johnston
DWC, Suarez-Almazor ME. Total joint
arthroplasties : Current concepts of patient
outcomes after surgery. Clinics in Geriatric
Medicine 2005; 21: 527-41.
12. Jones CA, Beaupre LA, Johnston DW,
Suarez-Almazor ME. Total joint arthroplasties:
current concepts of patient outcomes after
surgery. Rheumatic Diseases Clinics of North
America 2007; 33(1): 71-86.
13. Bhandari M, Devereaux PJ, Tornetta P et
al. Operative management of displaced
femoral neck fractures in elderly patients.
An international study. Journal of Bone and
Joint Surgery 2005; 87A (9): 2122-30.
14. Healy WL, Iorio R. Total hip arthroplasty.
Optimal treatment for displaced femoral
neck fractures in elderly patients. Clinical
Orthopaedics and Related Research 2004;
429: 43-8.
15. Lyons AR. Clinical outcomes and treatment
of hip fractures. The American Journal of
Medicine, 1997; 103: 51S-64S.
16. Hsiu Su BS, Aharonoff GB, Hiebert R,
et al. In-hospital mortlaity after femoral
neck fracture: Do internal fixation and
hemiarthroplasty differ? The American
Journal of Orthopedics 2003; 32(3): 151-5.
17. Sikand M, Wenn R, Moran CG. Mortality
following surgery for undisplaced intracapsular
hip fractures. Injury, International Journal of
Care for the Injured 2004; 35: 1015-9.
18. Bhandari M, Devereaux PJ, Swiontkowski
MF, et al. Internal fixation compared with
athroplasty for displaced fractures of the
femoral neck: A meta-analysis. Journal
of Bone and Joint Surgery 2003; 85A(8):
1673-81.
19. Blomfeldt R, Tornkvist H, Ponzer S et al.
Comparison of internal fixation with total
hip replacement for displaced femoral neck
fractures. The Journal of Bone and Joint
Surgery 2005; 87-A(8): 1680-8.
20. Statistics Canada and Canadian Institute for
Health Information. A Framework for Health
Outcome Analysis: Diabetes and Depression
Case Studies. Ottawa: Canadian Institute for
Health Information, 2008.
21. Beland Y. Canadian Community Health
Survey—methodological overview. Health
Reports (Statistics Canada, Catalogue 82-003)
2002; 13(3): 9-14.
22. Statistics Canada. Health Person Oriented
Information – External Linkage Production
Report (Data Years F1991-F2005). Ottawa:
Health Statistics Division, 2007 (internal
document)
23. Nadeau C. Linking HPOI 1992-2005 to CCHS.
Ottawa: Statistics Canada, Household Survey
Methods Division (internal document).
24. Statistics Canada. Generalized Record Linkage
Software. Available at: http://gensys-sysgen/
DesktopDefault.aspx?lang=en&tabid=106.
25. Feeney D, Furlong W, Torrence GW, et al.
Multiattribute and single-attribute utility
functions for the health utilities index mark
3 system. Medical Care 2002; 40(2): 113-28.
26. Feeny D, Furlong W, Boyle M, et al.
Multi-attribute health status classification
systems. Health Utilities Index.
Pharmacoeconomics 1995; 7(6): 490-502.
27. Erickson P, Kendall EA, Anderson JP, et al.
Using composite health status measures to
asses the nation’s health. Medical Care 1989;
27(suppl 3): S66-76.
28. Luo N, Johnson JA, Shaw JW, et al.
Self-reported health status of the general adult
U.S. population as assessed by the EQ-5D and
Health Utilities Index. Medical Care 2005;
43(11): 1078-86.
29. Kopec JA, Williams JI, To T, et al.
Cross-Cultural comparisons of health status
in Canada using the Health Utilities Index,
ethnicity and health. Ethnicity and Health
2001; 6(1): 41-50.
30. Grootendorst P, Feeny D, Furlong W. Health
Utilities Index Mark 3: evidence of construct
validity for stroke and arthritis in a population
health survey. Medical Care 2000; 38(3):
290-9.
31. Su H, Aharanoff GB, Hiebert R, et
al. In-hospital mortality after femoral
neck fracture: do internal fixation and
hemiarthroplasty differ? American Journal
of Orthopedics 2003; 32 (3): 151-5.
32. Richmond J, Aharonoff GB, Zuckerman JD,
et al. Mortality risk after hip fracture. Journal
of Othopaedic Trauma, 2003; 17 (1): 53-6.
33. Canadian Institute for Health Information.
Hip and Knee Replacements in Canada 2005,
Annual Report. Ottawa: Canadian Institute
for Health Information, 2006.
34. Creditor MC. Hazards of hospitalization of
the elderly. Annals of Internal Medicine 1993;
118: 219-23.
35. Covinsky KE, Palmer RM, Fortinsky RH, et
al. Loss of independence in activities of daily
living in older adults hospitalized with medical
illnesses: Increased vulnerability with age.
Journal of the American Geriatrics Society
2003; 51: 451-8.
36. Friedman SM, Mendelson DA, Bingham KW,
McCann RM. Hazards of hospitalization:
Residence prior to admission predicts
outcomes. The Gerontologist 2008; 48:
537-41.
37. Cree M, Yang Q, Sclater A, et al. Continuity
of care and health decline associated with a hip
fracture. Journal of Aging and Health 2002;
14(3): 385-98.
38. Wolinsky FD, Fitzgerald JF, Stump TE.
The effect of hip fracture on mortality,
hospitalization and functional status: A
prospective study. American Journal of Public
Health 1997; 87 (3): 398-403.
30 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Using population data to measure outcomes of care: The case of hip and knee replacements • Research article
39. Young NL, Cheah D, Waddell JP, Wright JG.
Patient characteristics that affect the outcome
of total hip arthroplasty: a review. Canadian
Journal of Surgery 1998; 41:188-95.
40. Jones CA, Beaupre LA, Johnston DW,
Suarez-Almazor ME. Total joint arthroplasties:
current concepts of patient outcomes after
surgery. Rheumatic Diseases Clinics of North
America 2007; 33(1): 71-86.
41. Lübbeke A, Katz JN, Perneger TV,
Hoffmeyer P. Primary and revision hip
arthroplasty: 5-year outcomes and influence of
age and comorbidity. Journal of Rheumatology
2007; 34(2): 394-400.
42. Escobar A, Quintana JM, Bilbao A, et al. Effect
of patient characteristics on reported outcomes
after total knee replacement. Rheumatology
2007; 46(1): 112-9.
43. Wilkins K, Shields M, Rotermann M.
Smokers’ use of acute care hospital – A
prospective study. Health Reports (Statistics
Canada, Catlogue 82-003) 2009; 20(4): 1-9.
44. Arvidsson S, Arvidsson B, Fridlund B,
Bergman S. Health prediting factors in
a general population over an eight-year
period in subjects with and without chronic
musculoskeletal pain. Health and Quality of
Life 2008; 6: 98.
45. Halpern MT, Warner KE. Differences in
former smokers’ beliefs and health status
following smoking cessation. American
Journal of Preventive Medicine 1994; 10(1):
31-7.
31
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Weight gain during pregnancy • Health matters
Weight gain during pregnancy:
Adherence to Health Canada’s guidelines
by Hélène Lowell and Doris C. Miller
Abstract
The 2006 Maternity Experiences Survey
provides information about women’s weight
before, during and after pregnancy. Using these
data, this study assessed Canadian women’s
adherence to the 1999 gestational weight gain
guidelines. Women with a higher pre-pregnancy
body mass index were more likely than normal
or underweight women to gain more than
recommended. Compared with older mothers,
a higher percentage of young mothers gained
more than recommended. Women who gave
birth for the fi rst time were more likely than those
who had had more than one birth to gain more
than recommended. A lower level of education
was associated with weight gain exceeding the
recommendations. Aboriginal women were more
likely than non-Aboriginal women to gain more
than recommended. Women who gained more
than recommended while they were pregnant
retained more weight fi ve to nine months post-
partum, compared with those who gained less
than or within the amount recommended.
Keywords
birth weight, body mass index, behaviour, body
weight changes, health surveys, pregnancy
outcomes
Authors
Hélène Lowell (1-613-948-4535; helene.lowell@
hc-sc.gc.ca) and Doris C. Miller (1-613-948-4534;
doris.miller@hc-sc.gc.ca) are with the Offi ce of
Nutrition Policy and Promotion, Health Products
and Food Branch at Health Canada, Ottawa,
Ontario, K1A 0K9.
anadian women’s adherence to Health Canada’s
gestational weight gain guidelines has not been
assessed since the recommendations were released
in 1999.1 Observational studies in countries with
similar guidelines have shown that women tend to
gain more weight than recommended while they are
pregnant.2-5 The release of perinatal health data6 for
a representative sample of Canadian women provides
an opportunity to determine if women in Canada also
gain more weight than is recommended.
C
This article describes Canadian women’s
adherence to Health Canada’s 1999
gestational weight gain guidelines, based
on an analysis of data from the 2006
Maternity Experiences Survey (see The
data). These guidelines for singleton
pregnancies vary according to the
mother’s pre-pregnancy body mass index
(BMI). At the time of the survey, the
recommended weight gain ranges were:
●12.5 to 18.0 kilograms for women
with a pre-pregnancy BMI less than
20;
●11.5 to 16.0 kilograms for women
with a pre-pregnancy BMI of 20 to
27; and
●7.0 to 11.5 kilograms for women
with a pre-pregnancy BMI greater
than 27.1
The ranges were adapted from the 1990
Institute of Medicine gestational weight
gain recommendations,7 which were
under review at the time of the analysis.8
The fi ndings are reported according
to whether the gestational weight
gain was below, within or above
the recommendations, by selected
socio-demographic and maternity
characteristics of the mother: pre-
pregnancy BMI; age; parity (number
of times the woman had given birth,
including stillbirths); education;
household income; Aboriginal status;
country of birth; and region of residence.
Results are also presented for two
health outcomes―post-partum weight
retention and infant birth weight―for the
three gestational weight gain categories.
32 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Weight gain during pregnancy • Health matters
The data
The 2006 Maternity Experiences Survey collected information about the experiences during pregnancy, birth and the early post-partum months of women
aged 15 or older at the time of their baby’s singleton live birth in Canada during the three-month period before the 2006 Census. They had to be living
with their infant at the time of the survey, which, for 96.9% of the mothers, was conducted fi ve to nine months post-partum. Mothers living in institutions
or on reserves were excluded from the survey. The survey was carried out by Statistics Canada on behalf of the Public Health Agency of Canada as an
initiative of the Canadian Perinatal Surveillance System. Detailed descriptions of the survey design and methods are available in a published report.9
The complete questionnaire is available online.10
A total of 6,421 women completed the survey, representing an estimated 76,508 women who gave birth during the target period, for a response rate
of 78%. Only those who gave birth to a full-term baby (37 to 41 weeks’ gestation) were included in this analysis; this excluded 474 women. As well, 24
women whose self-reported pre-pregnancy weight, gestational weight gain and post-pregnancy weight could not be reconciled were excluded; these were
women who may have reported their pregnancy weight instead of their weight gain, or who had a large relative difference between their pre-pregnancy
and post-pregnancy weights. Women with missing values for length of gestation or pre-pregnancy BMI were also excluded. A total of 5,554 women
remained in the analysis.
To take account of the complex survey design, the bootstrap method 11,12 was used to estimate standard deviations, coeffi cients of variation and
confi dence intervals. The signifi cance level was set at p<0.05. The Bonferroni correction13 was used for multiple comparisons.
In addition to descriptive statistics, a separate logistic regression was performed to identify signifi cant associations between socio-demographic/
maternity characteristics (pre-pregnancy BMI, mother’s age, parity, education, household income, born in Canada by Aboriginal status or born outside of
Canada, and region of residence) and gaining more weight than recommended, compared with gaining within the recommendations.
The mother’s pre-pregnancy BMI was obtained from self-reported height and weight. The mothers were also asked about their gestational weight gain
and their weight at the time of interview:
●“How tall are you without shoes on?”
●“Just before your pregnancy with (your baby), how much did you weigh?”
