Weight loss by mobile phone: A 1-year effectiveness study

Abstract

To investigate the short- and long-term effectiveness and the predictors of weight loss in a mobile phone weight-loss programme among healthy overweight adults.
One hundred and twenty-five healthy, overweight (BMI = 26-36 kg/m2), 25-44-year-old, screened volunteers were randomized to an experimental group (n 62) to use a mobile phone-operated weight-loss programme or to a control group (n 63) with no intervention. Via text messaging, the programme instructed a staggered reduction of food intake and daily weight reporting with immediate tailored feedback. Assessments were at 0, 3, 6, 9 and 12 months for the experimental group; at 0 and 12 months for the control group. Main outcome variables were changes in body weight and waist circumference.
By 12 months the experimental group had lost significantly more weight than the control group (4.5 (sd 5.0) v. 1.1 (sd 5.8) kg; F(1,80) = 8.0, P = 0.006) and had a greater reduction in waist circumference (6.3 (sd 5.3) v. 2.4 (sd 5.4) cm; F(1,80) = 55.2, P = 0.0001). Early weight loss, self-efficacy, contact frequency, attitudes towards the medium, changes in work and family life and changes made in dietary habits were the strongest predictors of weight loss.
This mobile phone weight-loss programme was effective in short- and long-term weight loss. As a minimum-advice, maximal-contact programme, it offers ideas for future weight-loss programmes.

Full text

Public Health Nutrition: 12(12), 2382 –2391 doi:10.1017/S1368980009005230
Weight loss by mobile phone: a 1-year effectiveness study
Irja Haapala
1,
*, Noe
¨l C Barengo
2
, Simon Biggs
3
, Leena Surakka
4
and Pirjo Manninen
5
1
Department of Public Health, University of Kuopio, Kuopio, Finland and Department of Education, University
of Joensuu, POB 86, 57101 Savonlinna, Finland:
2
Department of Public Health, University of Helsinki, Helsinki,
Finland and Unit of Epidemiology and Clinical Research, University Hospital La Paz, Madrid, Spain:
3
School of
Social Studies, King’s College London, London, UK:
4
Department of Public Health, University of Kuopio, Kuopio,
Finland and Finnish Institute of Occupational Health, Helsinki, Finland:
5
Department of Clinical Nutrition and
Department of Medicine, University of Kuopio, Kuopio, Finland
Submitted 30 May 2008: Accepted 21 January 2009: First published online 27 March 2009
Abstract
Objective: To investigate the short- and long-term effectiveness and the predictors
of weight loss in a mobile phone weight-loss programme among healthy over-
weight adults.
Design: One hundred and twenty-five healthy, overweight (BMI 526–36 kg/m
2
),
25–44-year-old, screened volunteers were randomized to an experimental group
(n62) to use a mobile phone-operated weight-loss programme or to a control
group (n63) with no intervention. Via text messaging, the programme instructed
a staggered reduction of food intake and daily weight reporting with immediate
tailored feedback. Assessments were at 0, 3, 6, 9 and 12 months for the experi-
mental group; at 0 and 12 months for the control group. Main outcome variables
were changes in body weight and waist circumference.
Results: By 12 months the experimental group had lost significantly more weight
than the control group (4?5(
SD 5?0) v.1?1(SD 5?8) kg; F(1,80) 58?0, P50?006)
and had a greater reduction in waist circumference (6?3(
SD 5?3) v.2?4(SD 5?4) cm;
F(1,80) 555?2, P50?0001). Early weight loss, self-efficacy, contact frequency,
attitudes towards the medium, changes in work and family life and changes made
in dietary habits were the strongest predictors of weight loss.
Conclusions: This mobile phone weight-loss programme was effective in short-
and long-term weight loss. As a minimum-advice, maximal-contact programme, it
offers ideas for future weight-loss programmes.
Keywords
Mobile phone
Internet
Contact
Weight maintenance
Interactive communication technology (using the Internet,
email and mobile phones) offers an innovative and attrac-
tive tool for weight-loss programme delivery. Traditional
face-to-face weight-loss programmes are increasingly being
enhanced by Internet support
(1–3)
. Specific mobile phone
applications for this purpose are emerging, but the effec-
tiveness of this medium and these new programmes in
supporting weight loss have yet to be reported.
Previous research into the effectiveness of Internet-based
weight-loss programmes
(4–7)
has presented short- to med-
ium-term results indicating slightly better performance or no
difference when compared with traditional programmes.
Many have used email or telephone to support a web-based
programme with weekly assignments. They have mostly
been labour-intensive, requiring considerable counselling
input from health-care professionals. In the current study
we investigated the effectiveness of a programme providing
minimal advice and no counselling but a maximum possi-
bility for user-initiated contact and connectedness via text
messaging.
Effective weight-loss programmes must support dieters
in the process of learning and adopting new dietary and
physical activity behaviours
(8,9)
.Atheoreticalmodel
(10)
to
guide research into educational/behavioural interventions
utilizing new, interactive media suggests that the amount,
frequency and type of use of the programme (contact)
influences learning effectiveness. This model, combined
with Bandura’s
(11)
self-efficacy theory, suggests that atti-
tudes to teletechnology and perceptions of personal self-
efficacy in dieting will influence contact and the use made
of the programme and thereby may affect weight loss.
External life-events and circumstances would exert an
additional influence. Guided by the theoretical model
presented in Fig. 1, we examined the possible influence on
weight loss exerted by selected background characteri-
stics (exogenous variables), process variables and contact
with the programme. The current paper reports short-term
(3-month) and long-term (12 months) results from a 1-year
study into the effectiveness of a mobile phone weight-loss
programme among healthy overweight adults.
*Corresponding author: Email irja.haapala@joensuu.fi rThe Authors 2009

