Content uploaded by Catherine Bouchard
Author content
All content in this area was uploaded by Catherine Bouchard on Aug 02, 2018
Content may be subject to copyright.
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
1
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Evidence of rapid changes in Lyme disease awareness in Canada
Cécile Aenishaenslin1,2§, Catherine Bouchard3, Jules K. Koffi4, Yann Pelcat3, Nicholas H.
Ogden3
1 Department of Epidemiology, Biostatistics and Occupational Health, McGill University, 1140
Pine Avenue, Montreal H3A 1A3, Québec, Canada
2 Groupe de Recherche en Épidémiologie des Zoonoses et Santé Publique (GREZOSP), Faculté
de médecine vétérinaire, Université de Montréal, CP 5000, Saint-Hyacinthe J2S 7C6, Québec,
Canada
3 National Microbiology Laboratory, Public Health Agency of Canada, CP 5000, Saint-Hyacinthe
H2S 7C6, Québec, Canada
4 Centre for Food-borne, Environmental & Zoonotic Infectious Diseases, Public Health Agency of
Canada, CP 5000, Saint-Hyacinthe H2S 7C6, Québec, Canada
§Corresponding author: cecile.aenishaenslin@mail.mcgill.ca
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
2
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Abstract
Lyme disease (LD) is emerging in Canada. A key preventive strategy is
promoting the adoption by the general public of personal preventive behaviors regarding
tick bites. The aim of this study was to measure the changes in public awareness toward
ticks and LD before and after the launch of a national communication campaign in
Canada using data from two surveys conducted in March and December 2014. The
results show a significant increase in awareness of LD after compared to before the
campaign, but also suggest that the importance of this increase is not equal amongst
Canadian regions. Moreover, respondents whose level of awareness increased most
significantly were those who lived in regions with low entomologic risk. The findings
underline the importance of risk communications for emerging diseases and reinforce the
need to understand the specific characteristics of the targeted populations before the
implementation of communication campaigns to increase their efficacy.
Keywords: Lyme Disease, Lyme Borreliosis, Awareness, Behavior, Prevention &
control, Ticks, Tick Bites, Tick-Borne Diseases, Canada, Risk, Communication
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
3
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Background
Lyme disease (LD), a tick-borne disease caused by infection with the bacterium
Borrelia burgdorferi, is the most frequently reported vector-borne disease in the
temperate zone of the northern hemisphere. Global climate change may intensify the
threat of this vector-borne disease to human health in northern countries (Gage et al.,
2008). In Canada, LD is an emerging disease, with the number of reported human cases
increasing from 40 in 2004 to 682 in 2013 (Bouchard et al., 2015; Ogden et al., 2009).
Currently, the LD risk varies greatly across the country. By 2013, LD risk areas in
Canada (where the tick vectors, Ixodes scapularis and I. pacificus are established, and
where transmission of B. burgdorferi is confirmed or assumed to occur) had been found
in southern British Columbia, southern Manitoba, south and southeast Ontario, southern
Quebec, southern New Brunswick and parts of Nova Scotia (Bouchard et al., 2015;
Ogden et al., 2014).
No vaccine is currently available to prevent LD. Preventive strategies include
attempts to raise awareness in health professionals and in exposed human populations and
to promote the adoption of personal preventive behaviors, and tick control interventions
that decrease the contact rate between infected ticks and humans. Tick control
interventions include actions aimed at reducing tick density in the environment, such as
the application of acaricides or landscaping, as well as actions targeted at tick hosts such
as treatment of deer or rodents with topical or oral acaricides, the exclusion of deer by
fencing, vaccination of rodents, among others (Piesman and Eisen, 2008). Preventive
behaviors, such as wearing long trousers when visiting wooded areas, applying tick
repellent, checking for and removing ticks after visiting wooded areas, have
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
4
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
demonstrated some efficacy in preventing LD in populations (Finch et al., 2014; Lane et
al., 1992; Malouin et al., 2003; Mowbray et al., 2012; Smith et al., 2001; Vazquez et al.,
2008). Raising public awareness on LD and promoting these behaviors are among the
main objectives of the Canadian Action Plan on Lyme Disease that was launched in May
2014 in response to LD emergence in the country (Government of Canada, 2016). In this
context, the aims of this study were to measure the changes in public awareness toward
LD before and after the launch of the 2014 pan-Canadian communication campaign, and
to compare these changes between populations living in different regions of southern
Canada where Lyme disease risk is present or emerging and between populations living
in high vs low entomologic risk zones.
Methods
Description of the 2014 communication campaign on Lyme disease
In February 2014, the Public Health Agency of Canada (PHAC) launched a 3-
year social marketing campaign to help it deliver the engagement, education and
awareness pillar of its pan-Canadian Action Plan on Lyme Disease. The campaign’s
objective was to increase awareness of the disease, including preventive measures
Canadians can take to reduce the likeliness of contracting the disease from infected ticks.
The campaign targeted health professionals and Canadians engaging in outdoor activities
at a higher exposure risk to ticks (e.g. camping, hunting, fishing, hiking, property
maintenance). These efforts were coordinated with the provinces and territories. Key
messages on Lyme disease were developed in collaboration with jurisdictions and shared
with partners in advance of the tick season to foster consistent communications.