●“How much weight did you gain during your pregnancy with (your baby)?”
●“How much do you weigh now?”
For parity, women were defi ned as either primiparous (their fi rst live birth with no previous stillbirths), or multiparous (had previous live births or
stillbirths).
The mother’s highest level of education was categorized into four levels: less secondary, secondary graduation, some postsecondary/diploma/
certifi cate, and university degree.14
The variable for household income was similar to a derived variable for income in the 2000/2001 (cycle 1.1) Canadian Community Health Survey,15
based on total household income and the number of people living in the household, collapsed into three categories: lowest/lower-middle, middle, and
upper-middle/highest.
Mothers were asked their country of birth. For those who were foreign-born, no adjustment was made for how long they had lived in Canada.
Although Aboriginal status was asked of respondents who were born in Canada, the United States and Greenland, in this analysis, Aboriginal was defi ned
as those who self-identifi ed as Aboriginal and were born in Canada.
One of the main limitations of the data is that height and weight were self-reported. However, the percentage distribution of pre-preganancy BMIs
among the weight classifi cation categories16 of women aged 18 to 50 in this analysis and that based on self-reported height and weight of non-pregnant
women aged 18 to 50 in the 2005 Canadian Community Health Survey were similar17 (5.8% versus 5.5% underweight; 60.3% versus 59.4% normal
weight; 21.0% versus 22.4% overweight; and 13.0% versus 12.7% obese).18 This similarity provides additional assurance that the fi ndings presented
here can be generalized to Canadian women of child-bearing age.
A systematic review of studies that compared directly measured with self-reported height, weight and BMI concluded that self-reported weight and
BMI were underestimated, and height was overestimated.19 Using data from the 2005 Canadian Community Survey, Shields et al20 quantifi ed the
bias associated with self-reported height, weight and BMI. Females’ average BMI was 1.2 kg/m2 less when calculated with self-reported height and
weight, compared with measured height and weight, and as weight increased so did the difference between self-reported and measured BMI. If BMI
was underestimated in the 2006 Maternity Experiences Survey, some women might actually be in a higher BMI category; that is, rather than being in
the “less than recommendations” group, they should be in the “within recommendations” group, or in the “more than recommendations” group rather
than the “within reccommendations” group. This implies that the percentages “within recommendations” and “more than recommendations” could be
underestimated for women whose pre-pregnancy BMI was 20 to 27 or more than 27.
Mothers in Nunavut, the Northwest Territories and Yukon were included in the sample, although they were interviewed nine to 14 months post-partum
rather than fi ve to nine months. As a result of the inclusion of these women, the data on average weight retention by pre-pregnancy BMI may be an
underestimation of weight retention at fi ve to nine months post-partum.
Factors that were not controlled for in the logistic regression (such as mother’s height, smoking status and alcohol use)8 may also predict gestational
weight gain.
33
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Weight gain during pregnancy • Health matters
Factors associated with
gestational weight gain
According to the 1990 Institute of
Medicine report, a woman’s pre-
pregnancy weight is a primary
determinant of how much weight she
will gain while she is pregnant.7 Indeed,
results from the Maternity Experiences
Survey show that 55% of overweight
women gained more than recommended
while they were pregnant, compared
with 41% of those who were in the
normal range and 26% of those who
were underweight (Figure 1). However,
in addition to pre-pregnancy weight,
factors such as age, parity, education
and income have also been identifi ed as
potential predictors of weight gain during
pregnancy.21,22
The percentage who gained more
weight than recommended declined at
older ages, from 56% of 15- to 19-year-
old mothers to 35% of those aged 35 to
39 (Table 1).
Women giving birth for the fi rst
time (primiparous) were more likely
than those who had previously given
birth (multiparous) to gain more than
recommended: 47% versus 37%. This
difference persisted when other variables,
including age, were taken into account.
Primiparous women’s adjusted odds of
exceeding rather than being within the
weight gain recommendations were 1.5
(95% CI of 1.3 to 1.7) times those of
multiparous women (data not shown).
The likelihood of gaining more weight
than recommended during pregnancy
was greater among women with less than
secondary education (53%), compared
with those who had some postsecondary
education (43%) or a university degree
(38%). This difference held when other
factors were taken into account. The
adjusted odds that women with less
than secondary education would exceed
rather than be within the weight gain
recommendations were 2.1 (95% CI of
1.6 to 3.0) times those of women with a
university degree (data not shown).
On the other hand, women with a
low household income were no different
from those with a high household income
in terms of gaining more weight than
recommended (43% versus 41%) during
pregnancy. However, a higher percentage
of women with a low household income
gained less than recommended, compared
with women who had a high household
income (27% versus 21%).
Women who self-identifi ed as
Aboriginal were more likely than non-
Aboriginal women to gain more than
recommended: 55% versus 44%. And
owing to post-partum weight retention,
excess weight gain during pregnancy has
the potential of further exacerbating the
current high prevalence of overweight
and obesity among Aboriginal women23
living off-reserve.
A higher percentage of women born in
Canada (44%) gained more weight than
recommended during their pregnancy,
compared with women not born in
Canada (33%).
Among the six regions, Ontario’s
percentage distribution of weight gain
during pregnancy in relation to the Health
Canada recommendations was very
close to the distribution for Canadian
women overall. And compared with
Ontario, only in the Atlantic region did a
signifi cantly higher percentage of women
gain more weight than recommended
while they were pregnant.
Gestational weight gain and
health outcomes
The weight gain guidelines refl ect
observations of healthy pregnancy
outcomes.7 Gaining insuffi cient weight
has been associated with low birth weight
(less than 2,500 grams), while gaining
too much weight has been associated
with both high birth weight (more than
4,000 grams) and post-partum weight
retention.24
According to the Maternity
Experiences Survey, women who gained
less weight than recommended when
they were pregnant were more likely to
have an infant weighing less than 2,500
grams than a normal weight full-term
infant: 44% versus 24% (Table 2). On
the other hand, the majority (58%) of
women who gained more weight than
Figure 1
Percentage of women who gained less than, within and more than Health
Canada’s gestational weight gain guidelines, by pre-pregnancy body mass
index (BMI), female household population aged 15 or older who gave birth
during three months before 2006 Census, Canada
* signifi cantly different from corresponding estimate for BMI more than 27 (p<0.05)
Source: 2006 Maternity Experiences Survey.
34
14
31
55
27* 26*
47*
25*
41*
BMI less than 20
BMI 20 to 27
BMI more than 27
Less than
recommendations
Within
recommendations
More than
recommendations
Weight gain during pregnancy
34 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Weight gain during pregnancy • Health matters
Table 1
Percentage of women who gained less than, within and more than Health
Canada’s gestational weight gain guidelines, by selected characteristics, female
household population aged 15 or older who gave birth during three months
before 2006 Census, Canada
Characteristics
Weight gain during pregnancy
Less than
recommendations
Within
recommendations
More than
recommendations
%
95%
confidence
interval
%
95%
confidence
interval
%
95%
confidence
interval
from to from to from to
Age at delivery
15 to 19 20 14 26 24 * 19 30 56 * 49 62
20 to 24 22 18 25 29 * 25 32 50 * 46 54
25 to 29 21 19 23 35 33 37 44 * 41 46
30 to 34 23 21 25 38 36 41 39 37 42
35 to 39†26 23 30 38 35 42 35 32 39
40 or older 24 17 31 40 32 49 36 28 44
Parity
Primiparous 19 * 18 21 33 * 31 35 47 * 45 50
Multiparous†25 24 27 38 36 39 37 35 39
Highest level of education
Less than secondary†21 17 25 26 22 31 53 48 58
Secondary graduation 25 22 28 30 26 33 45 41 49
Some postsecondary/diploma/
certifi cate
22 20 24 36 * 34 38 43 * 40 45
University degree 22 20 24 40 * 37 42 38 * 35 40
Household income
Low†27 23 31 30 26 34 43 39 47
Medium 23 21 24 35 * 34 37 42 40 44
High 21 * 19 23 38 * 36 40 41 39 44
Aboriginal status
Non-Aboriginal 21 19 22 36 * 34 37 44 * 42 45
Aboriginal†16 12 20 29 23 35 55 49 61
Country of birth
Canada 21 * 19 22 35 34 37 44 * 43 46
Other†29 26 32 38 35 41 33 30 36
Region
Canada 22 21 24 36 34 37 42 40 43
Atlantic 15 13 18 35 32 37 50 * 47 53
Quebec 22 20 25 39 36 42 39 36 41
Ontario†23 21 25 35 32 37 42 40 44
Prairies 22 20 25 34 31 37 44 41 47
British Columbia 24 20 27 35 31 39 41 37 45
Territories 29 25 33 34 29 38 37 33 42
† reference category
* signifi cantly different from estimate for reference category (p<0.05)
Source: 2006 Maternity Experiences Survey.
recommended gave birth to an infant
weighing 4,000 grams or more. These
fi ndings mirror results from a systematic
review by Viswanathan et al,24 who
found moderate-to-strong evidence
of an association between gestational
weight gains below the 1990 Institute
of Medicine recommendations and low
birth weight, and strong evidence to
support an association between gains
above the recommendations and high
birth weight.
Five to nine months after they had
given birth, women who gained more
weight than recommended during their
pregnancy retained more weight (an
average of 4.5 kg) than did women
who gained within or less than the
recommendations (averages of 2.0 kg
and 0.5 kg, respectively) (Table 3).
Viswanathan et al24 also found moderate
evidence supporting an association
between weight gain above the Institute
of Medicine recommendations and post-
partum weight retention three months to
three years later.
Conclusion
Information from the 2006 Maternity
Experiences Survey suggests that
relatively high percentages of women
who are young, primiparous, less
educated or Aboriginal gain more weight
than recommended while they are
pregnant. ■
35
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Weight gain during pregnancy • Health matters
Table 2
Percentage of women who gained less than, within and more than Health
Canada’s gestational weight gain guidelines, by baby’s birth weight, female
household population aged 15 or older who gave birth during three months
before 2006 Census, Canada
Baby’s birth
weight (grams)
Weight gain during pregnancy
Less than
recommendations
Within
recommendations
More than
recommendations
%
95%
confidence
interval
%
95%
confidence
interval
%
95%
confidence
interval
from to from to from to
Low (less than 2,500) 44* 33 55 F ... ... F ... ...
Normal (2,500 to less than 4,000) 24 * 22 25 37* 35 38 40* 38 41
High (4,000 or more)†12 914302734585462
† reference category
* signifi cantly different from estimate for reference category (p<0.05)
F too unreliable to be published (coeffi cient of variation more than 33.3%)
... not applicable
Source: 2006 Maternity Experiences Survey.
Table 3
Average weight retention of women who gained less than, within and more than
Health Canada’s gestational weight gain guidelines, by pre-pregnancy body
mass index, 5 to 9 months postpartum,
female household population aged 15 or
older who gave birth during three months before 2006 Census
, Canada
Pre-pregnancy
body mass index
Weight gain during pregnancy
Less than
recommendations
Within
recommendations
More than
recommendations
Mean
weight
retention
(kg)
95%
confidence
interval
Mean
weight
retention
(kg)
95%
confidence
interval
Mean
weight
retention
(kg)
95%
confidence
interval
from to from to from to
Total 0.5E0.2 0.8 2.0 1.8 2.3 4.5 4.3 4.8
Less than 20 1.8 1.3 2.3 3.0 2.6 3.4 5.8 5.1 6.5
20 to 27 1.0E0.6 1.3 2.2 1.9 2.5 5.0 4.7 5.3
More than 27 -3.7 -4.7 -2.6 F ... ... 3.2 2.6 3.9
E interpret with caution (coeffi cient of variation 16.6% to 33.3%)
F too unreliable to be published (coeffi cient of variation more than 33.3%)
... not applicable
Source: 2006 Maternity Experiences Survey.
36 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Weight gain during pregnancy • Health matters
References
1. Health Canada. Nutrition for a Healthy
Pregnancy: National Guidelines for the
Childbearing Years. Ottawa: Minister of
Public Works and Government Services
Canada, 1999.