Methods
Participants
One hundred and fifty-six healthy adult volunteers (120
women and thirty-six men) were recruited via newspaper
advertisement and telephone screening. The chosen
sample size allowed for 20 % ineligible volunteers and
30 % attrition rate to give a large enough sample to detect
large effects (effect size 50?40) with a50?05 and power
of 0?80 in a 2 (treatment group) 32 (pre-test/post-test)
repeated-measures ANOVA. A total of 125 volunteers who
met the eligibility criteria (age 25–44 years, BMI 525–36
kg/m
2
, access to a mobile phone and an Internet con-
nection, no diagnosed chronic disease, no major psy-
chiatric disease and no current, planned or previous
pregnancy within 6 months) started in the study. Disease
and general health-related data were based upon self-
report and were discussed with the study nurse. The
study nurse was blind to the randomization procedure,
which was performed within gender, to an experimental
group (n62) or a control group (n63). Signed informed
consent was obtained from all participants. The partici-
pant flow is presented in Fig. 2 and the participants’
characteristics in Table 1.
Design
The randomized controlled study ran from June 2001 to
June 2002. To ensure objectivity and validity of weight
loss, the experimental group was invited to the study
centre at 3-month intervals during the 12 months of the
study. The control group was invited for the baseline and
the 12-month visit. Its purpose was to control for threats to
internal validity (caused by selection or history) and for
possible concomitant launch of a new weight-loss pro-
gramme in the area. The control group received no inter-
vention but was offered the studied weight-loss programme
free of charge after the 12-month visit. No specific instruc-
tion on diet or exercise was given to either group. Because
the programme was intended to serve as a support to self-
directed dieters, self-directed dieting or joining another
weight-loss programme was not forbidden in either group.
Ethical approval for the study was obtained from the ethical
committee for human research at the University of Kuopio
and the Kuopio University Hospital.
Measurements
Outcome variables
Weight, height and waist circumference measurements
were performed at each follow-up by two study nurses
according to standardized procedures
(12)
.
User opinions about the programme’s operation and
usefulness included a grade (mark) given on a scale from
4 to 10 as in the Finnish school system. Users’ liking/
attitudes towards mobile phones and the Internet were
assessed with a yes/no question.
Amount, frequency and type of use of the programme
were assessed through self-reported frequency of weight
reporting via text messaging or to the website, and the
use of different parts of the programme. Options were:
15‘every day or more’ (scored as 7 times/week); 2 5‘2–3
times per week’ (scored as 2); 3 5‘once per week’
(scored as 1); 4 5‘1–2 times per month’ (scored as 0?5);
Changes in process variables
Change in:
• Dietary habits and intake
• Physical activity
• Perceived self-efficacy in
dieting
• Life situation at home/work
Mobile phone weight-loss programme Background characteristics =
Exogenous variables
• Education, age, marital status
• Technology liking and use
• Perceived self-efficacy in dieting,
at baseline (indication of
motivation)
• Previous dieting, preference for
group/individual dieting
• Baseline diet and physical activity
Amount, frequency and type of
use of the programme = Contact
• Reporting of weight
• Use of programme
components
• Contact with other dieters
Outcome variables
• Weight loss
• User satisfaction
Technology
• Access
• Usability
• Reliability
Programme
• Content
• Type of
support
Fig. 1 Contingency model in mobile phone weight loss
Mobile weight loss 2383

55‘less than once per month’ (scored as 0?2); 6 5‘not at
all’ (scored as zero). Further feedback was collected with
multiple choice questions to assess the type of use and
the importance of different aspects of the programme.
Process variables
Dietary habits were assessed at 0, 6 and 12 months with
questions related to the self-reported frequency of consum-
ing eight energy-dense foods. The options for each ranged
from 1 5‘less than once per month or never’ to 5 5‘once
per day or more often’. The scores were summed up to
form the ‘energy-dense food score’ (internal consistency
coefficient a50?71). Self-reported changes in dietary habits
were assessed also with open-ended questions at 3 months.
Dietary intake was assessed with 3 d dietary records by
household measures and analysed using the Nutrica

R
nutrient analysis program version 3?1 (The Social Insurance
Institution of Finland, Turku, Finland, 2000).
Frequency of leisure-time physical activity was assessed
with a question adopted from the Finnish national health
surveys (conducted by the National Public Health Institute),
using seven categories ranging from 0 5‘cannot’ and 1 5‘a
couple of times per year or less’ to 6 5‘daily’.
Self-efficacy in dieting, denoting trust in one’s cap-
ability in achieving self-set goals for weight loss, reducing
food intake, increasing physical activity and maintaining
the weight loss, was assessed with ten items (a50?84)
with a scale ranging from 0 5‘not at all certain’ to
95‘absolutely certain’ (adapted from Bandura
(11)
).
Changes at work and in family life were assessed with
open-ended questions. The use of additional sources of
information on nutrition or physical activity, and partici-
pation in other weight-loss programmes during the past
year, were also assessed.
The mobile phone weight-loss programme
The present study investigated the effectiveness of a
mobile phone-operated weight-loss programme, Weight
Balance