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
5
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Following an initial phase started in February 2014 to increase Canadian health
professionals’ awareness of Lyme disease, a second phase, which ran from May to
September 2014, targeted the public in order to increase Lyme disease awareness among
Canadians engaging in outdoor activities. It consisted of advertising; web content
enhancements to healthycanadians.gc.ca (now Canada.ca/Lyme); production of printed
collateral and web-based educational material; development and promotion of a
stakeholder toolkit; media relations; a partnership with Parks Canada to extend the reach
and visibility of messaging; and social media postings on Facebook, Twitter, Pinterest
and YouTube. Additional messaging also addressed increasing the audience’s knowledge
of preventive behaviors (e.g. wearing protective clothing, applying insect repellent
containing DEET or Icaridin, checks for and removal of ticks). The 2014 advertising
components included television, mobile app and geographically defined web-based
advertising; search engine marketing; paid social media posts; articles in online
magazines and both print ads in medical journals and web ads on websites targeting
health professionals. These ads generated 29.6 million impressions (the number of times
an ad appeared online) over the various platforms, which may approximate to the number
of individuals potentially exposed to an advertisement (Public Health Agency of Canada,
unpublished data).
Another key component of the campaign included the development and
distribution of a suite of awareness and educational material. Posters, handouts, brochures
and a stakeholders’ toolkit were made available free of charge online and in print for
individuals, stakeholders, healthcare professionals and other intermediaries. In total,
53,800 brochures, 35,500 handouts and 750 posters (including 150 specifically for pet
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
6
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
owners) were sent across the country, including a collaboration with Parks Canada which
placed on-site materials in 18 National Parks and Historic Sites. Close to 500,000 website
visits were registered to Lyme disease sections of the PHAC websites in 2014. In
addition, complementary social media messages received a high level of engagement
from its postings to the Healthy Canadians Facebook (6 posts), Twitter (19 tweets),
Pinterest (3 pins) and YouTube (1 video) platforms. Lastly, a series of articles, audio
news releases and a general public awareness video distributed to Canadian newspaper,
broadcast and online editors were published 175 times, reaching approximately 24
million users (Public Health Agency of Canada, unpublished data).
Study design
This study used data collected through two web-based surveys, one administered
in March (time t1, prior to the launch of the communication campaign in May 2014) and
one in December 2014 (time t2). The target population for this study consisted of
individuals living in the five study regions. The inclusion criteria for the two surveys
were to be 18 years or older, to be a resident of one of the five study regions and to speak
French or English. Respondents were recruited randomly from panels administered by
Canadian polling firms (Léger Marketing, 2016), from 11 to 25 March 2014 at t1, and
SOM (SOM, 2016), from 15 to 20 December at t2). The firms pre-tested the
questionnaires using a small group of panelists to verify understanding of questions. For
both administrations, samples were stratified on the five Canadian regions where there is
evidence of existing of emerging Lyme disease risk (Bouchard et al. 2015): Quebec,
Ontario, Atlantic (Prince Edward Island, New Brunswick, Nova Scotia and
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
7
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Newfoundland and Labrador), the Prairies (Manitoba, Saskatchewan and Alberta) and
British Columbia. This protocol was reviewed and approved by the Ethical Committee
for Health Research of the University of Montreal (Certificate number 14-088-CERES-
D).
Data collection
The questionnaires used for the surveys were developed based on the theories of
health behaviors, and more specifically on the Health Belief Model (Conner M., 2005)
and on a previous questionnaire used in Quebec to measure LD knowledge, attitudes and
behaviors (Aenishaenslin et al., 2015). The Health Belief Model is one widely used
theoretical model developed to study the main factors that can predict the adoption of
healthy behaviors, including risk perceptions, awareness of a disease and other cues to
actions (Conner M., 2005). The questionnaires, available in English and French, included
questions aiming to evaluate LD awareness: having heard of LD before the survey,
knowledge of symptoms (fever, headaches and fatigue, erythema migrans, arthritis,
manifestations of disseminated or post-treatment health problems), existence of a
treatment, knowledge of the transmission route, knowledge of the risk period and
knowledge of seven LD preventive behaviors (examining yourself for ticks and removing
them, wearing long clothing that covers the legs all the way down to your shoe, using
insect repellents containing DEET on skin and/or clothing, avoiding wooded areas during
the high-risk period, putting pesticides on your property, putting up barriers to keep deer
off your property, regularly mowing the lawn on your property). Note that it was decided
to use the term “chronic health problems” in the questionnaires for posttreatment health
problems as this term was considered more understandable by the general public,
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
8
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
however possible confusion with the controversial condition known as “chronic Lyme
disease” (Feder et al., 2007) was considered in interpretation of the results of analyses.
Demographic characteristics (gender, age, education level, family income, and province
of residency) and history with LD (having had LD, knowing someone with LD, perceived
LD risk) were also assessed in both questionnaires. Both were reviewed by a committee
of experts in public health and Lyme disease in Canada.
Data analysis
A global level of knowledge score (GKS, from 0 to 9) was computed based on the
sum of scores obtained on nine selected knowledge questions (one point per correct
answer): 1) fever, headaches and fatigue, 2) erythema migrans, 3) arthritis, 4)
manifestations of disseminated or post-treatment health problems, 5) transmission mode,
6) risk period, and knowledge of the following behavior as preventive measure for LD: 7)
examining yourself for ticks and removing them, 8) wearing long clothing , 9) using
insect repellents.