2. Schieve LA, Cogswell ME, Scanlon KS.
Trends in pregnancy weight gain within and
outside ranges recommended by the Institute
of Medicine in a WIC population. Maternal
and Child Health Journal 1998; 2(2): 111-6.
3. Kiel DW, Dodson EA, Artal R, et al.
Gestational weight gain and pregnancy
outcomes in obese women: how much is
enough? Obstetrics and Gynecology 2007;
110(4): 752-8.
4. Lederman SA, Paxton A, Heymsfield SB, et al.
Body fat and water changes during pregnancy
in women with different body weight and
weight gain Obstetrics and Gynecology 1997;
90: (4 Pt 1): 483-8.
5. Devader SR, Neeley HL, Myles TD,
Leet TL. Evaluation of gestational weight
gain guidelines for women with normal
prepregnancy body mass index Obstetrics
and Gynecology 2007; 110(4): 745-51.
6. Public Health Agency of Canada. 2009.
What Mothers Say: The Canadian Maternity
Experiences Survey. Available at: http://www.
publichealth.gc.ca/mes. Accessed May 13,
2009.
7. Institute of Medicine. Nutrition During
Pregnancy. Part I, Weight Gain. Washington,
DC: National Academy Press, 1990.
8. Committee on the Impact of Pregnancy
Weight on Maternal and Child Health,
National Research Council. Influence of
Pregnancy Weight on Maternal and Child
Health: Workshop Report. Available at: http://
www.nap.edu/catalog/11817.html. Accessed
January 11, 2009.
9. Dzakpasu S, Kaczorowski J, Chalmers B, et al.
The Canadian maternity experiences survey:
design and methods. Journal of Obstetrics and
Gynaecology Canada 2008; 30(3): 207-16.
10. Public Health Agency of Canada. Maternity
Experiences Survey, 2006 Questionnaire.
Available at: http://www.publichealth.gc.ca/
mes. Accessed January 8, 2009.
11. Rao JNK, Wu CFJ, Yue K. Some recent
work on resampling methods for complex
surveys using replication techniques. Survey
Methodology (Statistics Canada, Catalogue
12-001) 1992; 18(2): 209-17.
12. Rust KF, Rao JNK. Variance estimation for
complex surveys using replication techniques.
Statistical Methods in Medical Research 1996;
5(3): 281-310.
13. Miller RG. Simultaneous Statistical Inference,
Second Edition. NewYork: Springer-Verlag,
1981.
14. Chalmers B, Dzakpasu S, Heaman M,
Kaczorowski J. The Canadian maternity
experiences survey: an overview of findings.
Journal of Obstetrics and Gynaecology
Canada 2008; 30(3): 217-28.
15. Statistics Canada. Canadian Community
Health Survey (CCHS) Cycle 1.1 Derived
Variable (DV) Specifications. Available
at: http://www.statcan.gc.ca/imdb-bmdi/
document/3226_D5_T9_V1-eng.pdf.
Accessed January 12, 2009.
16. Health Canada. Canadian Guidelines for
Body Weight Classification in Adults. Ottawa:
Minister of Public Works and Government
Services Canada, 2003.
17. Statistics Canada. 2005 (3.1) Canadian
Community Health Survey. Available at:
http://www.statcan.gc.ca/cgi-bin/imdb/p2SV.
pl?Function=getSurvey&SurvId=3226&Surv
Ver=0&InstaId=15282&InstaVer=3&SDDS
=3226&lang=en&db=imdb&adm=8&dis=2.
Accessed March 30, 2009.
18. Miller D. Unpublished data.
19. Connor Gorber S, Tremblay M, Moher
D, Gorber B. A comparison of direct vs.
self-report measures for assessing height,
weight and body mass index: a systematic
review. Obesity Reviews 2007; 8(4): 307-26.
20. Shields M, Connor Gorber S, Tremblay M.
Estimates of obesity based on self-report
versus direct measures. Health Reports
(Statistics Canada, Catalogue 82-003) 2008;
19(2): 61-7.
21. Olson CM, Strawderman, MS. Modifiable
behavioral factors in a biopsychosocial model
predict inadequate and excessive gestational
weight gain. Journal of the American Dietetic
Association 1999; 94(4): 616-22.
22. Brawarsky P, Stotland NE, Jackson RA,
et al. Pre-pregnancy and pregnancy-related
factors and the risk of excessive or inadequate
gestational weight gain. International Journal
of Gynaecology and Obstetrics 2005; 91(2):
125-31. Epub October 3.
23. Garriguet D. Obesity and the eating habits
of the Aboriginal population. Health Reports
(Statistics Canada, Catalogue 82-003) 2008;
19(1): 21-35.
24. Viswanathan M, Siega-Riz AM, Moos M-K,
et al. Outcomes of Maternal Weight Gain,
Evidence Report/Technology Assessment No.
168. Prepared by RTI International-University
of North Carolina Evidence-based Practice
Center under Contract No. 290-02-0016
(AHRQ Publication No. 08-E009) Rockville,
Maryland: Agency for Healthcare Research
and Quality, 2008.
37
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Manitoba human papillomavirus surveillance • Methodological insights
The Manitoba Human Papillomavirus
vaccine surveillance and evaluation system
by Erich V. Kliewer, Alain A. Demers, Marc Brisson, Alberto Severini, Robert Lotocki, Brenda Elias,
Gregory Hammond, George Wurtak and the Manitoba HPV Research Group
Abstract
Background
With the recent introduction of a human
papillomavirus (HPV) vaccine in Canada, it is
important to establish surveillance and evaluation
programs that not only track the uptake of the
vaccine, but also assess its safety and its impact
on: distribution of HPV type, cervical cancer
screening programs, the incidence of anogenital
warts, precancerous lesions and various cancers,
and sexual behaviour.
Data sources and methods
Administrative databases, registries and
questionnaire information are being linked to
identify people receiving the HPV vaccine and to
develop an evaluation system.
Interpretation
The availability of extensive linkable databases
in Manitoba allows for the development of a
comprehensive HPV vaccine surveillance and
evaluation system that can address many of the
questions related to the HPV vaccine. Aspects of
the Manitoba surveillance and evaluation system
could be implemented in other provinces that have
similar databases.
Keywords
human papillomavirus vaccine, surveillance,
evaluation, record linkage
Authors
Erich V. Kliewer (1-604 675-8000, ext. 7076;
Erich.Kliewer@cancercare.mb.ca) and Alain
A. Demers are with CancerCare Manitoba in
Winnipeg, Manitoba; Marc Brisson is with Laval
University in Quebec City, Quebec; Alberto
Severini is with the Public Health Agency of
Canada in Winnipeg, Manitoba; Robert Lotocki,
Brenda Elias and Gregory Hammond are with the
University of Manitoba, and George Wurtak is with
the International Centre for Infectious Diseases in
Winnipeg, Manitoba.
quadrivalent human papillomavirus (HPV)
vaccine was approved for sale in Canada in
July 2006 for females aged 9 to 26 years. This
vaccine protects against infection from HPV types
6, 11, 16 and 18. Types 16 and 18 are responsible
for approximately 70% of all cervical cancers,
while types 6 and 11 are responsible for over 90%
of anogenital warts.1-4 Clinical trials have shown
that the vaccine is effective in preventing anogenital
warts and precancerous cervical, vulvar and vaginal
lesions.5-9 A bivalent (types 16 and 18) HPV
vaccine is currently going through the Canadian
regulatory approval process, and other HPV vaccines
that protect against an increased number of HPV
genotypes are being evaluated.
A
Because most provinces and territories
have implemented voluntary school-based
vaccination, it is important to establish a
surveillance and evaluation program that
not only tracks uptake of the vaccine, but
also assesses its safety and its impact on
the distribution of HPV type, on cervical
cancer screening, on the incidence of
anogenital warts, precancerous lesions
and various cancers, and on sexual
behaviour.
The Canadian National Advisory
Committee on Immunization statement
on HPV vaccine noted an infrastructure
gap in Canada, and that to evaluate the
vaccine’s effectiveness and impact,
databases and registries must be
developed and linked.10 Others have
also recognized the potential of linkable
databases for evaluating vaccines.11-15
Such databases allow for evaluation
at a population level, as opposed to
the restrictive setting of clinical trials.
Through partnerships with Manitoba
Health, CancerCare Manitoba and the
Public Health Agency of Canada’s
38 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Manitoba human papillomavirus surveillance • Methodological insights
Figure 1
Human papillomavirus (HPV) vaccine surveillance and evaluation system
Rectangular boxes = inputs
Round boxes = outputs
Dashed boxes = dependent on funding
National Microbiology Laboratory, and
with access to extensive linkable data
resources, Manitoba is well-positioned
to develop such a surveillance and
evaluation system.
This paper describes specifi c aspects
of the surveillance and evaluation system
(Figure 1) that is being implemented in
Manitoba (population 1.15 million).
HPV immunization registry
The backbone of any vaccine
surveillance and evaluation program is
an immunization registry. In Manitoba,
such a registry is being developed
from information in the Manitoba
Immunization Monitoring System
(MIMS - see www.gov.mb.ca/health/
publichealth/cdc/surveillance/mims07.
pdf), the Drug Program Information
Network (DPIN), and medical claims.
Females receiving the HPV vaccine
through the school-based program are
captured in MIMS. Those obtaining
the vaccine outside the school-based
program usually require a physician’s
prescription; the DPIN database
includes most prescriptions fi lled in
the province. This database allows for
the identifi cation of those who fi lled
a prescription for the vaccine, but it is
not possible to determine if they were
actually vaccinated. However, given the
cost of the vaccine (approximately $400
for three doses), it is unlikely that those
who purchased it did not use it.
Anecdotal reports suggest that some
Manitoba physicians provide the vaccine
to their patients without a prescription.
A potential source for identifying
these patients is the medical claims
database, which includes records of all
claims submitted to Manitoba Health
by physicians for payment for services.
The tariff (billing) code 8891 has been
specifi cally assigned to the HPV vaccine,
although this code was not implemented
until late 2008. Before that, physicians
Drug Program
Information
Network Questionnaire
Immunization
Registry
Population Registry
Pap Registry
Impact Outcomes
Cancer Registry
Pap Registry
Medical claims
Typing study HPV type
Communicable
Diseases
Registry
Cancer
Precancerous
lesions
Anogenital warts
Sexual behaviour
Vaccine uptake
Screening
Medical claims
Indian Registry
System
Metis Population
Data Base
Vaccine safety
Questionnaire
Hospitalizations
Manitoba
Immunization
Monitoring System
39
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Manitoba human papillomavirus surveillance • Methodological insights
could use 8800, which is a billable
tariff code for any immunization.
Physicians could indicate in the claim’s
comment fi eld that the HPV vaccine was
administered, but not all may have done
so. Thus, the immunization registry will
miss individuals whose physician gave
them the vaccine without a prescription
and submitted a claim with tariff code
8800, but did not specify HPV.
The immunization registry contains
only non-identifying information such
as the scrambled unique personal health
identifi cation number, date of birth,
region of residence, date the prescription
was fi lled, and date the vaccine was
administered.
Aside from being essential for an
effective evaluation of the vaccine, the
registry will also be a means of contacting
vaccinated individuals if safety issues
arise or if booster doses are required.
This is more effective than relying on the
media or health professionals.10
Non-vaccinated females
The Manitoba surveillance system
allows for the follow-up and comparison
of outcomes in vaccinated and
non-vaccinated females. All residents
covered by the provincial health
insurance are included in the Manitoba
Population Registry (MPR), which
is maintained by Manitoba Health to
administer the insurance program.
Since health insurance is provided free
of charge, it covers more than 99% of
the population. By linking the HPV
immunization registry to the MPR, it is
possible to identify females who have
not been vaccinated. Loss of follow-up
can be determined for both vaccinated
and non-vaccinated females, as the MPR
contains dates of termination of coverage
through emigration or death.