R
(GeraCap Invia Ltd, Seina
¨joki, Finland), laun-
ched in Finland in 2001. All expenses accrued due to this
programme were covered. The programme calculated the
dieter’s daily energy requirement using an equation of
31 excluded
23 ineligible subjects
8 absent
1 refused to
participate
- excluded from
analyses due to
incomplete
baseline data
63 assigned to control group
62 assigned to experimental group
At 3 months, n 56 completers*
5 discontinued† intervention (8 %)
1 due to long distance
1 unhappy with programme
3 unknown reasons
1 unable to attend
At 6 months, n 45 completers*
15 discontinued intervention (24%)
5 as at 3 months
5 due to family/work stress
1 disliked technology at free time
4 unknown reasons
2 unable to attend
At 9 months, n 45 completers*
16 discontinued intervention (26%)
15 as at 6 months + 1 felt unmotivated
1 unable to attend
At 12 months, n 40 completers‡
22 discontinued intervention (35%)
At 12 months, n 45 completers*
17 discontinued intervention (27 %)
16 as at 9 months + 1 unknown reasons
156 eligible via telephone screening invited for assessment
125 randomized
Fig. 2 Participant flow in the study (*three excluded from all analyses due to long-term medical problems (not reported at baseline);
†discontinue 5drop-out 5withdraw from the study; ‡one excluded from all analyses due to use of the commercial weight-loss
programme under study)
2384 I Haapala et al.

Owen et al.
(13)
and physical activity coefficients adapted
from Shetty et al.
(14)
. It was designed to discourage daily
energy intakes below 800 kcal, participation by children
(younger than 18 years of age) and participation by
anyone with a BMI below 18 kg/m
2
. After receiving
information on the dieter’s current weight, the pro-
gramme sent a text indicating the percentage dieters had
reached for the day’s target weight; the extent to which
they had reached their daily weight goal; the amount of
food to be consumed in proportion to the subject’s nor-
mal diet, as a fraction, percentage and as energy (3/4,
75 %, 1500 kcal); and the days remaining until the target.
The programme was based on text messages; there were
no phone calls made. The initiation was made by the
study participant who sent the first text message. All
messages sent by the study participant led to an auto-
matically generated, tailored response text message.
The programme advised the dieters to start reducing their
food intake by leaving out ‘unnecessary foods’ high in sugar
and/or fat and to cut down on alcohol. It encouraged an
increase in daily physical activity and emphasized the need
for regular weight reporting, either via text messaging or
through the programme’s password-protected website. The
website provided a personal (password-protected) web-
space for dietary record keeping and tracking one’s weight
loss in visual form. It also offered links to reliable sources
of information on healthy nutrition and physical activity.
TheprogrammeiscurrentlyavailableonlyinFinnishat
http://www.weightbalance.fi.
Dieters in the present study were allowed to set their
target weight either as a short- or a long-term goal and to
adjust it as needed at every 3-month visit. After the user
reached the set target weight, he/she could still use the
programme for weight-loss maintenance. As a rule, weight
loss was started at 2 kg/month. Those who wished to start at
a faster pace (the fastest being 4?8 kg/month) were closely
monitored via the web tracking system, which provided the
research team access to the dieter’s personal weight charts.
Data analysis
Repeated-measures ANOVA was used to test for changes in
dependent, normally distributed continuous variables over
time within and between groups. An intention-to-treat ana-
lysis was also run using baseline weight, or weight carried
forward from last observation if it was higher, for any missing
data for those who withdrew from the study. Bivariate
correlation and linear regression analyses were run to assess
the relationship between contact with the programme and
background, process and outcome variables. In keeping
with our theoretical model, we first evaluated how well the
background and process variables predicted average weekly
contact with the programme at 3 months. Variables with a
bivariate correlation to the criterion were entered stepwise
into the models in sets of background and process variables
using SPSS PC

R
for Windows statistical software package
release 10?0?5 (SPSS Inc., Chicago, IL, USA).
Results
Background characteristics and withdrawal
from the study
The experimental (n62) and control group (n62) did not
differ on any of the background characteristics or baseline
Table 1 Baseline characteristics by group: overweight healthy adult volunteers, Finland, June 2001 to June 2002
EG (n62) CG (n62)
Variable Mean or nSD or % Mean or nSD or %
Age (years) 38?14?738?04?7
Weight (kg) 87?512?686?412?5
BMI (kg/m
2
)30?62?730?42?8
Waist circumference (cm) 98?510?396?610?4
Sex
Females 49 79 47 76
Males 13 21 15 24
Education*
Vocational school 10 16 11 18
College degree/baccalaureate 44 71 31 50
Graduate degree 5 8 14 23
Marital status
Married/co-habiting 53 85 50 81
Like using mobile phones (yes)-56 90 52 85
Used mobile phones for .2 years (yes)-54 87 48 79
Like using the Internet (yes)*-60 97 49 80
Years of Internet use-
-
3?31?72?61?9
Have prior experience with dieting (yes)y54 87 50 81
Prefer dieting alone v. in a group (yes)y28 54 27 49
EG, experimental group; CG, control group.
*P.0?05, x
2
test between groups.
-Missing data for one subject in CG.
-
-
P.0?05, ttest between groups.
yMissing data for ten subjects in EG, seven in CG.
Mobile weight loss 2385