To evaluate the level of entomological risk for each respondent, the number of
tick submitted from 1991 to 2012 through the passive surveillance system per 100 000
inhabitants for each census subdivision (CSD) was used as an indicator. Only submitted
Ixodes scapularis ticks found on humans with known likely locations where they were
acquired were included for a total of 9,715 ticks (more details on passive surveillance
data can be found in Bouchard et al. (2015). Respondents postal codes were used to
match respondents to CSD. To compare groups and to perform multivariable regression,
the number of ticks submitted per 100,000 inhabitants was then dichotomized as high
entomological risk if in the upper 50th percentile and as low entomological risk if not.
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
9
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
To measure the changes in LD knowledge before and after the communication
campaign, differences in proportions were calculated for each knowledge item using
weighted data on gender, age, province and education level. Post stratification weights
were computed for each respondent using the 2011 Canadian census data (Statistics
Canada, 2016). This step led to the exclusion of respondents with missing data on gender,
age or education level (59 respondents excluded at t1 and 83 at t2). Pearson Chi-square
statistics were calculated to assess significant differences (p < 0.05) between regions and
time periods.
Multivariable logistic regression was used to measure the effect of the region, the
time period (t1 vs t2), and the entomological risk zone (high vs low) on the GKS
(high=GKS ≥6). Respondents from British Columbia were excluded from this analysis
because there were an insufficient number of respondents classified in high risk CSD (0
at t1 and 1 at t2). This is explained by the fact that only very few ticks were submitted to
the federal passive surveillance system as a province –specific surveillance system
operated here. Furthermore, I. scapularis are not present in this region and I. pacificus,
the main LD vector in British Columbia, is a species that does not quest high in the
herbage as does I. scapularis (Morshed et al., 2015), and are more difficult to collect in
passive surveillance. Therefore, using numbers of ticks submitted in passive surveillance
as an index of risk based on passive surveillance in regions where I. scapularis occurs is
not relevant for regions where I. pacificus occurs. Respondents that had reported that they
never heard about LD before the survey (527 respondents at t1 and 303 at t2),
respondents with missing data for gender, age or education level, and respondents with
invalid postal codes (197 respondents at t1 and 63 at t2) were excluded for this analysis
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
10
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
(final sample size for multivariable analysis=4173). Gender, age, and education level
were forced in the model as potential confounders. Interactions were tested between
included variables (shown in the result section only if significant). Descriptive and
multivariable statistical analyses were performed using IBM SPSS Statistics 19 (IBM
Corporation, Armonk, NY).
Results
A total of 2529 respondents participated at t1 and 2876 at t2 (geographical
distribution of respondents is represented in Figure 1). Differences were present in the
distribution of gender, age groups, family income and education level amongst regions
and between time periods (Table 1). Participants with a university degree were
overrepresented in both time periods and in all regions when compared to Canadian
census data (Statistics Canada, 2016). The proportion of respondents that declared they
never heard about LD decreased significantly between t1 and t2 in all regions (-12% for
all regions), the decrease was highest in the Atlantic region with -21%, and lowest in
British Columbia with -6% (Table 2). The pre and post proportions that declared they
never heard about LD also varied between regions, with a higher proportion observed in
Quebec at t1 with 50% of the respondents declaring they never heard about LD, and the
lowest being in the Prairies and Atlantic with 2% at t2 for both regions (Figure 1, Table
2).
When considering respondents who heard about LD before the survey, the
proportions with correct answers to 13 knowledge questions about LD also varied
between time periods and regions (Table 2). Significant increases in the proportions
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
11
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
between time periods were more frequent in the Prairie region and in Ontario, both with
seven knowledge items, followed by British Columbia with six items. In Quebec, the
proportions of respondents with correct answers increased for only one item
(transmission mode), and no significant increase was observed in Atlantic region.
Interestingly, proportions of respondents with correct answers on some preventive
measures questions decreased between t1 and t2 as shown in the overall statistics for:
avoiding wooded areas, putting pesticides on your property and mowing the lawn
regularly.
Overall, at t2, the proportions of respondents with correct answers reached at least
75% for 6 knowledge items: fever, headaches and fatigue (77%) as well as manifestations
of disseminated and post-treatment health problems (75%) as LD symptoms, the
transmission mode (the best known item with 91%), the risk period (78%), and tick check
(80%) and protective clothing (87%) as preventive measures. These levels of knowledge
varied also between regions at t2: respondents in British Columbia reached 75% for 7 of
13 items, those in the Prairies and Ontario for 6 items, those in the Atlantic region for 4
and for those in Quebec for only 2 items (Table 2). Confusion of “chronic” health
problems with “chronic Lyme disease” was not considered influential in the analyses
because the responses to questions regarding “chronic” health problems were always
significantly and positively correlated with responses to questions on other manifestations
of fever, headaches and fatigue, erythema migrans and arthritis (Pearson correlation
coefficient > 0.45, p<0.001 in all cases).
In the multivariable logistic regression model, some demographic variables were
significantly associated with a high GKS: being in the 18-34 age group was negatively
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
12
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
associated with high GKS, and having a university degree was positively associated with
high GKS (Table 3). Consistent with the descriptive analysis presented in Table 2, living
in Quebec or in the Atlantic region was negatively associated with a high GKS when
compared to the Prairie region. Regarding the variables of interest, living in a high
entomological risk CSD as well as the time period t2 were both associated with high
GKS, respectively with OR=1.82 (95%CI=1.39-2.40) and 1.97 (95%CI=1.52-2.56). The
interaction between living in a high entomological risk and t2 was negatively associated
with a high GKS with OR=0.69 (95%CI=0.48-0.99), suggesting that an effect of the
awareness campaigns by t2 was decreased knowledge in respondents living in a high risk
zone (Table 3).