Aboriginal peoples
Although HPV infection rates16-18 and
cervical cancer incidence and mortality
rates19-21 are higher in Aboriginal than
non-Aboriginal females, little is know
about the epidemiology of HPV among
Aboriginal peoples. The uptake and
impact of the vaccine may be different in
Aboriginal populations.22
First Nations
As part of a Health Disparity Research
Program at the Manitoba First Nations
Centre for Aboriginal Health Research at
the University of Manitoba, permission
has been received from the federal
Department of Indian and Northern
Affairs to link the Indian Registry
System (IRS) to the MPR. The IRS
contains information on all registered
First Nations as defi ned by the Indian
Act, including reinstated First Nations
under federal Bill C-31 legislation. With
this link it is possible to undertake studies
on vaccinated and non-vaccinated
cohorts that include Registered First
Nation status. However, approval
must fi rst be obtained from Manitoba’s
institutional review boards, which
include First Nations ethical and health
information decision-making bodies.
To date, permission has been received
to investigate HPV vaccine uptake,
comparing Registered First Nations
in Manitoba and all Manitobans, and
permissions will be sought to examine
broader aspects of the HPV vaccine
surveillance program.
Métis
The Manitoba Metis Federation (MMF)
Health and Wellness Department, in
partnership with the Manitoba Centre
for Health Policy and Manitoba Health,
produced a province-wide Metis Health
Status and Health Services Utilization
study that created a large permanent
updatable Metis Population Data
Base (MPDB). The MPDB identifi es
the Manitoba Metis and exists under
MMF ownership, control, access and
stewardship. It is, in principle, possible
to link the MPDB with the MPR to
undertake a Metis-specifi c HPV vaccine
surveillance and evaluation program.
However, an agreement outlining the
details of the program and authority from
the MMF would be required, along with
ethics and privacy approvals.
Vaccine uptake
With the development of the
immunization registry, uptake of the
vaccine in Manitoba is being tracked on a
population basis. Specifi c questions that
are being examined include:
●What are the overall and
age-specific vaccination rates?
● How has uptake changed over
time?
●What percentage of females receive
fewer than the three recommended
doses?
●Is uptake highest in areas of greatest
need (for example, with the highest
cervical cancer rates or lowest
screening rates)?
●Does uptake vary by income
quintile?
Among individuals with lower income,
cervical cancer screening rates tend to be
lower,23-26 and cervical cancer incidence
and mortality rates higher.27,28 Cost is
an important determinant of women’s
attitudes about receiving the vaccine, and
household income has been associated
with uptake.29-31 Given the high cost,
it would be expected that vaccination
rates outside the school-based program
would be lower among individuals
with low income. Such inequity in
access may well widen the difference
between low- and high-income women
in rates of anogenital warts and cervical
abnormalities.
Vaccine impact
Cervical screening program
Some vaccinated females may develop
a false sense of protection that could
result in their no longer seeking
screening.32-34 Although the vaccine
targets the oncogenic HPV types 16 and
18, it is essential that vaccinated females
continue to be screened, as only about
70% of cervical cancers are caused
by these two HPV types.1-3 And some
females may have been infected by these
two types before vaccination or infected
by other types of oncogenic HPV.
The Manitoba Cervical Cancer
Screening Program was established in
40 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Manitoba human papillomavirus surveillance • Methodological insights
January 2000, and the reporting of all
cervical cancer screening tests to the
program was mandated by law in 2001.
A registry was established that contains
demographic information for all women
aged 18 to 69, and all Pap test, colposcopy
and biopsy results. The registry also
includes results for females outside
the program’s age range. By linking
the immunization registry to the Pap
registry, it will be possible to determine
the impact of the vaccine on screening,
in particular, whether screening rates of
vaccinated and non-vaccinated women
differ.
If females receiving the vaccine are
those who would have been screened
regularly, the vaccine will have less
impact on reducing rates of cervical
cancer.35 Because a substantial number
of Manitoba women aged 18 or older
are being vaccinated, this possibility
can be investigated by using the linked
databases to examine the screening
history of vaccinated and non-vaccinated
women.
The vaccine will likely reduce the
prevalence of cytological abnormalities,
which, in turn, will lead to a decrease
of the positive predictive value of Pap
cytology.33,36 Research is needed to
evaluate the performance of cytology
and HPV testing among vaccinated and
non-vaccinated women, although because
of ethical concerns, randomized trials
may not be possible.37 As described in
the next section, a province-wide survey
of HPV type was undertaken in Manitoba,
the results of which will be included in the
Pap registry. If such surveys continue,
the accuracy of cytology versus HPV
testing in vaccinated and non-vaccinated
women can be determined by linking the
immunization registry to the Pap registry.
It would also be possible to determine if
the HPV type is changing over time in
women with lesions.
HPV type
A pilot study conducted in Winnipeg
in 2007 and a larger province-wide
study in 2008 collected HPV samples
from approximately 900 women. HPV
type is being determined by the Public
Health Agency of Canada’s National
Microbiology Laboratory and the
Cadham Provincial Laboratory of
Manitoba Health. The results of the
HPV tests are being entered into the Pap
registry. Participants also completed a
questionnaire on demographic, socio-
economic, reproductive and lifestyle
characteristics (http://www.cancercare.
mb.ca/resource/File/Epi-Cancer_
Registry/Questionnaire_For_Risk_
Factors_Associated_With_Cervical_
Cancer.pdf).
These studies will provide preliminary
estimates of the prevalence of HPV
types in Manitoba before widespread
HPV vaccination. The intention is to
repeat the survey periodically, although
the frequency will depend on funding.
These surveys will make it possible to
determine whether the vaccine alters the
infection rate and distribution of other
HPV types, particularly other oncogenic
types.38 Based on the questionnaire
information, differences in HPV type
by the personal characteristics of survey
participants will be examined.
Sexual behaviour
Concern has been expressed that
HPV vaccination may lead to an
increase in premature sexual activity
and risky sexual behaviour.33,39,40 The
questionnaire for the Manitoba HPV
typing study, which asks about sexual
behaviour, could provide information
on the sexual behaviour of vaccinated
versus non-vaccinated females.
Although the impact of the vaccine
on sexual behaviour cannot be directly
assessed using the Manitoba databases,
differences in pregnancy or birth rates
between vaccinated and non-vaccinated
women may be an indirect measure.
Because virtually all births occur in
hospital, linked immunization registry
and hospital data can be used to determine
birth rates in the two cohorts of women.
And by including information from
medical claims, pregnancy rates could
also be estimated, although this would be
less accurate than the data for births.
Differences between vaccinated and
non-vaccinated women in the incidence
of notifi able sexually transmitted
infections may also provide indirect
evidence of how the vaccine affected
sexual behaviour. This information will
be derived by linking the vaccine registry
to the Manitoba communicable diseases
registry.
If a suffi cient number of older women
are vaccinated, it will be possible to
compare these indicators of sexual
behaviour before and after vaccination.
Vaccine outcomes
Cancer
The Manitoba cancer registry was
established in the 1930s and has been
population-based since 1956. Because
cancer is a notifi able disease and multiple
sources of ascertainment are used,
completeness in the recording of cases is
considered to be very high.
In addition to causing most cervical
cancer, HPV 16 and 18 are responsible
for 80% to 90% of anal cancers. As well,
varying proportions of vulvar, vaginal,
urethral and head and neck cancers
contain oncogenic HPV types.33 Risk
for these cancers can be determined by
linking the Manitoba cancer registry
to the cohorts of vaccinated and
non-vaccinated females. However, given
the rarity of these diseases, the cohorts
must be followed for a substantial period
before enough cases have occurred to
test for differences. On the other hand,
Manitoba may be able to contribute data
to existing effi cacy trials, such as the
Nordic HPV vaccine trials, for a possible
pooled analysis.41
Precancerous cervical lesions
While vaccination should, in the long-
term, lead to a decrease in cervical cancer
caused by HPV 16 and 18, in the short-
term, a reduction in atypical squamous
cells of undetermined signifi cance and
squamous intraepithelial lesions would
be expected because of the shorter latency
between HPV infection and development
of these abnormalities.36 Because the Pap
registry includes cytological results for
all Pap tests undertaken in Manitoba and
colposcopy and histological information,
abnormality rates among the vaccinated
and non-vaccinated can be calculated.
41
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Manitoba human papillomavirus surveillance • Methodological insights
Vaccine safety
After reviewing data on events occurring
up to six years after vaccination, the
World Health Organization concluded
that the evidence for the safety of the HPV
vaccines was “reassuring.”44,45 However,
“long-term safety data are essential
for an HPV vaccine, since it will likely
target hundreds of millions of young,
healthy individuals worldwide who are
otherwise not subject to epidemiological
surveillance…”38 Furthermore, the
safety results to date are based on
carefully controlled clinical trials, the
participants in which are subjected to
strict eligibility criteria. Studies that
examine the safety of the vaccine in
real world population-based settings are
required.
Because many vaccinated females
will be in, or about to enter, their
reproductive years, it is important
to determine if the vaccine results in
reproductive toxicities or increases the
risk of adverse pregnancy outcomes.33,38
It has been suggested that the vaccine
may have a positive impact on pregnancy
outcomes by reducing the number of
women treated for precancerous cervical
lesions.33 Procedures used to treat these
lesions, such as loop electrosurgical
excision and cold knife conization, have
been associated with preterm delivery,
low birth weight, caesarean section, and
premature rupture of membranes.46
Canada, like many other countries,
has a surveillance system that tracks
adverse events following vaccination.12
A recent report47 based on the American
system found that, except for syncope
and venous thromboembolic events, the
rates of adverse events after receiving
the HPV vaccine were no greater than
those for other vaccines. However, these
results tend to be based on voluntary
notifi cation and underestimate the actual
number of events. And because no
information is available for a comparative
non-vaccinated cohort, determining
causality is diffi cult. By linking
medical and hospitalizations records to
the immunization registry, as has been
called for by Brotherton et al.,48 it will be
possible to undertake a long-term follow-
up on a population basis to determine if
the vaccinated group is at increased risk
for any medical conditions. A similar
method is being used in the Nordic
trials.41
Mathematical modeling
Mathematical models, such as those
developed by Brisson and colleagues,49,50
are currently part of the overall evidence
base used to inform decision-making
about HPV vaccination and cervical
cancer screening programs in Canada.22
Models can also be an intrinsic part of
an ongoing HPV vaccine surveillance
program, particularly the long-term
impact of the vaccine. An individual-
based dynamic model of HPV
transmission, infection and disease,
including screening and vaccination, can
be developed with data from the various
Manitoba databases and registries.
Integration of models and surveillance
will allow:
●better understanding of emerging
epidemiologic trends after
vaccination (for example, changes
in age at infection, waning
effectiveness, herd-immunity, HPV
type replacement).
●improved predictions of the
effectiveness and cost-effectiveness
of HPV vaccination and cervical
cancer screening (for example,
projections based on up-to-date
data).
●adjustment and optimization of
HPV vaccination and cervical
cancer screening strategies (for
example, reduce number of doses,
change vaccine schedule, revisit
screening paradigms).
Conclusion
Surveillance of vaccine coverage
and safety is critical for a successful
immunization program.51 Erickson
et al. have outlined the requirements
for an evaluation of an immunization
program, which include the availability
of information systems to measure
coverage, reduction of disease incidence,
complications, sequelae and mortality,
What is already
known on this
subject?
■A quadrivalent HPV vaccine was
approved for sale in Canada in July
2006.
■Most provinces and territories
have implemented school-based
vaccination programs.
■Questions remain about the vaccine’s
safety and its impact on anogenital
warts, cervical abnormalities, cervical
cancer screening, HPV type, and
sexual behaviour.
What does this study
add?
■This article explains how linkable
databases and registries available
in Manitoba and other Canadian
provinces and territories can be used
to address questions about the HPV
vaccine.
Anogenital warts
Given the short time between exposure to
HPV and the development of anogenital
warts, they are one of the fi rst indicators
of the success of a vaccination program.42
For the 1985 to 2004 period, medical
claims and hospitalization records were
linked to identify men and women with
anogenital warts for a study of incidence
and prevalence trends in Manitoba.43
The methodology developed in that
study will be employed to create an
on-going registry of cases of anogenital
warts. This registry will be employed
to document the impact of the vaccine
on the incidence and prevalence of
anogenital warts, and by linking it to the
immunization registry, to determine the
effectiveness of the vaccine in preventing
anogenital warts. Although the vaccine
has not been recommended for males,
an anogenital warts registry will make
it possible to determine if vaccinating
females reduces the incidence of
anogenital warts in males.