measurements (Table 1). In the experimental group,
subjects who withdrew (discontinued) from the study
(see Fig. 2) did not significantly differ on any background
variables but lost less weight by 3 months than those who
continued in the study (1?0(
SD 3?4) v.5?3(SD 3?5) %,
t53?7, P,0?0001). Completers of the 12 months and
those who discontinued the study and provided feedback
(n14) reported similar high grades for programme opera-
tion and usefulness: 7?3(
SD 1?2) v.7?9(SD 1?0), P,0?053
(on a scale from 4 to 10). Reasons for discontinuing in the
study included increased stress at work or studying (n3),
changes in personal life situation (n3), not feeling up to the
challenge alone (n2) and preferring to turn the mobile
phone and computer off after work (n1).
Programme effectiveness
Short- and long-term results
Repeated-measures ANOVA indicated a significant time
effect for weight loss across the 3-month intervals (F(4,38) 5
24?5, P,0?0001) and a significant time by group interaction
at 12 months in favour of the experimental group (by 4?1
(SD 1?4) %; F(1,80) 58?0, P50?006; Table 2). Most of the
weight loss in the experimental group took place during the
first three months (4?5(
SD 3?1) kg) while the cumulative
reduction was highest at 6 months (5?2(
SD 4?4) kg). By
12 months, the experimental group had lost 4?5(
SD 5?0) kg
(t55?8, P,0?0001) while the weight loss among the con-
trols was non-significant (1?1(
SD 5?8) kg; t51?2, P50?247;
Table 2). Adding participation in other weight-loss pro-
grammes as a cofactor in repeated-measures ANOVA had
no significant effect on the results in weight loss. In the
experimental group, 24 % (n10) of the subjects lost at least
10 % of their initial weight by 12 months, while 10 % (n4)
of controls succeeded in this. The percentage achieving at
least 5 % weight loss and keeping it off for 12 months was
45 % (n19) and 20 % (n8) in the experimental and control
group, respectively. The reduction in waist circumference
showed a similar pattern, with a significant reduction by 12
months in both groups that was greater in the experimental
group: 6?3(
SD 5?3) v.2?4(SD 5?4) cm (Table 2).
Intention-to-treat analyses indicated a significant
time effect (F(1,118) 518?8, P,0?0001) and time by
group interaction in favour of the experimental group
(F(1,118) 57?4, P50?008): reduction of 3?1(
SD 4?9) kg
(t54?9, P,0?0001) v.0?7(SD 4?7) kg (t51?2, P50?245).
Similarly, intention-to-treat analysis for the reduction in
waist circumference indicated a significant time effect
(F(1,18) 546?0, P,0?0001) and time by group interac-
tion in favour of the experimental group (F(1,118) 511?0,
P50?002): reduction of 4?5(
SD 5?3) cm (t56?5,
P50?0001) v.1?6(
SD 4?5) cm (t52?8, P50?008).
User satisfaction
The participants gave the programme a relatively high score
on a scale from 4 to 10: 7?8(
SD 0?8) at 12 months (n42).
Amount, frequency and type of use
Overall frequency of use of the programme faded from
8 times per week to 3–4 times per week by 12 months
Table 2 Outcome variables by group at 3-month intervals for completers of 12 months in both groups: overweight healthy adult volunteers,
Finland, June 2001 to June 2002
Baseline 3 months 6 months 9 months 12 months
Variable Mean SD Mean SD Mean SD Mean SD Mean SD
Body weight (kg)*
EG (n42) 86?612?782?012?981?413?681?813?882?114?1
CG (n40) 85?112?5––––––84?013?2
Percentage weight loss-
EG (n42) – – 5?33?56?15?15?65?65?45?8
CG (n40) ––––––––1?36?5
Waist circumference (cm)-
-
EG (n42) 97?610?591?710?490?911?291?111?691?311?7
CG (n40) 95?710?9––––––93?311?1
Self-efficacy in dietingy
EG (n40) 7?01?17?01?26?71?16?61?36?41?7
CG (n40) 7?01?0––––––6?61?4
Energy-dense food score||
EG (n41) 2?90?6– – 2?40?6– – 2?60?6
CG (n40) 2?70?7––––––2?60?7
EG, experimental group; CG, control group.
*Time effect: F(4,38) 524?5, P50?0001; time by group interaction: F(1,80) 58?0, P50?006. For EG, significant difference from baseline at each time point
(P,0?0001); for CG, non-significant change.
-Significant difference between groups at 12 months: t53?0, P50?003.
-
-
Time effect: F(4,38) 530?1, P50?0001; time by group interaction: F(1,80) 555?2, P50?0001. For EG, significant difference from baseline at each time point
(P,0?0001); for CG, significant change: t52?8, P50?0008.
yTrust in one’s capability of achieving the self-set goals for weight loss, reducing food intake, increasing physical activity and maintaining the weight loss on
10-point scales: 0 5‘I am not at all certain’ to 9 5‘I am absolutely certain’. Significant decrease for EG, Friedman test and Kendall’s W:x
2
510?2, P50?05.
Significant decrease in CG, Wilcoxon test: Z522?08, P50?04. In EG, significant change only between 3-month and 12-month scores: Z52?05, P50?05.
||Energy-dense foods score scale, consumption frequency for eight food items (internal consistency coefficient 50?71, n116): 1 5‘less than once per month or
never’, 2 5‘once or twice per month’, 3 5‘once per week’, 4 5‘once or twice per week’, 5 5‘once per day or more often’. Time effect: F(2,39) 527?6,
P50?0001; time by group interaction: F(1,80) 55?6, P50?03. For EG, significant difference from baseline at each time point (P,0?0001); for CG, non-
significant change.
2386 I Haapala et al.