This effect is illustrated in table 4: the changes in proportions of respondents with
high GKS and with good answers to the nine knowledge items included in the GKS are
presented separately in respondents living in low vs high entomologic risk CSDs. The
level of knowledge on LD was lower in lower vs higher risk zone at t1 for most items.
The increase in the respondents’ level of knowledge between t1 and t2 was greater for
respondents living in the low risk CSDs for all items (Table 4). Moreover, for 6
knowledge items of 9, the difference in proportion was equal or close to 0 at t2 indicating
an absence of changes for these items in this subgroup (Table 4).
Discussion
This study is the first to evaluate and compare the level of LD awareness and to
assess the impact of LD risk communications in the five regions of Canada. Overall, the
results showed that Canadians are aware of LD with almost 9 people in 10 having heard
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
13
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
about the disease in December 2014, a significant increase since March 2014, even
though the incidence of Lyme disease in Canada was still nationally low. Interestingly, a
national study conducted in the United States in 2004 found that 86.4% of the
participants have heard about LD, a comparable level to what was found in Canada at t2
(Herrington, 2004). Moreover, when considering those who heard about LD before the
study, 78% had a high level of knowledge at t2. Although it is not possible to attribute all
these changes to the 2014 national communication campaign, this is good news from a
public health perspective.
On the other hand, unexpected results were found regarding the differences in the
levels of knowledge amongst regions, and in the magnitude of the changes in knowledge
between regions and between respondents in low versus high entomologic risk CSD.
First, why the level of knowledge for some preventive measures such as avoiding
wooded areas, putting pesticides on its property, putting barriers and mowing the lawn
decreased is unclear. One hypothesis is that most of these preventive strategies were not
explicitly part of the communication campaign messages (except for keeping the grass
mowed), while other individual preventive strategies were explicitly mentioned. Also,
perhaps the absence of these measures from the campaign had an impact on perceptions
of their efficacy. Further research using a qualitative methodology would be needed to
explain this finding.
Second, in the Quebec region, our results suggest that the level of awareness on
LD in respondents from this region was lower in general at t1 than other regions, and that
it is also in this region that the change by t2 in the level of knowledge was the smallest.
At t2, still 33% of respondents from this region reported that they never heard about LD,
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
14
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
compared to only 2 to 5% in all four other regions. This observation is surprising given
that Quebec was the province with the third highest number of reported human cases of
LD after Ontario and Nova Scotia in 2012 (Public Health Agency of Canada, 2015). One
hypothesis that could explain this finding is that the majority of Quebec residents are
francophone (78 % in 2011) and while the 2014 awareness campaign was bilingual,
Quebec residents may have been less exposed to English-language media from the United
States, which may have been a much longer-term source of information on LD given the
history of LD in the United States.
The results also suggest that the effect of risk communications by t2 on LD
awareness was higher in lower entomological risk CSD. Public health authorities may
want to target populations that are at most risk of LD in a risk communication campaign.
However, in an emerging context and considering that the level of awareness in the
higher risk CSD at t1 was already good with 74% with a high GKS, this finding can be
interpreted as a positive outcome. Indeed, new high risk zones will probably be identified
in the next years in Canada (Leighton et al., 2012), and the fact that the campaigns by t2
had an effect of “pulling” the level of LD awareness in lower risk zones to the level
found in higher risk zones could accelerate the adoption of preventive behaviors in new
emerging high risk regions. Herrington (2004) also compared knowledge, perceptions
and behaviors regarding LD between low and high incidence States in the United States
and found inconsistent differences in proportions between groups, but no statistical tests
were performed to test significance of these differences.
Considering the findings globally, this study underlines the importance of
unerstanding the specific characteristics of the targeted populations before the
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
15
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
implementation of communication programs to increase its outreach and its efficacy. This
conclusion is consistent with previous research comparing the knowledge, risk perception
and behaviors toward LD between one Canadian and one Swiss regions with different
epidemiologic and socio-cultural context (Aenishaenslin et al., 2015, 2014). This study is
a first step in understanding differences in LD awareness among Canadians, and results
serve as a baseline for more research on the evolution of the level of awareness, and on
the explanations of discrepancies that were found among subgroups of respondents.
This study has limitations. First of all, web-panels from two different firms were
used to recruit participants. The study was then restricted to internet users, and the source
population for our samples was different at t1 and t2. However, this is unlikely to have
had an important impact on our results. Both firms used a similar methodology for their
web-based surveys: respondents were recruited randomly from each firm’s web-panels.
Web-panelists are recruited by the firm on a voluntary basis and through random phone
surveys in order to have a representative sample of the Canadian population in terms of
the main demographic characteristics. Phone surveys are still regarded by many as the
reference method for obtaining a random sample, but due to current changes in telephony
networks, including the transition to mobile phones, there is uncertainty that phone
survey is the methodology that produces the least biased results. Some studies suggest
that methods using internet questionnaires have as good or better external validity
(Gosling et al. 2004; Braunsberger et al. 2007), when compared to other conventional
methods. Moreover, we used analytical control techniques to correct for differences in
demographic variables that may have an impact on the results: post-stratification weights
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
16
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
computed on these variables were applied on proportions, and these variables were
included in the multivariable model as potential confounders.