42 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Manitoba human papillomavirus surveillance • Methodological insights
Acknowledgements
The authors thank Judy Bartlett, Sheila
Carter and Julianne Sanguins of the
Manitoba Metis Federation for input on
the section of the paper about the Metis.
The members of the Manitoba
HPV Research Group are: James R.G.
Butler (Australian National University,
Canberra, Australia); Magdy Dawood
(Cadham Provincial Laboratory,
Winnipeg, Manitoba); Lawrence
J. Elliott (University of Manitoba,
Winnipeg, Manitoba); Marion Harrison
(CancerCare Manitoba, Winnipeg,
Manitoba); Allan Ronald (International
Centre for Infectious Diseases, Winnipeg,
Manitoba); Brenna Shearer (International
Centre for Infectious Diseases, Winnipeg,
Manitoba), and Shelley Stopera
(Manitoba Health,Winnipeg, Manitoba).
and adverse events associated with
vaccination, and to link health outcomes
databases, immunization registries and
population registries.14 The essential
role of linked databases in evaluating
the HPV vaccine’s effectiveness has
also been noted by others in Canada and
elsewhere.10-13,22,52,53 The participants in
the Canadian HPV Vaccine Research
Priorities Workshop rated the importance
of such linkages as high, but they
considered feasibility to be low.
Manitoba has a long history of record
linkage, facilitated by the inclusion of
a unique Personal Health Identifi cation
Number in most databases. The Manitoba
databases are as comprehensive as those
being used in the Phase III and IV Nordic
trials. Information arising from the
References
1. Smith JS, Lindsay L, Hoots B, et al. Human
papillomavirus type distribution in invasive
cervical cancer and high-grade cervical
lesions: a meta-analysis update. International
Journal of Cancer 2007; 121: 621-32.
2. Muñoz N, Bosch FX, de Sanjose S, et al.
Epidemiologic classification of human
papillomavirus types associated with cervical
cancer. New England Journal of Medicine
2003; 348: 518-27.
3. Franceschi S, Clifford GM. Fraction of cervical
neoplasias due to human papillomavirus 16
and 18 in vaccine trials. International Journal
of Cancer 2008; 122: 719-20.
4. Greer CE, Wheeler CM, Ladner MB, et
al. Human papillomavirus (HPV) type
distribution and serological response to
HPV type 6 virus-like particles in patients
with genital warts. Journal of Clinical
Microbiology 1995; 33: 2058-63.
5. Joura EA, Leodolter S, Hernandez-Avila M,
et al. Efficacy of a quadrivalent prophylactic
human papillomavirus (types 6, 11, 16, and
18) L1 virus-like-particle vaccine against
high-grade vulval and vaginal lesions: a
combined analysis of three randomised clinical
trials. Lancet 2007; 369: 1693-702.
6. Garland SM, Hernandez-Avila M,
Wheeler CM, et al. Quadrivalent vaccine
against human papillomavirus to prevent
anogenital diseases. New England Journal
of Medicine 2007; 356: 1928-43.
7. Future II Study Group. Quadrivalent vaccine
against human papillomavirus to prevent
high-grade cervical lesions. New England
Journal of Medicine 2007; 356: 1915-27.
8. Future II Study Group. Prophylactic efficacy of
a quadrivalent human papillomavirus (HPV)
vaccine in women with virological evidence of
HPV infection. Journal of Infectious Diseases
2007; 196: 1438-46.
9. Paavonen J, Jenkins D, Bosch FX, et al.
Efficacy of a prophylactic adjuvanted bivalent
L1 virus-like-particle vaccine against infection
with human papillomavirus types 16 and 18 in
young women: an interim analysis of a phase
III double-blind, randomised controlled trial.
Lancet 2007; 369: 2161-70.
10. National Advisory Committee on
Immunization. Statement on human
papillomavirus vaccine. An Advisory
Committee Statement (ACS). Canada
Communicable Disease Report 2007; 33: 1-32.
11. Howlett RI, Miller AB, Pasut G, et al.
Defining a strategy to evaluate cervical cancer
prevention and early detection in the era of
HPV vaccination. Preventive Medicine 2009;
48(5): 405-6.
12. Shefer A, Markowitz L, Deeks S, et al. Early
experience with human papillomavirus
vaccine introduction in the United States,
Canada and Australia. Vaccine 2008; 26
(Suppl. 10): K68-75.
13. Singh A, Wong T, Howlett RI. Human
papillomavirus vaccines: Why the time is right
to implement immunization and surveillance
programs in Canada. Canadian Journal of
Infectious Diseases and Medical Microbiology
2008; 19: 294-6.
14. Erickson LJ, De Wals P, Farand L. An
analytical framework for immunization
programs in Canada. Vaccine 2005; 23:
2470-6.
surveillance and evaluation system will
provide data on many questions related to
the vaccines’ uptake, impact and safety.
Aspects of the Manitoba surveillance and
evaluation system could be implemented
in other provinces that have similar
databases. Also, policy makers and/or
researchers who have questions about
the HPV vaccine that are not being
addressed by the current surveillance
system and evaluation program could,
with the appropriate local ethics and
privacy approvals, obtain access to the
necessary information from the registries
and databases that form the basis of the
Manitoba surveillance system. ■
15. Chen RT, Glasser JW, Rhodes PH, et al.
Vaccine Safety Datalink project: a new tool
for improving vaccine safety monitoring in the
United States. The Vaccine Safety Datalink
Team. Pediatrics 1997; 99: 765-73.
16. Young TK, McNicol P, Beauvais J. Factors
associated with human papillomavirus
infection detected by polymerase chain
reaction among urban Canadian aboriginal and
non-aboriginal women. Sexually Transmited
Diseases 1997; 24: 293-8.
17. Hamlin-Douglas LK, Coutlee F, Roger M, et
al. Prevalence and age distribution of human
papillomavirus infection in a population of
Inuit women in Nunavik, Quebec. Cancer
Epidemiology, Biomarkers and Prevention
2008; 17: 3141-9.
18. Morris SK, Nguyen CK. The human
papillomavirus vaccine in Canada. Canadian
Journal of Public Health 2008; 99: 114-6.
19. Band PR, Gallagher RP, Threlfall WJ, et al.
Rate of death from cervical cancer among
native Indian women in British Columbia.
Canadian Medical Association Journal 1992;
147: 1802-4.
20. Young TK, Kliewer E, Blanchard J, et al.
Monitoring disease burden and preventive
behavior with data linkage: cervical cancer
among aboriginal people in Manitoba, Canada.
American Journal of Public Health 2000; 90:
1466-8.
21. Marrett LD, Chaudhry M. Cancer incidence
and mortality in Ontario First Nations,
1968-1991 (Canada). Cancer Causes and
Control 2003; 14: 259-68.
43
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Manitoba human papillomavirus surveillance • Methodological insights
33. Saslow D, Castle PE, Cox JT, et al. American
Cancer Society Guideline for Human
Papillomavirus (HPV) Vaccine Use to Prevent
Cervical Cancer and Its Precursors. CA: A
Cancer Journal for Clinicians 2007; 57: 7-28.
34. Herzog TJ, Huh WK, Downs LS, et al.
Initial lessons learned in HPV vaccination.
Gynecologic Oncology 2008; 109: S4-11.
35. Franco EL, Cuzick J. Cervical cancer screening
following prophylactic human papillomavirus
vaccination. Vaccine 2008; 26(Suppl. 1):
A16-23.
36. Franco EL, Cuzick J, Hildesheim A, et al.
Chapter 20: Issues in planning cervical cancer
screening in the era of HPV vaccination.
Vaccine 2006; 24(Suppl. 3): S171-7.
37. Franco EL, Bosch FX, Cuzick J, et al. Chapter
29: Knowledge gaps and priorities for research
on prevention of HPV infection and cervical
cancer. Vaccine 2006; 24(Suppl. 3): S242-9.
38. Hildesheim A, Markowitz L, Avila MH, et al.
Chapter 27: Research needs following initial
licensure of virus-like particle HPV vaccines.
Vaccine 2006; 24(Suppl. 3): S227-32.
39. Ogilvie GS, Remple VP, Marra F, et al.
Parental intention to have daughters receive
the human papillomavirus vaccine. Canadian
Medical Association Journal 2007; 177:
1506-12.
40. Zimet GD, Liddon N, Rosenthal SL, et al.
Chapter 24: Psychosocial aspects of vaccine
acceptability. Vaccine 2006; 24(Suppl. 3):
S201-9.
41. Lehtinen M, Herrero R, Mayaud P, et al.
Chapter 28: Studies to assess the long-term
efficacy and effectiveness of HPV vaccination
in developed and developing countries.
Vaccine 2006; 24(Suppl. 3): S233-41.
42. Lehtinen M, Dillner J. Long-term follow-up
(community randomized) vaccination trials
facilitates the implementation of human
papillomavirus vaccination. HPV Today
2008; 14: 10-1.
43. Kliewer EV, Demers AA, Elliott L, et al.
Twenty-year trends in the incidence and
prevalence of diagnosed anogenital warts in
Canada. Sexually Transmitted Diseases 2009;
36: 380-6.
22. Public Health Agency of Canada. Canadian
Immunization Committee. Recommendations
on a Human Papillomavirus Immunization
Program. Ottawa: Public Health Agency of
Canada, 2008.
23. Lee J, Parsons GF, Gentleman JF. Falling
short of Pap test guidelines. Health Reports
(Statistics Canada, Catalogue 82-003) 1998;
10: 9-19.
24. Snider JA, Beauvais JE. Pap smear utilization
in Canada: estimates after adjusting the eligible
population for hysterectomy status. Chronic
Diseases in Canada 1998; 19: 19-24.
25. Maxwell CJ, Bancej CM, Snider J, et al.
Factors important in promoting cervical cancer
screening among Canadian women: findings
from the 1996-97 National Population Health
Survey (NPHS). Canadian Journal of Public
Health 2001; 92: 127-33.
26. Johnston GM, Boyd CJ, MacIsaac MA.
Community-based cultural predictors of Pap
smear screening in Nova Scotia. Canadian
Journal of Public Health 2004; 95: 95-8.
27. Mackillop WJ, Zhang-Salomons J, Boyd CJ, et
al. Associations between community income
and cancer incidence in Canada and the United
States. Cancer 2000; 89: 901-12.
28. Ng E, Wilkins R, Fung MF, et al. Cervical
cancer mortality by neighbourhood income in
urban Canada from 1971 to 1996. Canadian
Medical Association Journal 2004; 170:
1545-9.
29. Sauvageau C, Duval B, Gilca V, et al. Human
papilloma virus vaccine and cervical cancer
screening acceptability among adults in
Quebec, Canada. BMC Public Health 2007;
7: 304.
30. Ferris DG, Waller JL, Owen A, et al. Midadult
women’s attitudes about receiving the
prophylactic human papillomavirus vaccine.
Journal of Lower Genital Tract Disease 2007;
11: 166-72.
31. Jain N, Euler GL, Shefer A, et al. Human
papillomavirus (HPV) awareness and
vaccination initiation among women in
the United States, National Immunization
Survey-Adult 2007. Preventive Medicine
2009; 48(5): 426-31.
32. Haug CJ. Human papillomavirus
vaccination--reasons for caution. New England
Journal of Medicine 2008; 359: 861-2.
44. World Health Organization. Global Advisory
Committee on Vaccine Safety, 12-13 June
2007. Weekly Epidemiological Record 2007;
82: 252-9.
45. World Health Organization. Global Advisory
Committee on Vaccine Safety, 17-18
December 2008. Weekly Epidemiological
Record 2009; 84: 37-40.
46. Kyrgiou M, Koliopoulos G, Martin-Hirsch P,
et al. Obstetric outcomes after conservative
treatment for intraepithelial or early invasive
cervical lesions: systematic review and
meta-analysis. Lancet 2006; 367: 489-98.