(Table 3). Those with more than 5 % weight loss at
12 months reported more frequent weekly contact at
3 months than those who had lost less than 5 % (9?7(
SD 3?7)
v.7?0(SD 3?8) times; t52?31, P,0?05). Mobile phones were
the predominant medium for weight reporting and keeping
in contact with the programme (Table 3).
The four most useful features of the programme listed
at 3 months were: the use of mobile phones and the
Internet (93 % agreement), that the programme was free
of charge (93 %), regular reporting of weight (91 %) and
immediate feedback (90 %). By 12 months, setting short-
term goals was also reported as useful (95 %).
Within the experimental group, 56 % reported at 3
months that they had searched for more information on
healthy diet, while only 12 % had looked up information on
physical activity to support weight loss. At 12 months, 33 %
of the participants reported having searched for more
information on physical activity. In the control group, 33 %
reported having searched for more information on a heal-
thy diet and 20 % on physical activity. The most common
sources reported were the Internet, books, brochures and
old notes from previous weight-loss attempts.
Process variables
Dietary habits, nutritional intake and physical activity
At 3 months, 83 % of the completers reported having
made some improvements to their diet. Most common
changes were: reduced fat intake (48 %), reduced amount
of sugar and sweets (33 %) or of food overall (29 %) and
increased amounts of vegetables (17 %). The average
number of positive changes reported at 3 months was 1?6
(SD 1?1). Energy-dense food scores indicated a significant
change in consumption of this type of food (Table 2).
Nutritional intake analysis did not indicate significant a
change in average daily energy intake from 7297 (SD 1975)
kJ/d in the experimental group (n25) or 7263 (SD 1937)
kJ/d in the control group (n21). Physical activity
increased on average in both groups, from 2–3 times per
month to once per week (P,0?05).
Changes in self-efficacy and life situation
At 12 months, perception of dietary self-efficacy showed a
significant decrease from baseline (Table 2). However, it
increased from 6?7(
SD 1?4) (on a scale from 0 to 9, high
scores indicating stronger trust in one’s capabilities in diet-
ing) among those who had lost at least 5 % of initial weight
by 12 months (by 0?3(SD 1?2), P50?46), but decreased from
7?3(SD 0?8) (by 1?3(SD 1?9), P50?008) among those who
had gained weight or lost less than 5% by 12 months. Life
situation at home had changed during the 12 months for
fifteen subjects due to a move, major celebration, injury,
illness in the family or other family reasons; thirty-two
subjects reported having experienced such stress at work
that it had negatively impacted on their weight loss.
Predictors of weight loss
Correlation analyses between contact with the pro-
gramme, background, process and outcome variables
(see Fig. 1 for theoretical modelling) are presented for
short-term (3 months) and long-term (12 months) data in
Table 4. Because of the strong correlation between short-
and longer-term weight loss results, we concentrated our
analysis on finding the predictors of weight loss at 3 and
12 months (Table 5) in addition to those for programme
contact (see Table 5).
For ‘contact with the programme’, the most parsimo-
nious prediction equation consisted of one background
(liking teletechnology) and three process variables
(number of positive changes made in the diet, seeking
more information on nutrition and self-reported changes
at work), which accounted for a significant amount (41 %)
of the average contact variability (Table 5). The best
prediction equation for weight loss at 12 months con-
sisted of percentage weight loss at 3 months and change
in self-efficacy from baseline, accounting for 65 % of the
variance in the criterion (Table 5). For the 3-month weight
loss, background variables did not significantly add to the
prediction equation over and above the two process
variables (change in self-efficacy in dieting and seeking
information on nutrition at 3 months), amount of use of
the programme and the grades given to the programme.
This model explained 62 % of the variance in the criterion
(weight loss at 3 months; Table 5).
Discussion
The present study showed that mobile phone delivery
can be considered an effective method for supporting
Table 3 Frequency of self-reported weight reporting and contact with the programme for completers of 12 months (n40): overweight
healthy adult volunteers, Finland, June 2001 to June 2002
3 months 6 months 9 months 12 months
Variable Mean SD Mean SD Mean SD Mean SD
Reporting via mobile phone per week* 4?42?82?62?81?62?41?11?8
Reporting via Internet per week-1?92?11?92?51?42?21?22?0
Total number of contacts per week-
-
8?24?05?74?63?73?53?13?5
*Time effect: F(3,37) 518?7, P50?0001.
-Significant time effect: F(3,37) 516?5, P50?0001.
-
-
The total amount of contact 5weight reporting via mobile phone and Internet, plus other use of the programme website.
Mobile weight loss 2387