Another important point is that the study design consisted of two cross-sectional
studies and the respondents included at t1 differ from those recruited at t2. In addition,
the exposure to the 2014 Canadian communication campaign on LD was not quantified
precisely in the surveys. Therefore, the association found between the time period t2 and
a higher level of knowledge cannot be attributed specifically to this campaign and we
cannot presume direct causal relationships. Provinces and regional health units can also
have their own Lyme disease communication programs that differ between them, and
respondents were probably exposed to and influenced by public messages that are not
linked to the federal campaign. Further studies, conducted at the provincial levels, are
needed to assess more specifically the changes in LD awareness following provincial and
regional communication campaigns. Secondly, results were adjusted for gender, age,
education level and province of residency, but other unmeasured factors may have an
impact on the results, such as cultural identity or history of travelling in higher risk
countries.
To evaluate the level of entomological risk for each respondent, we used the
number of ticks submitted per 100,000 inhabitants through the passive surveillance
system in the respondents’ CSD. However, it is important to underline that active
surveillance strategies are needed to confirm the establishment of I. scapularis ticks in a
region and to identify entomological risk zones (Ogden et al., 2010). As active
surveillance has not been performed in a systematic way in all CSD across provinces, we
chose to use the passive tick surveillance indicator instead. More work is needed to
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
17
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
evaluate how well this indicator corresponds to regions where vector populations are
truly established.
Finally, we have to keep in mind that the ultimate objective of the communication
campaign is to reduce the risk of LD in the Canadian population, but that the effect of the
campaign on the incidence of LD in humans is difficult to assess. The major reason is
that LD is emerging in Canada with the expansion of the geographic range of established
blacklegged tick populations, therefore it is expected that the incidence will continue to
rise in the near future. Longitudinal monitoring of changes in both awareness and
incidence in different Canadian regions would be of great use to help understand the
efficacy of communications to prevent LD.
Conclusion
This study revealed a rapid and significant increase in LD awareness in Canada
following the launch of the 2014 national communication campaign. These results
reiterate the impact and the importance of risk communications when implementing a
prevention program for emerging diseases. The differences found between respondents
living in higher and lower risk zones and the particular profile of Quebec also underline
the crucial importance of understanding the context and specific characteristics of the
populations in order to establish effective interventions that are well adapted to targeted
populations and their epidemiological contexts.
Disclosures
This work was funded by the Public Health Agency of Canada (PHAC). The
views expressed in the publication are those of the authors and not necessarily those of
the PHAC or the institutions with which the authors are affiliated.
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
18
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Acknowledgements
The authors wish to thank Pascal Michel for its contribution to the development of
the questionnaire, and Daniel Morier, from Health Canada and the Public Health Agency
of Canada’s Marketing and Creative Services Team, for details and data on the 2014
communication campaign on LD. We also thank federal and provincial partners across
the country who participated in the validation of the questionnaire.
References
Aenishaenslin, C., Michel, P., Ravel, A., Gern, L., Milord, F., Waaub, J.P., Belanger,
D., 2015. Factors associated with preventive behaviors regarding Lyme disease in
Canada and Switzerland: a comparative study. BMC Public Health 15, 185.
doi:10.1186/s12889-015-1539-2
Aenishaenslin, C., Ravel, A., Michel, P., Gern, L., Milord, F., Waaub, J.P., Belanger,
D., 2014. From Lyme disease emergence to endemicity: a cross sectional comparative
study of risk perceptions in different populations. BMC Public Health 14, 1298.
doi:10.1186/1471-2458-14-1298
Bouchard, C., Leonard, E., Koffi, J.K., Pelcat, Y., Peregrine, A., Chilton, N., Rochon,
K., Lysyk, T., Lindsay, L.R., Ogden, N.H., 2015. The increasing risk of Lyme disease
in Canada. Can Vet J 56, 693–9.
Braunsberger, K., Wybenga, H., Gates. R. 2007. A comparison of reliability between
telephone and web-based surveys. J Bus Res 60: 758–64.
doi:10.1016/j.jbusres.2007.02.015.
Conner M., N.P., 2005. Predicting Health Behaviour. Open University Press.
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
19
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Feder, H.M., Johnson, B.J.B., O’Connell, S., Shapiro, E.D., Steere, A.C., Wormser,
G.P., 2007. A Critical Appraisal of “Chronic Lyme Disease.” N. Engl. J. Med. 357,
1422–1430. doi:10.1056/NEJMra072023
Finch, C., Al-Damluji, M.S., Krause, P.J., Niccolai, L., Steeves, T., O’Keefe, C.F.,
Diuk-Wasser, M.A., 2014. Integrated assessment of behavioral and environmental
risk factors for Lyme disease infection on Block Island, Rhode Island. PLoS One,
e84758. doi:10.1371/journal.pone.0084758
Gage, K.L., Burkot, T.R., Eisen, R.J., Hayes, E.B., 2008. Climate and vectorborne
diseases. Am J Prev Med 35, 436–50. doi:10.1016/j.amepre.2008.08.030
Government of Canada. Action plan on Lyme disease. Available online
http://healthycanadians.gc.ca/diseases-conditions-maladies-affections/disease-
maladie/lyme/plan-eng.php (last accessed 29.03.16).