47. Slade BA, Leidel L, Vellozzi C, et al.
Postlicensure safety surveillance for
quadrivalent human papillomavirus
recombinant vaccine. Journal of the American
Medical Association 2009; 302: 750-7.
48. Brotherton JM, Gold MS. Monitoring
vaccine safety: a critical component of every
immunisation program. Medical Journal of
Australia 2008; 189: 243-4.
49. Brisson M, Van de Velde N, De Wals P,
et al. The potential cost-effectiveness of
prophylactic human papillomavirus vaccines
in Canada. Vaccine 2007; 25: 5399-408.
50. Brisson M, Van de Velde N, De Wals P, et al.
Estimating the number needed to vaccinate to
prevent diseases and death related to human
papillomavirus infection. Canadian Medical
Association Journal 2007; 177: 464-8.
51. Wright TC, Bosch FX, Franco EL, et al.
Chapter 30: HPV vaccines and screening in
the prevention of cervical cancer; conclusions
from a 2006 workshop of international experts.
Vaccine 2006; 24(Suppl. 3): S251-61.
52. Stanley M, Villa LL. Monitoring HPV
vaccination. Vaccine 2008;26 Suppl. 1: A24-7.
53. Wheeler CM, Franceschi S. EUROGIN
2007 roadmap—conclusion. Vaccine 2008;
26(Suppl. 1): A28-31.
45
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Hyperactivity/Inattention Subscale • Methodological Insights
Evaluating the Hyperactivity/Inattention
Subscale of the National Longitudinal
Survey of Children and Youth
by Alice Charach, Elizabeth Lin and Teresa To
Abstract
Background
High scores on the National Longitudinal Survey
of Children and Youth Hyperactivity/Inattention
Subscale (NLSCY H/I Scale) have been used
to indicate severe inattention and overactivity
representing Attention Defi cit/Hyperactivity Disorder
(ADHD) symptoms. However, a threshold on
the scale has not been identifi ed for use as an
epidemiological marker for clinically signifi cant
disorder.
Data and methods
The NLSCY H/I Scale is evaluated in a subsample
of the cycle 1 NLSCY population (n=10,498),
weighted to represent 2.36 million children aged 6
to 11 in 1994/1995. Logistic regression measured
the association of scores on the scale against three
potential criteria, adjusting for age, sex and socio-
economic status: 1) current methylphenidate use,
2) diagnosed emotional disorder, and 3) functional
impairment. Sensitivity analyses identifi ed threshold
scores where false positives and false negatives
were most nearly equivalent. The preferred criterion
provides the greatest area under the Receiver
Operating Characteristic (ROC) curve and the
highest specifi city at the identifi ed threshold.
Results
Current methylphenidate use and diagnosed
emotional disorder yielded essentially identical
models, with thresholds of 14 or more and nearly
overlapping ROC curves. High scores on the
NLSCY H/I Scale are associated with current
methylphenidate use and diagnosed emotional
disorder.
Interpretation
The parent-reported NLSCY H/I Scale can be used
in population studies as a highly specifi c indicator
of clinically signifi cant ADHD symptoms.
Keywords
Attention Defi cit/Hyperactivity Disorder,
epidemiology, hyperactivity, inattention, National
Longitudinal Survey of Children and Youth
Authors
Alice Charach (1-416-813-6600;
alice.charach@sickkids.ca),Teresa To and
Elizabeth Lin are with the Faculty of Medicine,
University of Toronto, Toronto, Ontario.
he National Longitudinal Survey of Children
and Youth (NLSCY) is a federally sponsored,
national prospective study designed to measure the
well-being, health and development of Canadian
children from birth through young adulthood. The
survey began in 1994/1995, and data collection has
occurred at two-year intervals since then. As part
of the interview, the parent (usually the biological
mother) was asked to describe the child’s behaviour
using the Children’s Behaviour Scale.
T
The entire scale is composed of
several subscales, one of which, the
Hyperactivity/Inattention Subscale
(H/I Scale), is designed to identify
hyperactive, inattentive and impulsive
behaviours in children aged 4 to 11 in
large, population-based studies. The
items were taken from the Ontario
Child Health Study1 and the Montreal
Longitudinal Study.2
Several researchers have used
high scores on this scale as a proxy
for clinically signifi cant symptoms
often identifi ed with Attention Defi cit
Hyperactivity Disorder (ADHD).3-5
However, comparisons across studies
are hampered by the lack of consistency
in classifying children likely to have a
clinically signifi cant disorder such as
ADHD. Two studies where the scale
has been dichotomized to distinguish
children with signifi cant diffi culties used
thresholds of 1.5 standard deviations
above the population mean,3,5 and another
used the top 10%.4
This article evaluates the parent-
reported NLSCY H/I Scale with data
from cycle 1 (1994/1995) of the survey.
The NLSCY H/I Scale is based on the
Ontario Child Health Study Survey
Diagnostic Instrument (OCHS SDI)1
hyperactivity scale, which was validated
against Diagnostic and Statistical
Manual of Mental Disorders, Third
Edition (DSM-III) diagnosis of ADHD,
and used a combination of parent- and
teacher-reports for case identifi cation.
However, DSM criteria for ADHD have
46 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Hyperactivity/Inattention Subscale • Methodological Insights
identify may overlap. However, one
represents a narrow, and the other, a
broad, conceptualization of disorder. In
1994/1995, methylphenidate accounted
for the vast majority of stimulant
medications prescribed to children,11,12
and was likely to be a marker for ADHD,
but not for other disorders. In addition,
the NLSCY question specifi ed current
medication use.13 By contrast, a diagnosis
of emotional, mental or nervous disorder
could apply to several conditions, only
one of which might be ADHD. As well,
the diagnosis was not specifi c to the time
of the interview.
The NLSCY includes items about
functional impairment (academic
performance, getting along with peers,
and getting along with parents), the third
element identifi ed by Boyle et al. As was
done by Boyle et al., this study combined
these items to indicate impaired
functioning in at least one domain.9
Another consideration in developing
a model to evaluate the H/I Scale is the
sensitivity and specifi city of the threshold
for case identifi cation. Sensitivity is the
percentage of cases that the threshold
identifi es as positive that truly have the
disorder (true positives/true positives
+ false negatives). Specifi city is the
percentage of cases that the threshold
identifi es as negative that do not have the
disorder (true negatives/true negatives
+ false positives).14 For diagnostic
screening, the ideal threshold maximizes
both sensitivity and specifi city, although
false positives and false negatives
may be common. But when a scale is
used to determine the prevalence of a
relatively rare disorder in a large non-
clinical population, priority should go to
minimizing the error rate.15
Like other measures designed for
use in population samples, the NLSCY
H/I Scale is abbreviated. Such scales
are highly sensitive, but not very
specifi c. A threshold chosen to balance
sensitivity and specifi city would yield
excessively high rates of false positives
in population samples. The overall error
rate is lowest when the threshold is set
where the numbers of false positives and
false negatives are closest to equivalent.
This results in greater specifi city and
less sensitivity, the strategy chosen for
this study because of the relatively low
prevalence of ADHD in the population.
The goal is to develop a model for
evaluating the NLSCY H/I Scale so that
it can be used to identify children with
clinically signifi cant ADHD symptoms
in large population-based studies. The
model is tested with 1994/1995 data, the
cycle with the most complete information
and that has not been subject to attrition
over time. The objectives are to: 1)
evaluate the strength of the association
between scores on the NLSCY H/I Scale
and each of three potential criteria for
ADHD: current methylphenidate use,
diagnosis of an emotional disorder, and
functional impairment; 2) identify the
criterion with the strongest association,
adjusting for age, sex and socio-economic
status; and 3) identify the threshold with
the most nearly equal false negatives and
false positives. The point prevalence of
clinically signifi cant ADHD symptoms
among Canadian children is also
estimated.
Methods
Sample
The NLSCY used a random sampling
frame of households with clusters within
age groups and large geographic areas
to be representative of children in the 10
provinces. Children in highly mobile,
transient or homeless families were
under-represented. Children living in
institutions and on Aboriginal reserves
were excluded. A full description of the
NLSCY is available elsewhere.16
In each household, Statistics Canada
interviewers administered a standardized
questionnaire to the person most
knowledgeable about the child (the
biological mother in 89.9% of cases).
(In this study, the term “mother” or
“parent” is used rather than person most
knowledgeable because the NLSCY H/I
Scale was designed to be parent-reported.)
The overall response rate was 87%. The
population analysed is the subset of the
NLSCY sample consisting of children
aged 6 to 11 in 1994/1995 whose parent
been substantially revised since DSM-III
was published.6 Nonetheless, since DSM-
II,7 the underlying conceptualization has
remained a long-standing childhood
disorder characterized by detrimental
levels of overactivity, impulsiveness and
distractibility, and a short attention span.
In addition to changes in diagnostic
criteria over time, another reason for
evaluating how well the parent-reported
NLSCY H/I Scale identifi es children at
risk of ADHD is that 52% of the teacher-
reported information was missing in
cycle 1.8 To determine whether the child
showed symptomatic behaviours in more
than one context (at home and at school),
the OCHS SDI required that both parent
and teacher rate the child. However, the
lack of teacher responses for slightly
more than half the NLSCY participants
substantially undermines this method of
identifying cases in the NLSCY data.
The method in the present study
is based on the work of Boyle et
al.,9 who recommended that survey
instruments designed for population
studies incorporate “elements of
distress, impairment and therapeutic
concern” in defi ning a case, rather than
simply applying a threshold number
of symptoms.9 In 1999, Goodman
demonstrated that a measure of child
“impact” that combined “distress” and
“social impairment” improved case
identifi cation, compared with parent and
teacher ratings alone.10
The NLSCY database contains two
variables that represent these elements:
current use of methylphenidate (Ritalin),
which is used almost exclusively to
treat childhood ADHD,11,12 and previous
diagnosis of emotional, psychological
or nervous disorder. In each instance,
the child’s parent would have sought
professional assistance. In the fi rst,
before prescribing methylphenidate,
a physician concurred that the child
required treatment, and in the second, for
the child to have a diagnosed emotional
disorder, a health professional perceived
enough impairment to warrant diagnosis
and treatment
To some extent, the groups of
children that these two variables
47
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Hyperactivity/Inattention Subscale • Methodological Insights
responded to the interview—a total
of 10,498, representing 2.36 million
children across Canada.
The statistical models were developed
based on a training sample derived using
a replicate sampling with replacement
strategy, 17,18 followed by testing the
statistical model in the NLSCY sample.
A random half-sample for model
development was not feasible because
of confi dentiality constraints imposed by
the small number of respondents scoring
positive for the clinical indicators.
The training sample was produced
by aggregating 10 replicate random
samples, each equivalent to 10% of the
NLSCY sample, a strategy similar to a
simplifi ed bootstrap17,18 procedure. The
resulting sample is comparable to the
NLSCY sample (Table 1). The fi nal
models reported here are those evaluated
in the NLSCY sample, specifi cally the
NLSCY cycle 1 subsample of children
whose parents answered the interview
questions.
Measures
The Hyperactivity/Inattention Subscale
of the parent-reported NLSCY Children’s
Behaviour Scale consists of 8 items (can’t
sit still, distractible, fi dgets, impulsive,
diffi culty sitting still, cannot settle for
long, can’t concentrate, inattentive)
scored as 0 (not true), 1 (sometimes true)
or 2 (often true), resulting in a continuous
scale with scores from 0 to 16. Internal
consistency on factor analysis is good
(Cronbach’s α =0.86).8
The covariates are the child’s age
and sex, low maternal education (did
not complete secondary school) and low
household income, based on Statistics
Canada’s derived variable of household
size and income (below the 1995 low
income cut-off).19
Three clinical indicators reported by
the parent were evaluated as potential
criteria for validity:
Current methylphenidate use: “Does
… (your child) … take any of the
following medications on a regular basis
… Ritalin?” (Yes/No)
Previous diagnosis of emotional
disorder: “ Does … (your child) …
have any of the following long-term
conditions that have been diagnosed by
a health professional? … Emotional,
psychological or nervous disorder?”