Table 4 Correlations* between percentage weight loss at 3 and 12 months and background and process variables for completers of 3 months and completers of 12 months: overweight healthy
adult volunteers, Finland, June 2001 to June 2002
Completers of 3 months Completers of 12 months
Percent weight loss at
Weekly contact Percent weight loss at 3 months Weekly contact during 12 months 3 months 12 months
rrnrnrrn
Background characteristics
Level of education 0?10 0?29 52 0?11 41 0?23 0?39 41
Liking teletechnology (1 5no, 2 5yes) 0?39 0?28 52 0?34 41 0?29 0?24 41
Energy-dense food score 0?02 20?24 52 20?19 41 20?31 20?43 41
Energy intake (kcal) at baseline 20?05 20?19 45 20?35 37 20?39 20?38 37
Process variables
Changes at work at the time point (1 5yes, 2 5no) 0?41 0?39 52 0?41-40 0?36 0?08 41
Changes at home at the time point (1 5yes, 2 5no) 0?26 0?20 52 0?41-40 0?32 0?16 41
Seeking for information on nutrition at 3 months 0?39 0?53 51 0?25 40 0?38 0?39 40
Number of self-reported changes in dietary habits at 3 months 0?43 0?51 52 0?21 41 0?44 0?11 41
Change in energy-dense food score at 12 months 20?07 n/a 0?17 41 0?16 0?03 41
Change in self-efficacy at the time point 0?29 0?51 52 0?15-
-
41 0?41 0?68 41
Weekly contact at the time point 1?00 0?62 52 0?35y41 0?57 0?38 41
Outcome variables
Grade for the programme at the time point 0?40 0?55 51 0?55|| 41 0?58 0?36 41
Percentage weight loss at 3 months 0?62 1?00 52 0?59 41 1?00 0?63 41
Percentage weight loss at 12 months 0?28 0?67 40 0?36 41 0?63 1?00 41
n/a, not applicable for the participants at 3 months.
*With n52, coefficients were significant at the P,0?05 level when r.0?28, at the P,0?01 level when r.0?39 and at the P,0?001 level when r.0?51. With n40, respectively when r.0?31, r.0?40 and r.0?54.
-Correlation between changes at work and home and contact at 3 months.
-
-
Correlation between the change in self-efficacy from baseline at either time point.
yCorrelation between percentage weight loss at 1 year and average weekly contact.
||Correlation between average 12-month grade and average weekly contact.
2388 I Haapala et al.

short- (3 months) and long-term (12 months) weight loss.
With mobile phone delivery, weight loss can be depen-
dent upon the amount and type of programme use and
learning (changes in health behaviour and self-efficacy)
taking place in that process and not solely on information
arising from the programme. Weight loss was positively
influenced by programme contact. Keeping in contact
with the programme, on the other hand, may be a process
influenced by social context, as indicated by the negative
impact of work stress on contact with the programme
under study.
Long-term (12-month) reductions in weight and waist
circumference in the experimental group were equal to or
greater than those previously reported from minimal-
contact or minimal-advice programmes
(15–19)
and from
more labour-intensive Internet-based programmes
(1,3–7)
.
Findings are also in line with a recent meta-analysis of US
studies, indicating that higher initial weight loss (.20 kg)
was associated with successful loss maintenance, and
only a 3 % reduction from the initial weight was main-
tained at 5 years after participation
(20)
.
The associations seen in our study between short-term
(3 months) weight loss and the amount of programme
use, satisfaction, change in self-efficacy and seeking
information on nutrition highlight the mediating role of
new technologies in supporting self-directed, active par-
ticipation in learning new health behaviours. Successful
self-directed learners are said to ‘show interest, personal
efficacy, enthusiasm, and even comfort in controlling
their own learning activities’
(21)
. Such active participation
in a weight-loss programme has been shown to predict
long-term weight-loss maintenance
(22,23)
. With the new
technology, attitudes towards teletechnology may serve
as strong predictors of health behaviour change because
they can influence the amount of programme use (con-
tact), as indicated in the present study.
In prior research, continued contact with the group,
therapist or the programme leader has been shown to be a
major predictor of success in weight loss
(4,24,25)
,andsuc-
cessful long-term weight-loss maintenance has been related
to frequent self-monitoring of body weight and food
intake
(26–28)
. In our study, the more frequent the reporting
of weight via text messaging had been, the more weight
was lost at 12 months as a percentage of initial weight. The
nature of this relationship is unclear. However, it is possible
to gain some insights from the participants’ responses. For
example, dieters seemed to appreciate the independence
of time and place in keeping in touch with the weight-loss
programme. The preferred medium for weight reporting in
the current study was the mobile phone (4?4(
SD 2?8) times
per week during the first three months). The Internet was
used once or twice weekly for weight reporting and other
programme activities, thus the average weekly contact with
the programme during the first twelve months was 5?0(
SD
3?2) times per week (median 4?6; range 0?2to12?7times
per week). This shows that the self-directed participation
Table 5 Multiple regression models for predicting contact, 3-month and 12-month weight loss: overweight healthy adult volunteers, Finland, June 2001 to June 2002
Variable Cumulative R
2
Bcoefficient sEE PUnivariate R
2
Bcoefficient SEE
Predicting weekly amount of contact at 3 months: R
2
50?41, R564, F(4,46) 58?0, P,0?0001
Number of positive self-reported changes in dietary habits 0?16 0?84 4?08 0?003 0?18 1?74 4?10
Baseline liking of the use of teletechnology (1 5no, 2 5yes) 0?25 27?27 3?90 0?020 0?15 28?82 4?18
Seeking more information on nutrition, reported at 3 months (15no, 2 5yes) 0?35 2?60 3?68 0?011 0?16 3?46 4?10
Changes at work by 3 months (1 5yes, 2 5no) 0?41 2?45 3?54 0?033 0?17 3?78 4?15
Predicting weight loss at 3 months: R
2
50?62, R579, F(4,46) 518?5, P,0?0001
Seeking more information on nutrition, reported at 3 months (15no, 2 5yes) 0?28 1?68 3?37 0?0001 0?28 4?11 3?37
Change in self-efficacy from baseline to 3 months (3 months minus baseline value) 0?42 0?77 3?04 0?001 0?26 1?35 3?41
Weekly amount of contact with the programme reported at 3 months 0?55 0?28 2?71 0?001 0?38 0?54 3?13
Grade for the programme at 3 months 0?62 1?06 2?53 0?007 0?31 2?03?30
Predicting weight loss at 12 months: R
2
50?65, R581, F(2,38) 535?2, P,0?0001
Change in self-efficacy from baseline to 12 months (12 months minus baseline value) 0?46 1?66 4?20 0?0001 0?46 2?13 4?20
Percentage weight loss from baseline 0?65 0?72 3?40 0?0001 0?40 1?04?50
Bcoefficient, unstandardized Bcoefficient; P, significance of contribution of each additional parameter to the stepwise multiple regression model; univariate R
2
, single variable entered into the prediction equation; SEE,
standard error of the estimate.
Mobile weight loss 2389