Government of Canada. Lyme disease awareness resources. Available online
http://healthycanadians.gc.ca/diseases-conditions-maladies-affections/disease-
maladie/lyme/resources-ressources/index-
eng.php?_ga=1.22863803.1836517318.1458045953 (last accessed 07.06.16)
Gosling, S. D., Vazire S., Srivastava S., John O.P. 2004. Should we trust web-based
studies? A comparative analysis of six preconceptions about internet questionnaires.
Am Psychol 59: 93–104. doi:10.1037/0003-066X.59.2.93.
Herrington, J. E., 2004. Risk perceptions regarding ticks and Lyme disease. Am J
Prev Med 26, 135–40. doi:10.1016/j.amepre.2003.10.010.
Hinckley, A.F., Meek, J.I., Ray, J.A.E., Niesobecki, S.A., Connally, N.P., Feldman,
K.A., Jones, E.H., Backenson, P.B., White, J.L., Lukacik, G., Kay, A.B., Miranda,
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
20
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
W.P., Mead, P.S., 2016. Effectiveness of residential acaricides to prevent Lyme and
other tick-borne diseases in humans. J. Infect. Dis. doi:10.1093/infdis/jiv775
Lane, R.S., Manweiler, S.A., Stubbs, H.A., Lennette, E.T., Madigan, J.E., Lavoie,
P.E., 1992. Risk factors for Lyme disease in a small rural community in northern
California. Am J Epidemiol 136, 1358–68.
Leger. Leger: the research intelligence group. Available online
http://leger360.com/en-ca/legerwebpanel.asp (last accessed 29.03.16).
Leighton, P.A., Koffi, J.K., Pelcat, Y., Lindsay, L.R., Ogden, N.H., 2012. Predicting
the speed of tick invasion: an empirical model of range expansion for the Lyme
disease vector Ixodes scapularis in Canada. J. Appl. Ecol. 49, 457–464.
doi:10.1111/j.1365-2664.2012.02112.x
Malouin, R., Winch, P., Leontsini, E., Glass, G., Simon, D., Hayes, E.B., Schwartz,
B.S., 2003. Longitudinal evaluation of an educational intervention for preventing tick
bites in an area with endemic Lyme disease in Baltimore County, Maryland. Am J
Epidemiol 157, 1039–51.
Morshed, M., Lee, M., Man, S., Fernando, K., Wong, Q., Hojgaard, A., Tang, P.,
Mak, S., Henry, B., Patrick, D., 2015. Surveillance for Borrelia burgdorferi in Ixodes
ticks and small rodents in British Columbia. Vector Borne Zoonotic Dis. 15, 701–
705.
Mowbray, F., Amlot, R., Rubin, G.J., 2012. Ticking all the boxes? A systematic
review of education and communication interventions to prevent tick-borne disease.
Vector Borne Zoonotic Dis 12, 817–25. doi:10.1089/vbz.2011.0774
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
21
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Ogden, N.H., Bouchard, C., Kurtenbach, K., Margos, G., Lindsay, L.R., Trudel, L.,
Nguon, S., Milord, F., 2010. Active and passive surveillance and phylogenetic
analysis of Borrelia burgdorferi elucidate the process of Lyme disease risk
emergence in Canada. Env. Health Perspect 118, 909–14. doi:10.1289/ehp.0901766
Ogden NH, Koffi JK, Pelcat Y, Lindsay LR. Environmental risk from Lyme disease
in central and eastern Canada: a summary of recent surveillance information. Can
Commun Dis Rep. 2014. Epub 2014 March 6.
Ogden, N.H., Lindsay, L.R., Morshed, M., Sockett, P.N., Artsob, H., 2009. The
emergence of Lyme disease in Canada. CMAJ 180. doi:10.1503/cmaj.080148
Piesman, J., 2006. Strategies for reducing the risk of Lyme borreliosis in North
America. Int J Med Microbiol 296 Suppl 40, 17–22. doi:10.1016/j.ijmm.2005.11.007
Piesman, J., Eisen, L., 2008. Prevention of tick-borne diseases. Annu Rev Entomol
53, 323–43. doi:10.1146/annurev.ento.53.103106.093429
Public Health Agency of Canada. National Lyme disease surveillance in Canada
2009-2012 Available online http://healthycanadians.gc.ca/publications/diseases-
conditions-maladies-affections/2009-2012-lyme/index-eng.php#a4 (last accessed
09.02.16).
Smith, G., Wileyto, E.P., Hopkins, R.B., Cherry, B.R., Maher, J.P., 2001. Risk factors
for Lyme disease in Chester County, Pennsylvania. Public Health Rep 116 Suppl 1,
146–56.
SOM. SOM Research and surveys Available online http://www.som.ca/en (last
accessed 29.03.16)
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
22
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Statistics Canada. The Canadian population in 2011 Available online
http://www12.statcan.ca/census-recensement/2011/as-sa/98-310-x/98-310-x2011001-
eng.cfm (last accessed 15.02.16)
Vazquez, M., Muehlenbein, C., Cartter, M., Hayes, E.B., Ertel, S., Shapiro, E.D.,
2008. Effectiveness of personal protective measures to prevent Lyme disease. Emerg
Infect Dis 14, 210–6.