(Yes/No)
Impairment in academic, social or
family functioning:
●“Based on your knowledge of his/
her school work, including report
cards,… how is he/she doing
overall?”
●“During the past six months, how
well has … (your child) … gotten
along with other kids such as friends
or classmates (excluding brothers or
sisters)?”
●“How well has … (your child) …
gotten along with his/her parents?”
Parents evaluated each of the three areas
on a 5-point scale; scores of 4 or more
(poor or very poor functioning) on any
of the three scales indicated functional
impairment in one or more areas.9
Data analysis
The research design is a retrospective
cross-sectional analysis. Logistic
regression analysis using backward
selection was applied to the training
sample to measure the association
of the H/I Scale against each of the
potential criterion variables, adjusted
for age, sex, low maternal education
and low household income. The
regression models included cross-
sectional population weights.8,20,21 Only
independent variables with a statistical
signifi cance of p < 0.01 were retained in
the fi nal models. Best-fi t statistical models
were chosen using the –2 Log Likelihood
statistic and the Hosmer-Lemeshow
goodness-of-fi t test. Receiver Operating
Characteristic (ROC) curves were
examined to identify the model with the
greatest area under the curve. Sensitivity
analyses throughout the full range of
scores were used to identify the threshold
scores with the most nearly equivalent
false negatives and false positives.
Frequency estimates were normalized to
adjust for missing values, and reported
with cross-sectional population weights
and variance estimates.8,20 The criterion
variable of choice was the one with the
greatest area under the ROC curve, with
the largest beta, and whose threshold has
the highest specifi city.
After the preferred criterion variable(s)
were determined, the statistical models
were tested in the NLSCY sample.
Using the threshold score identifi ed
during model development, the
association of the binary NLSCY H/I
Scale was measured against the chosen
criteria in the NLSCY sample (including
cross-sectional population weights)
to understand the properties of case
identifi cation. In addition, an estimate of
the population prevalence of clinically
signifi cant ADHD symptoms was
generated. All analyses were performed
using SAS version 8.2.22
Table 1
Selected characteristics of training sample and National Longitudinal Survey
of Children and Youth (NLSCY) sample, household population aged 6 to 11,
Canada excluding territories, 1994/1995
Characteristics
Training sample NLSCY sample
%
99%
confidence
interval
%
99%
confidence
interval
from to from to
Male sex 51.3 48.8 53.7 51.3 48.0 53.7
Low maternal education 16.7 15.1 18.3 17.1 15.4 18.7
Low household income 16.3 14.7 17.9 17.0 15.4 18.7
Current methylphenidate use 2.0 1.3 2.6 1.9 1.4 2.5
Diagnosed emotional disorder†1.9 1.3 2.4 1.7 1.1 2.3
Impairment in academic, social or family functioning 4.9 4.0 5.8 4.7 3.7 5.7
† emotional, psychological or nervous disorder
Notes: Determined using cross-sectional weights from Statistics Canada, normalized for missing values. Training sample = 10,370
observations, representing 2,354,000; NLSCY sample = 10,498 observations, representing 2,360,300.
Source: 1994/1995 National Longitudinal Survey of Children and Youth.
48 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Hyperactivity/Inattention Subscale • Methodological Insights
Table 2
Logistic regression models of current methylphenidate use, diagnosed emotional disorder and functional impairment,
household population aged 6 to 11, Canada excluding territories, 1994/1995
Characteristics
Current
methylphenidate use
Diagnosed
emotional disorder
Impairment in academic,
social or family functioning
beta†
Standard
deviation P value beta†
Standard
deviation P value beta†
Standard
deviation P value
Hyperactivity/Inattention Scale
(continuous) 0.30 0.02 < 0.0001 0.31 0.02 < 0.0001 0.29 0.01 < 0.0001
Male sex 0.52 0.10 < 0.0001 ... ... ... ... ... ...
Age (years) 0.13 0.04 0.0046 0.26 0.05 < 0.0001 0.16 0.03 < 0.0001
Low household income ... ... ... 0.32 0.09 0.0002 0.39 0.05 < 0.0001
† beta coeffi cient of parameter
... not applicable
Notes: Multivariable regression models chosen by backwards selection. Model chosen using backwards selection with p < 0.01 to stay and p < 0.10 to go. Best model chosen using -2 Log Likelihood
statistic and Hosmer/Lemeshow Goodness-of-Fit tests.
Source: 1994/1995 National Longitudinal Survey of Children and Youth.
Table 3
Sensitivity and specifi city values for threshold on National Longitudinal
Survey of Children and Youth Hyperactivity/Inattention Subscale, by
current methylphenidate use, diagnosed emotional disorder and functional
impairment, household population aged 6 to 11, Canada excluding territories,
1994/1995
Threshold
True
positives
True
negatives
False
positives
False
negatives Sensitivity Specifi city
%%
Current methylphenidate use
7 or more 33,110 1,674,677 592,445 11,738 74 74
8 or more 30,901 1,823,860 443,262 13,947 69 80
9 or more 28,100 1,961,000 306,100 16,800 63 86
10 or more 26,100 2,056,900 210,200 18,700 58 91
11 or more 20,900 2,121,000 146,100 23,900 47 94
12 or more 16,400E2,169,000 98,100 28,400 37 96
13 or more 13,700E2,201,500 65,600 31,200 31 97
14 or more* 11,200E2,230,200 36,900 33,600 25 98
15 or more 8,000E2,247,800 19,300 36,900 18 99
Diagnosed emotional disorder
7 or more 28,932 1,674,227 596,623 10,692 73 74
8 or more 27,644 1,824,331 446,519 11,980 70 80
9 or more 24,200 1,960,900 309,900 15,400 61 86
10 or more 21,800 22,056,300 214,600 17,900 55 91
11 or more 17,800 2,121,600 149,200 21,800 45 93
12 or more 14,800E2,171,100 99,700 24,900 37 96
13 or more 11,700E2,203,200 67,700 28,000 29 97
14 or more* 8,900E2,231,700 39,200 30,700 23 98
15 or more 6,400E2,250,000 20,900 33,200 16 99
Impairment in academic, social
or family functioning
9 or more 59,300 1,927,700 274,800 49,900 54 88
10 or more 48,800 2,015,000 187,500 60,400 45 91
11 or more 39,600 2,075,100 127,500 69,700 36 94
12 or more* 31,200 2,119,300 83,300 78,000 29 96
13 or more 25,307 2,148,500 54,000 83,900 23 98
14 or more 17,700 2,172,100 30,400 91,600 16 99
* threshold value with most nearly equivalent false positives and false negatives
E interpret with caution (coeffi cient of variation 16.6% to 33.3%)
Note: Normalized weighted population frequencies; N=2,312,000 for current methylphenidate use; N=3,210,500 for diagnosed
emotional disorder; N=2,311,800 for impairment in academic, social or family functioning.
Source: 1994/1995 National Longitudinal Survey of Children and Youth.
Results
Scores on the parent-reported NLSCY
H/I Scale was associated with each of
the three clinical indicators: current
methylphenidate use; previous diagnosis
of emotional disorder; and impaired
functioning in academic, social or family
domains among children aged 6 to 11
(Table 2). Statistical overlap among the
three models was substantial, with beta
values ranging narrowly from 0.29 (SE
= 0.01) for impaired functioning to 0.31
(SE = 0.02) for emotional disorder.
The sensitivity analyses and resulting
ROC curves show that methylphenidate
use and emotional disorder produced
essentially the same statistical model—
that is, a model with a greater area under
the ROC curve and higher specifi city at
the identifi ed threshold compared with
impaired functioning (Table 3, Figure
1). With either methylphenidate use or
emotional disorder as the criterion, a H/I
Scale threshold of 14 or more out of 16
distinguished cases from non-cases. With
methylphenidate use as the criterion,
sensitivity = 0.25 and specifi city = 0.98;
with emotional disorder, sensitivity
= 0.23 and specifi city = 0.98. This
threshold resulted in highly specifi c, but
not very sensitive, case identifi cation.
The logistic regression models for the
continuous scale highlight similarities
and differences between the models
(Table 2). If methylphenidate use is
the criterion, the best-fi t model for the
49
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Hyperactivity/Inattention Subscale • Methodological Insights
continuous H/I Scale includes age and
sex as modifi
ers, with older children
and boys more likely to be taking
methylphenidate. If emotional disorder
is the criterion, the best-fi t model for
the continuous H/I Scale includes age
and low household income, with older
children and those from low-income
households more likely to have been
diagnosed.
To better understand the properties of
case identifi cation using the H/I Scale,
the association of the binary variable with
methylphenidate use and with emotional
disorder was examined (data not shown).
Children using methylphenidate were
more likely to: show high rather than low
levels of parent-reported hyperactivity
and inattention (OR = 16.2; 99% CI
= 11.2 to 23.5); be boys rather than
girls (OR = 3.8; 99% CI = 2.6 to 5.4);
and be near the older end of the 6 to 11
age range (OR = 1.1; CI = 1.0 to 1.2).
Children with a previous diagnosis of
emotional, psychological or nervous
disorder were more likely to: show high
levels of parent-reported hyperactivity
and inattention (OR = 16.9; CI = 11.3 to
25.2); come from households with a low
income (OR = 2.2; CI = 1.6 to 3.1), and
be near the older end of the age range
(OR = 1. 3; CI = 1.2 to 1.4).
Based on a threshold of 14 or more
on the NLSCY H/I Scale, an estimated
2.1% (99% CI = 1.5 to 2.7) of Canadian
children aged 6 to 11 had clinically
signifi cant ADHD symptoms.
Discussion
The current study demonstrates that
the NLSCY Hyperactivity/Inattention
Subscale was associated with two
clinical indicators of ADHD in Canadian
children aged 6 to 11 in 1994/1995:
methylphenidate use, adjusted for age
and sex, and previous diagnosis of
emotional disorder, adjusted for age and
household income. Earlier studies based
on the NLSCY have shown an association
between high levels of parent-reported
hyperactivity and methylphenidate use
among school-aged boys,3,5 but this is
the fi rst to examine the association of
hyperactivity with emotional disorder,
and to develop a model to determine
a threshold for use as a marker for
identifying children with ADHD.
Although there is no clear statistical
advantage to choosing either current
methylphenidate use or previous
diagnosis of emotional disorder as the
criterion for evaluating the NLSCY H/I
Scale, there may be broad conceptual
value in choosing the latter. While it is
no surprise that methylphenidate use can
be a criterion for ADHD, it is somewhat
more novel that a history of emotional
disorder can be used as a criterion as well.
The ROC curves for methylphenidate
use and for emotional disorder appear to
be essentially the same statistical model.
This is consistent with the likelihood that
children taking methylphenidate were
diagnosed before they began taking it.
That is, the basic construct of “caseness”
is met a child came to the attention of
a health professional because of parental
concern, and that professional agreed
that therapeutic attention was warranted.
However, diagnosis of emotional
disorder represents a wide array of
potential disorders. As such, it could
be considered for use as a criterion to
evaluate other subscales of the NLSCY
Children’s Behaviour Scale as potential
measures of mental health disorders.
The similarity in the statistical
models raises the question of whether
methylphenidate use and diagnosed
emotional disorder represent the same
children. As noted earlier, the two
groups may overlap, but only partially.
For example, boys were more likely
than girls to use methylphenidate, but
sex was not signifi cantly associated with
diagnosed emotional disorder. And
while children in low-income households
were more likely to have been diagnosed
with an emotional disorder, household
income was not associated with
methylphenidate use. The lack of
overlap may be attributable to several
factors. The NLSCY question about
emotional disorder asked if the child had
ever received a diagnosis. Therefore,
Figure 1
Receiver Operating Characteristic (ROC) curves from sensitivity analyses
for National Longitudinal Survey of Children and Youth Hyperactivity/
Inattention Subscale plotted against current methylphenidate use, diagnosed
emotional disorder and functional impairment, household population aged 6 to
11, Canada excluding territories, 1994/1995
Source: 1994/1995 National Longitudinal Survey of Children and Youth.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 0.05 0.1 0.15 0.2 0.25 0.3
1 - Specificity
Sensibility
Current methylphenidate use
Diagnosed emotional disorder
Impairment in academic, social
or family functioning
Threshold
50 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Hyperactivity/Inattention Subscale • Methodological Insights
a child may have been diagnosed with
ADHD in the past, but was not taking
medication at the time of the interview.