(contact) in the current study was more than double the
log-in frequency in a recent study of an Internet-only
group
(7)
and considerably higher than in previously
reported Internet-based programmes
(4,6)
.
Our finding that baseline self-efficacy was not as strong
a predictor as change during the programme indicated
that the programme may have succeeded in providing
positive opportunities for performance accomplishment,
verbal persuasion, vicarious performance and physiological
or affective arousal: the sources of self-efficacy identified
by Bandura
(11)
. Supported by previous research
(29,30)
,our
findings also suggest that baseline self-efficacy appraisals
may be overly optimistic and unconnected to practical skills
or opportunities to overcome possible obstacles encoun-
tered in personal life or at work. In our study, starting from
a lower level of baseline self-efficacy, those who achieved
and maintained at least 5% weight loss reported sig-
nificantly higher 12-month self-efficacy than those who lost
less than 5 % or ga ined weight by 12 months.
Our results suggest that life circumstances may have a
significant impact on willingness to persevere with the
programme. The amount of use of the programme was
negatively influenced by self-reported negative changes
at work at all time points, and this was, in fact, the main
reason reported (n6) for discontinuing in the study.
The present study examined two aspects of self-directed
behaviour: (i) the micro qualities of the medium through
which the communication of messages and feedback takes
place; and (ii) the macro qualities of the life-space of the
participant. Further research into the micro qualities of the
medium requires further longitudinal studies with qualita-
tive methodology to capture the participants’ perceptions
of the usefulness of programme features. Answers to
the macro qualities may require connecting weight-loss
programmes with life counselling, as also suggested by
others
(22,31)
.
As a limitation of our study, the 12-month results may
overestimate effects of the programme itself on weight
loss because of the 3-monthly in-person weigh-ins. How-
ever, even with these short visits to the study centre, the
programme remained a minimal-advice and minimally
labour-intensive intervention that could cost-effectively be
incorporated into everyday health-care practice. Another
limitation may be that the data regarding physical activity
and dietary habits were collected using a self-administered
questionnaire and thus constitute self-reported data. How-
ever, this is the usual method for nutritional research which
enables comparisons with prior research. Furthermore, at
the time of our study, we could not identify a previously
validated measure of energy-dense foods in the Finnish
population.
The important new approach of the present study is
that it combined existing behaviour change theory with a
theoretical model of adult learning and interaction with
new media. This allowed us to critically assess the power
of the feedback medium itself. Mobile phones have the
advantage of being spatially flexible and providing suc-
cinct messages, which are easily assimilated. Also, people
associate certain qualities to the use of mobile phones:
it is a medium for two-way communication with others
and is a self-adopted medium, recognizing the dieter as
an active participant.
In conclusion, the current results indicate that mobile
phone delivery, in conjunction with user-determined
Internet access, may be used to develop an effective
medium for short- and long-term weight loss.
Acknowledgements
Sources of funding: This research was partly funded by
GeraCap Invia Ltd, Seina
¨joki, Finland. Conflict of interest
declaration: A consultation fee was received by I.H. from
GeraCap Invia Ltd, producer of the Weight Balance

R
pro-
gramme. Following this research and since spring 2008, I.H.
has served on the executive board of the Weight Balance

R
programme. Authorship responsibilities: I.H. designed
the study, analysed and interpreted the results, and
drafted the manuscript. L.S. and P.M. participated in data
collection, while all authors participated in critical revi-
sion of the first and second versions of the manuscript.
Acknowledgements: We thank Pasi Juntunen, Marja
Leena Karhunen and Tiina Anttonen for their expertise in
the initial phases of data collection in this study.
References
1. Tsai AG & Wadden TA (2005) Systematic review: An
evaluation of major commercial weight loss programs in
the United States. Ann Intern Med 142, 6–66.
2. Rothert K, Strecher VJ, Doyle LA, Caplan WM, Joyce JS,
Jimison HB, Karm LM, Mims AD & Roth MA (2006)
Web-based weight management programs in an integrated
health care setting: a randomized, controlled trial. Obesity
14, 266–272.
3. Gold BC, Burke S, Pintauro S, Buzzell P & Harvey-Berino J
(2007) Weight loss on the web: a pilot study comparing
a structured behavioral intervention to a commercial
program. Obesity 15, 155–164.
4. Tate DF, Wing RR & Winett RA (2001) Using Internet
technology to deliver a behavioral weight loss program.
JAMA 285, 1172–1177.
5. Harvey-Berino J, Pintauro S, Buzzell P et al. (2002) Does
using the Internet facilitate the maintenance of weight loss?
Int J Obes Res 26, 1254–1260.
6. Womble LG, Wadden TA, McGuckin BG, Sargent SL,
Rothman RA & Krauthamer-Ewing ES (2004) A randomized
controlled trial of a commercial internet weight loss
program. Obes Res 12, 1011–1018.
7. Tate DF, Jackvony EH & Wing RR (2006) A randomized trail
comparing human e-mail counseling, computer-automated
tailored counseling, and no counseling in an internet
weight loss program. Arch Intern Med 166, 1620–1625.
8. National Heart, Lung, and Blood Institute & North
American Association for the Study of Obesity (2000) The
Practical Guide to the Identification, Evaluation, and
Treatment of Overweight and Obesity in Adults. Bethesda,
MD: National Institutes of Health.
2390 I Haapala et al.