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
23
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Table 1. Demographic characteristics of respondents (unweighted data)
British Columbia
Prairies
Ontario
March
December
March
December
March
December
n
%
n
%
n
%
n
%
n
%
n
%
Total
302
100
355
100
426
100
494
100
1008
100
1015
100
Gender
Women
156
52
220
62
185
43
262
53
497
49
637
63
Men
146
48
134
38
241
57
230
47
511
51
376
37
Age
18-25y
29
10
0
0
44
10
2
0
93
9
1
0
25 -34y
50
17
17
5
48
11
24
5
134
13
68
7
35-44y
50
17
50
14
60
14
67
14
167
17
193
19
45-54y
67
22
77
22
105
25
93
19
244
24
207
20
55-64y
50
17
106
30
86
20
144
29
175
17
288
28
65-74y
47
16
85
24
59
14
115
23
174
17
186
18
75y+
9
3
20
6
24
6
47
10
21
2
65
6
Total
302
102
355
101
426
100
492
100
1008
99
1008
98
Prefer not to answer
na
na
0
0
na
na
2
0
na
na
7
1
Family income
Less than 20 000$
40
16
19
7
31
9
19
5
99
12
34
4
20 000-39 999$
81
32
51
18
56
17
63
16
198
23
120
15
40 000-59 999$
42
17
53
19
64
19
56
14
144
17
126
16
60 000-79 999$
36
14
54
19
43
13
76
20
128
15
128
16
80 000-99 999$
25
10
35
12
42
12
51
13
109
13
98
12
100 000-119 999$
16
6
24
8
48
14
38
10
76
9
100
13
120 000 $ and more
13
5
48
17
53
16
85
22
95
11
182
23
Total
253
100
284
100
337
100
388
100
849
100
788
100
Prefer not to answer
49
na
71
na
89
na
106
na
159
na
227
na
Education level
Less than high school
4
1
2
1
3
1
1
0
9
1
3
0
High school completed
and equivalent
87
30
47
14
114
28
66
14
271
28
137
14
College completed and
equivalent1
100
34
213
62
152
37
303
63
355
36
538
54
University degree and
equivalent
101
35
84
24
145
35
114
24
349
35
310
31
Total
292
100
346
100
414
100
484
100
984
100
988
100
Other
1
Na
3
na
5
na
3
na
11
na
6
na
Prefer not to answer
9
Na
6
na
7
na
7
na
13
na
21
na
1In Canada, except in Quebec, college is a non-mandatory post-secondary education school offering applied
career-oriented programs (vs university which offers academic education). In Quebec, two years in college
are mandatory for students before starting university.
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
24
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Table 1. (continued)
Québec
Atlantic
Overall
March
December
March
December
March
December
n
%
n
%
n
%
n
%
n
%
n
%
Total
631
100
826
100
162
100
186
100
2529
100
2876
100
Gender
Women
314
50
428
52
74
46
112
60
1226
48
1659
58
Men
317
50
398
48
88
54
74
40
1303
52
1212
42
Age
18-25y
103
16
78
9
13
8
0
0
282
11
81
3
25 -34y
79
13
132
16
21
13
12
6
332
13
253
9
35-44y
87
14
155
19
30
19
22
12
394
16
487
17
45-54y
125
20
203
25
38
23
52
28
579
23
632
22
55-64y
111
18
161
19
29
18
59
32
451
18
758
26
65-74y
104
16
76
9
23
14
33
18
407
16
495
17
75y+
22
3
19
2
8
5
8
4
84
3
159
6
Total
631
100
824
99
162
100
186
100
2529
100
2865
100
Prefer not to answer
na
na
2
0
na
na
0
0
na
na
11
0
Family income
Less than 20 000$
63
12
61
9
15
11
15
10
248
10
148
6
20 000-39 999$
111
20
118
17
54
38
34
23
500
20
386
17
40 000-59 999$
120
22
137
20
21
15
33
22
391
15
405
18
60 000-79 999$
72
13
123
18
19
13
22
15
298
12
403
18
80 000-99 999$
74
14
87
13
14
10
11
7
264
10
282
12
100 000-119 999$
54
10
69
10
8
6
14
9
202
8
245
11
120 000 $ and more
51
9
97
14
10
7
20
13
222
9
432
19
Total
545
100
692
100
141
100
149
100
2125
84
2301
100
Prefer not to answer
86
na
134
na
21
na
37
na
404
16
575
na
Education level
Less than high school
14
2
6
1
1
1
2
1
31
1
14
0
High school completed
and equivalent
193
31
171
21
59
37
22
12
724
29
443
16
College completed1 and
equivalent
212
34
327
40
54
34
114
64
873
35
1495
53
University degree and
equivalent
201
32
304
38
46
29
41
23
842
33
853
30
Total
620
100
808
100
160
100
179
100
2470
98
2805
100
Other
5
na
11
na
2
na
3
na
24
1
26
na
Prefer not to answer
6
na
7
na
0
na
4
na
35
1
45
na
1In Canada, except in Quebec, college is a non-mandatory post-secondary education school offering applied
career-oriented programs (vs university which offers academic education). In Quebec, two years in college
are mandatory for students before starting university.