As well, parent reports that a professional
had diagnosed the child with an
emotional disorder could refer to a wide
range of conditions, including cognitive
and learning problems identifi ed by
educators, and other behaviour disorders
for which medication is not the treatment
of choice.
The prevalence estimate of 2.1%
for ADHD among 6- to 11-year-olds
in Canada is low compared with other
estimates. The Ontario Child Health
Study reported 6.1% among children aged
4 to 16,23 and the Quebec Child Mental
Health Survey, 5.4% among children
aged 6 to 14. 24 These two estimates
were based on combined parent and
teacher information about symptoms and
a measure of impairment. In a systematic
review of studies using this combination
of case identifi cation methods, Waddell
et al. generated a summary prevalence
estimate of childhood ADHD of 4.8%
(95% CI: 2.7 to 7.3).25
The low estimate from the NLSCY
H/I Scale may refl ect the use of only
parent information. However, it may also
refl ect the input of clinical professionals.
In 2000, British researchers, Goodman et
al., estimated that 2.4 % of children aged
5 to 15 had ADHD according to DSM-IV
criteria.26 Their method for identifying
cases included parent and teacher reports
and measures of impairment, but in
addition, a clinician reviewed all material
to decide if the child met diagnostic
criteria. Health professionals examine
children with behavioural problems to
fi nd explanations other than ADHD, a
judgment not available from surveys,
and one that could infl uence rates of case
identifi cation.27 Therefore, an alternative
explanation for the low NLSCY estimate
is that it may refl ect the practice of
Canadian health professionals in
1994/1995. For instance, in 1995/1996,
administrative data from Manitoba
identifi ed 2.9% of children aged 7 to
9 and 2.2% of children aged 10 to 13
with ADHD, a rate similar to that of the
NLSCY.27
Limitations
An important question raised by this
study is whether parent-reported clinical
case markers are the “gold standard”
for identifi cation of childhood ADHD.
A strong argument can be made that
parent reports introduce multiple sources
of potential error. Waddell et al.25
recommended independent professional
input for population-based studies. The
original design of the NLSCY would
have offered the opportunity to use
this method, but missing teacher data
preclude this option.
The issue is how best to take advantage
of the strengths of NLSCY data to
examine predictors and consequences
of severe childhood hyperactivity and
inattention. The clinical markers are
useful target criteria for severe behaviour
problems. Specifi cally, parent-reported
history of emotional disorder can be
used to set thresholds for the children’s
behaviour questionnaire subscales in the
NLSCY.
The replicate sampling strategy with
replacement rather than a random half
sample to create the development set
could be a limitation, because individual
participants might appear in the dataset
more than once. Although this may
seem to interfere with the independence
of observations (and potentially bias
derived estimates), the strategy is a
simplifi ed version of the bootstrapping
procedure used to provide reliable
variance estimates and confi dence
intervals around values derived in
population samples.17,18 In fact, estimates
from the development sample were
highly comparable to estimates in the
NLSCY parent sample (Table 1) .
Some researchers have suggested
that NLSCY data can be used without
population weights for studies where
population estimates are not the primary
focus. While the methodological gap
addressed here may appear to be such
a study, it is important to consider the
likelihood of geographic variability.
Differences in rates of ADHD diagnosis
and psychostimulant prescriptions in
adminstrative data suggest that Canadian
children experience differential access
to specialists and differences in clinical
practice.27,28 For the current study, the
prevalence of clinical markers is too
low to generate provincial estimates.
Statistics Canada’s cross-sectional
population weighting strategies were
used to resolve the population distribution
issues and provide a national estimate.
An additional limitation is that the
scale was developed to elicit information
about children aged 4 to 11, but previous
diagnosis of emotional disorder and
What is already
known on this
subject?
■The parent-reported Hyperactivity/
Inattention Subscale in the National
Longitudinal Survey of Children
and Youth was designed to identify
children with severe symptoms of
hyperactivity and inattention.
■A threshold score on the scale that
identifies children likely to have
clinically significant disorder has not
been determined.
■Previous studies using the Scale
have not been consistent in
how children with high levels of
hyperactivity were defined.
What does this study
add?
■Variables collected by the NLSCY
that represent therapeutic concern by
parents and health professionals—
methylphenidate use and diagnosis
of emotional disorder—can be
used as criteria to evaluate the
Hyperactivity/Inattention Scale and
set a threshold that identifies clinical
“cases” requiring intervention.
■The threshold where false positives
and false negatives are nearly
equivalent is a highly specific, but
not very sensitive, marker of clinical
“caseness.”
51
Statistics Canada, Catalogue no. 82-003-XPE • Health Reports, Vol. 21, no. 2, June 2010
Hyperactivity/Inattention Subscale • Methodological Insights
impairment in school functioning were
asked only for children aged 6 to 11.
Also, the sample size was not large
enough to perform separate sensitivity
analyses by sex and age, which are
both known modifi ers of hyperactivity
and attention span. It is plausible
that separate threshold values should
be set by sex or by age. Finally, the
small number of children reported as
using methylphenidate or having been
diagnosed with an emotional disorder
introduces uncertainty. However, a
conservative approach was used in
the regression analyses, retaining only
variables with statistical signifi cance
<0.01 and thereby increasing confi dence
in the results.
disorder can be used as a target criterion
to evaluate other subscales of the NLSCY
children’s behaviour questionnaire.
With a common method of using the
behaviour subscales as clinical markers,
NLSCY data can be applied to the study
of childhood mental health disorders. ■
Acknowledgements
Access to the National Longitudinal
Survey of Children and Youth was made
possible through the Toronto Region
Research Data Center of Statistics
Canada at the University of Toronto.
Future directions
With the method described in this paper,
the parent-reported NLSCY H/I Scale can
be used to identify clinically signifi cant
ADHD symptoms in Canadian children
aged 6 to 11, either as an outcome
measure for investigating developmental
antecedents of such symptoms, or as
an independent variable predicting
adolescent and adult outcomes in the
NLSCY sample. Even without teacher
information, a score of 14 or more on the
scale identifi es children likely to have
clinically signifi cant diffi culties. As
well, previous diagnosis of emotional
52 Health Reports, Vol. 21, no. 2, June 2010 • Statistics Canada, Catalogue no. 82-003-XPE
Hyperactivity/Inattention Subscale • Methodological Insights
References
1. Boyle MH, Offord DR, Hofmann HG, et al.
Ontario Child Health Study. I. Methodology.
Archives of General Psychiatry 1987; 44(9):
826-31.
2. Tremblay RE, Loeber R, Gagnon C, et al.
Disruptive boys with stable and unstable
high fighting behavior patterns during junior
elementary school. Journal of Abnormal Child
Psychology 1991;19(3): 285-300.
3. Charach A, Cao H, Schachar R, To T.
Correlates of methylphenidate use in Canadian
children: a cross-sectional study. Can J
Psychiatry 2006;51(1):17-26.
4. Elgar FJ, Curtis LJ, McGrath PJ, et al.
Antecedent-consequence conditions in
maternal mood and child adjustment: a
four-year cross-lagged study. Journal of
Clinical Child and Adolescent Psychology
2003; 32(3): 362-74.
5. Miller AR, Kohen D, Johnston C. Child
characteristics and receipt of stimulant
medications: a population-based study.
Ambulatory Pediatrics 2008; 8(3): 175-81.
6. American Psychatric Association. Diagnostic
and Statistical Manual of Mental Disorders.
Fourth ed. Washington, DC: American
Psychiatric Association, 1994.
7. American Psychatric Association. Diagnostic
and Statistical Manual of Mental Disorders.
Second ed. Washington, DC: American
Psychiatric Association, 1968.
8. Statistics Canada and Human Resources
Development Canada. National Longitudinal
Survey of Children: User’s Handbook and
Microdata Guide Cycle 1. Ottawa: Minister
of Industry, 1997.
9. Boyle MH, Offord DR, Racine Y, et al.
Identifying thresholds for classifying
childhood psychiatric disorder: issues and
prospects. Journal of the American Academy
of Child and Adolescent Psychiatry 1996;
35(11): 1440-8.
10. Goodman R. The extended version of the
Strengths and Difficulties Questionnaire
as a guide to child psychiatric caseness
and consequent burden. Journal of Child
Psychology and Psychiatry 1999; 40(5):
791-9.
11. LeFever GB, Dawson KV, Morrow AL.
The extent of drug therapy for attention
deficit-hyperactivity disorder among children
in public schools. American Journal of Public
Health 1999; 89(9): 1359-64.
12. Wolraich ML, Lindgren S, Stromquist A, et
al. Stimulant medication use by primary care
physicians in the treatment of attention deficit
hyperactivity disorder. Pediatrics 1990; 86(1):
95-101.
13. Charach A, Gajaria A. Improving
psychostimulant adherence in children with
ADHD. Expert Review of Neurotherapeutics
2008; 8(10): 1563-71.
14. Kelsey JL, Whittemore AS, Evans AS,
Thompson WD. Methods in Observational
Epidemiolgy. Second ed. New York: Oxford
Uniersity Press, 1996.
15. Robins LN. Epidemiology: Reflections on
testing the validity of psychiatric interviews
Archives of General Psychiatry 1985; 42(9):
918-24.
16. Statistics Canada and Human Resources
Development Canada. National Longitudinal
Survey of Children: Overview of Survey
Instruments for 1994-1995 Data Collection
Cycle 1. Ottawa: Minister of Industry, 1996.
17. Rust KF, Rao JNK. Variance estimation for
complex surveys using replication techniques.
Statistical Methods in Medical Research 1996;
5: 281-310.
18. Feinstein AR. Chapter 6: Probabilities and
Standardized Indexes. Principles of Medical
Statistics. Boca Raton, Fl: Chapman and Hall/
CRC, 2002: 81-4.
19. Statistics Canada. Low Income Cut-offs.
Ottawa: Minister of Industry, 1996.
20. Statistics Canada. Special Surveys Division.
Microdata User Guide National Longitudinal
Survey of Children and Youth Cycle 6,
September 2004-June 2005. Statistics Canada:
Ottawa, 2006.
21. Roberts G, Kovacevic M, Phillips O,
Gentleman J. Bridging the Gap between the
Theory and Practice of Analysis of Data from
Complex Surveys - Some Statistics Canada
Experiences. Washington, DC: Federal
Committee on Statistical Methodology
Conference. 1999.
22. SAS. Proprietary Software Release 8.2.
Cary, North Carolina: SAS Institute, Inc.,
1999-2001.
23. Offord DR, Boyle MH, Szatmari P, et al.
Ontario Child Health Study. II. Six-month
prevalence of disorder and rates of service
utilization. Archives of General Psychiatry
1987; 44(9): 832-6.
24. Breton JJ, Bergeron L, Valla JP, et al. Quebec
child mental health survey: prevalence of
DSM-III-R mental health disorders. Journal
of Child Psychology and Psychiatry 1999;
40(3): 375-84.
25. Waddell C, Offord DR, Shepherd CA, et al.
Child psychiatric epidemiology and Canadian
public policy-making: the state of the science
and the art of the possible. Canadian Journal
of Psychiatry 2002; 47(9): 825-32.
26. Goodman R, Ford T, Richards H, et al. The
Development and Well-Being Assessment:
description and initial validation of an
integrated assessment of child and adolescent
psychopathology. Journal of Child Psychology
and Psychiatry 2000; 41(5): 645-55.
27. Brownell MD, Yogendran MS.
Attention-deficit hyperactivity disorder in
Manitoba children: medical diagnosis and
psychostimulant treatment rates. Canadian
Journal of Psychiatry 2001; 46(3): 264-72.
28. Miller AR, Lalonde CE, et al. Prescription
of methylphenidate to children and youth,
1990-1996. Canadian Medical Association
Journal 2001; 165(11): 1489-94.