9. World Health Organization (2000) Obesity: Preventing and
Managing the Global Epidemic. WHO Technical Report
Series no. 894. Geneva: WHO.
10. Hiltz SR (1988) Productivity enhancement from computer-
mediated communication: a systems contingency approach.
Commun ACM 31, 1438–1454.
11. Bandura A (1997) Self-efficacy. The Exercise of Control.
New York: W.H. Freeman and Company.
12. Lohman TG, Roche AF & Martorell R (1991) Anthropo-
metric Standardization Reference Manual, abridged edi-
tion. Champaign, IL: Human Kinetics Books.
13. Owen OE, Holup JL, D’Alessio DA et al. (1987) A
reappraisal of the caloric requirements of men. Am J Clin
Nutr 46, 875–885.
14. Shetty PS, Henry CJ, Black AE & Prentice AM (1996) Energy
requirements of adults: an update on basal metabolic rates
(BMRs) and physical activity levels (PALs). Eur J Clin Nutr
50, Suppl., S11–S23.
15. Cameron R, MacDonald MA, Schlegel RP, Young CI, Fisher
SE, Killen JD, Rogers T, Horlick L & Shepel LF (1990)
Toward the development of self-help health behaviour
change programs: weight loss by correspondence. Can J
Public Health 81, 275–279.
16. Miller WC, Eggert KE, Wallace JP, Lindeman AK &
Jastremski C (1993) Successful weight loss in a self-taught,
self-administered program. Int J Sports Med 14, 401–405.
17. Hellerstedt WL & Jeffery RW (1997) The effects of a
telephone-based intervention on weight loss. Am J Health
Promot 11, 177–182.
18. Latner JD, Wilson GT, Stunkard AJ & Jackson ML (2002)
Self-help and long-term behavior therapy for obesity.
Behav Res Ther 40, 805–812.
19. Heshka S, Anderson JW, Atkinson RL, Greenway FL, Hill
JO, Phinney SD, Kolotkin RL, Miller-Kovach K & Pi-Sunyer
FX (2003) Weight loss with self-help compared with a
structured commercial program: a randomized trial. JAMA
289, 1792–1798.
20. Anderson JW, Konz EC, Frederich RC & Wood CL (2001)
Long-term weight-loss maintenance: a meta-analysis of US
studies. Am J Clin Nutr 74, 579–584.
21. Hiemstra R (2006) Has the Internet Changed Self-Directed
Learning? Paper presented at the 20th International Self-
Directed Learning Symposium, Cocoa Beach, FL, 8–11
February 2006. http://www-distance.syr.edu/InternetandSDL.
html (accessed October 2007).
22. Jeffery RW, Drewnowski A, Epstein LH, Stunkard AJ,
Wilson GT, Wing RR & Hill DR (2000) Long-term
maintenance of weight loss: current status. Health Psychol
19, Suppl., 5S–16S.
23. Teixeira PJ, Going SB, Sardinha LB & Lohman TG (2005) A
review of psychosocial pre-treatment predictors of weight
control. Obes Rev 6, 43–65.
24. Perri MG, McAllister DA, Gange JJ, Jordan RC, McAdoo G &
Nezu AM (1988) Effects of four maintenance programs on
the long-term management of obesity. J Consult Clin
Psychol 56, 529–534.
25. Wadden TA, Butryn ML & Byrne KJ (2004) Efficacy of
lifestyle modification for long-term weight control. Obes
Res 12, Suppl., 151S–162S.
26. Wing RR & Hill JO (2001) Successful weight loss
maintenance. Annu Rev Nutr 21, 323–341.
27. Linde JA, Jeffery RW, French SA, Pronk NP & Boyle RG
(2005) Self-weighing in weight gain prevention and weight
loss trials. Ann Behav Med 30, 210–216.
28. White MA, Martin PD, Newton RL, Walden HM,
York-Crowe EE, Gordon ST, Ryan DH & Williamson DA
(2004) Mediators of weight loss in a family-based
intervention presented over the Internet. Obes Res 12,
1050–1059.
29. Maibach E, Flora JA & Nass C (1991) Changes in self-
efficacy and health behavior in response to a minimal
contact community health campaign. Health Commun 3,
1–15.
30. Palmeira AL, Teixeira PJ, Branco TL, Martins SS, Minderico
CS, Barata JT, Serpa SO & Sardinha LB (2007) Predicting
short-term weight loss using four leading health behavior
change theories. Int J Behav Nutr Phys Act 4, 14.
31. Wing RR, Tate DF, Gorin AA, Raynor HA & Fava JL (2006) A
self-regulation program for maintenance of weight loss. N
Engl J Med 355, 1563–1571.
Mobile weight loss 2391