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
25
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Table 2. Changes in knowledge on Lyme disease (LD) by region and time period (data
represent the weighted proportion of respondents with good answer for each knowledge
item among the respondents who declared that they heard about LD before the surveys
(n=4433), except for ‘never heard of LD’ which is the weighted proportion on the entire
sample, excluding respondents with missing data on age, gender, or education level
(n=5263))
British Columbia (%)
Prairies (%)
Ontario (%)
t1
(%)
t2
(%)
Δ%
t1
(%)
t2
(%)
Δ%
t
1
(
%
)
t2
(%)
Δ%
Never heard of Lyme disease
11
5
-6*
16
2
-14***
1
5
5
-10***
Subsample (n)
267
270
278
304
8
9
0
971
Knowledge of symptoms
Fever, headaches and fatigue
68
80
11**
73
79
6
7
3
80
7***
Erythema migrans
59
60
1
59
65
6
6
1
66
4
Arthritis
54
73
19**
*
58
69
11**
6
1
70
9***
Disseminated/post-treatment health problems
68
83
16**
*
64
80
16***
6
8
78
10***
Transmission and risk period
Transmission mode
82
92
10**
*
85
92
7**
8
3
94
11***
Risk period
58
70
12**
72
88
16***
7
1
78
8***
Knowledge of preventive measures
Examining yourself for ticks
77
83
6
79
88
9**
7
8
84
5***
Wearing long clothing
73
89
17**
*
77
87
10**
7
9
88
9***
Using insect repellents
46
48
2
55
64
10*
6
4
65
1
Avoiding wooded areas
75
76
1
80
81
1
8
0
73
-6***
Putting pesticides on your property
23
16
-8*
21
22
1
2
7
20
-8***
Putting up barriers
30
35
5
31
33
2
3
6
33
-3
Regularly mowing the lawn on your property
28
36
7
43
45
2
4
4
35
-9***
Global level of knowledge (GKS ≥6)
63
77
14**
*
71
82
11**
7
2
81
9***
*p < 0.05,** p < 0.01, *** p < 0.001 (Pearson Chi-square, between t1 and t2)
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
26
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Table 2. (continued)
Quebec (%)
Atlantic (%)
Overall (%)
t1
(%)
t2
(%)
Δ%
t1
(%)
t2
(%
)
Δ%
t1
(%)
t2
(%)
Δ%
Never heard of Lyme disease
50
33
-17***
23
2
-21***
24
12
-12***
Subsample (n)
337
576
73
83
2002
2573
Knowledge of symptoms
Fever, headaches and fatigue
74
70
-4
74
73
0
72
77
5***
Erythema migrans
61
58
-3
64
71
7
61
64
3
Arthritis
55
57
2
59
70
11
58
68
10***
Disseminated/post-treatment health problems
62
58
-3
68
73
5
66
75
9***
Transmission and risk period
Transmission mode
75
83
8**
88
88
0
82
91
9***
Risk period
75
75
0
81
80
-1
70
78
8***
Knowledge of preventive measures
Examining yourself for ticks
57
61
4
85
84
-1
75
80
5***
Wearing long clothing
79
84
5
79
88
8
78
87
9***
Using insect repellents
54
56
2
55
58
3
57
60
3
Avoiding wooded areas
76
69
-7
84
73
-10
79
74
-5**
Putting pesticides on your property
19
16
-4
25
20
-4
24
19
-5***
Putting up barriers
18
12
-6*
45
36
-9
31
29
-2
Regularly mowing the lawn on your property
31
26
-5
41
45
3
39
35
-4*
Global level of knowledge (GKS ≥6)
62
70
8**
73
77
4
69
78
9***
*p < 0.05;** p < 0.01; *** p < 0.001 (Pearson Chi-square, between t1 and t2)
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
27
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Table 3. Results of multivariable logistic regression measuring the association
between time period and high entomological risk CSD on global knowledge score
(dependant variable is GKS≥6=1) (n=3547, excluding British Columbia
respondents)
OR
95%CI
p
Gender (men=reference)
1.04
(0.87-1.25)
0.7
Age group
18-34yr
0.65
(0.52-0.83)
.000
35-54yr
0.91
(0.74-1.12)
.4
55+ (ref)
1
-
-
University degree
1.51
(1.24-1.83)
0.000
Region of residency
Atlantic
0.63
(0.40-0.99)
0.04
Quebec
0.39
(0.29-0.53)
.000
Ontario
0.79
(0.61-1.03)
0.08
Prairies (ref)
1
-
-
High entomological risk CSD¶
1.82
(1.39-2.40)
.000
Time period 2
1.97
(1.52-2.56)
.000
t2 * High entomological risk CSD¶
0.69
(0.48-0.99)
0.046
¶Census subdivision
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
28
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Table 4. Weighted proportions of respondents with high global knowledge score
(GKS≥6) and with good answers to 10 knowledge questions regarding Lyme
disease (LD) between t1 and t2 in populations living in lower and higher
entomologic risk CDS. Analysis was restricted to respondents who declared that
they have heard about LD before the survey, and excluded respondents from BC,
and with missing data for age, education level and postal codes (n=2714).
Lower entomological risk CSD
(%)
Higher entomological risk CSD
(%)
t1
t2
Δ
t1
t2
Δ
High GKS
67
80
13
74
79
5
Fever, headache, fatigue
70
80
10
77
76
-1
Erythema migrans
57
67
10
66
63
-3
Arthritis
58
69
11
64
66
2
Disseminated/post-treatment health
problems
62
76
14
72
72
0
Transmission by ticks
81
90
9
84
92
8
Risk period
68
79
9
79
79
0
Tick check
74
82
8
76
76
0
Protective clothing
76
87
11
82
87
5
Tick repellent
56
67
9
61
59
-2
Accepted manuscript : http://dx.doi.org/10.1016/j.ttbdis.2016.09.007
29
© 2016. This manuscript version is made available under the CC-BY-NC-ND 4.0 license
http://creativecommons.org/licenses/by-nc-nd/4.0/
Figure 1. Geographical distribution of respondents, and differences in LD awareness
between regions.