Behavioral Weight Control for Overweight
Adolescents Initiated in Primary Care
Brian E. Saelens,*† James F. Sallis,† Denise E. Wilfley,† Kevin Patrick,‡ John A. Cella,§
and Richard Buchta¶
SAELENS, BRIAN E., JAMES F. SALLIS, DENISE E.
WILFLEY, KEVIN PATRICK, JOHN A. CELLA, AND
RICHARD BUCHTA. Behavioral weight control for
overweight adolescents initiated in primary care. Obes Res.
Objective: This study evaluates the post-treatment and
short-term follow-up efficacy of, as well as participant
satisfaction for, a 4-month behavioral weight control pro-
gram for overweight adolescents initiated in a primary care
setting and extended through telephone and mail contact.
Research Methods and Procedures: 44 overweight adoles-
cents were randomly assigned to either a multiple compo-
nent behavioral weight control intervention (Healthy Habits
[HH]; n ? 23) or a single session of physician weight
counseling (typical care [TC]; n ? 21). Weight, height,
dietary intake, physical activity, sedentary behavior, and
problematic weight-related and eating behaviors and beliefs
were assessed before treatment, after the 4-month treatment,
and at 3-month follow-up. Participant satisfaction and be-
havioral skills use were measured.
Results: HH adolescents evidenced better change in body
mass index z scores to post-treatment than TC adolescents.
Body mass index z scores changed similarly in the condi-
tions from post-treatment through follow-up. Behavioral
skills use was higher among HH than TC adolescents, and
higher behavioral skills use was related to better weight
outcome. Energy intake, percentage of calories from fat,
physical activity, sedentary behavior, and problematic
weight-related or eating behaviors/beliefs did not differ by
condition or significantly change over time independent of
condition. The behavioral intervention evidenced good fea-
sibility and participant satisfaction.
Discussion: A telephone- and mail-based behavioral inter-
vention initiated in primary care resulted in better weight
control efficacy relative to care typically provided to over-
weight adolescents. Innovative and efficacious weight con-
trol intervention delivery approaches could decrease pro-
vider and participant burden and improve dissemination to
the increasing population of overweight youth.
Key words: adolescence, telephone, mail, intervention,
The prevalence of obesity in young childhood through
adolescence continues to increase in the United States (1).
Population trends suggest that overweight children are
heavier than overweight youth in previous decades (2). Such
trends are disconcerting given the psychosocial and physical
health risk associated with being overweight in childhood
(3). Long-term health risk seems greater for overweight
adolescents than for younger overweight children, indepen-
dent of adult weight (4). Obese adolescents are also more
likely to track obesity into adulthood than younger chil-
dren (5). These factors make adolescent weight control
a high priority (6).
Traditional weight control clinic-based interventions for
adolescent obesity have demonstrated some success (7).
However, the paucity of obesity treatment research for
overweight adolescents is remarkable relative to the quan-
tity of treatment research directed at overweight younger
children and adults (8–10). Adolescent treatment research is
especially limited regarding the format of treatment deliv-
ery, which has included only more costly and labor-inten-
sive hospital in-patient or weight control clinic-based for-
mats (9). Adult weight-loss and weight-gain prevention
programs have extended into applications without in-person
Submitted for publication May 3, 2001.
Accepted for publication in final form September 25, 2001.
*Department of Pediatrics, Division of Psychology, Children’s Hospital Medical Center,
Cincinnati, Ohio; †Department of Psychology, San Diego State University, San Diego,
California; ‡Graduate School of Public Health and Student Health Services, San Diego State
University, San Diego, California; §Department of Pediatrics, Southern California Kaiser
Permanente, La Mesa, California; and the ¶Department of Pediatrics, Scripps Clinic, La
Address correspondence to Dr. Brian E. Saelens, Children’s Hospital Medical Center,
Division of Psychology, 3333 Burnet Avenue, Cincinnati, OH 45229.
Copyright © 2002 NAASO
22 OBESITY RESEARCH Vol. 10 No. 1 January 2002
contact (e.g., telephone, mail, and Internet) (11,12), while
still maintaining primarily behavioral and cognitive–behav-
ioral components. Clinic-based family behavioral weight
control interventions for younger children have strong evi-
dence of long-term efficacy (8), but have generally not
tested innovative delivery formats that would increase cost
effectiveness and be more easily disseminated.
Pediatric primary care clinics are ideal settings for iden-
tifying overweight adolescents and initiating delivery of
more readily disseminated approaches for adolescent weight
control. The majority of adolescents visit a physician in any
given year (13,14). Overweight youth may be even more
likely to visit their pediatrician than non-overweight chil-
dren (15). Although health care providers sometimes fail to
capitalize on opportunities to address weight-related issues
with overweight patients (16), adolescents perceive their
providers as a valuable source of such information (13,17).
Indeed, adults who receive weight control advice from
physicians are more likely to attempt weight loss (18).
Health care providers seem amenable to changing current
practices to increase and improve their weight-related coun-
seling (19). Professional-based recommendations and in-
volvement in adolescents’ weight control efforts may also
decrease the likelihood that overweight youth will engage in
problematic weight-related and eating behaviors (e.g., skip-
ping meals, and purging). This is particularly relevant for
the overweight child population, who could be at increased
risk of developing disordered eating (20).
The present study was designed to evaluate the accept-
ability and efficacy of a multi-component behavioral inter-
vention for weight control among overweight adolescents.
The intervention includes computer interaction and physi-
cian counseling in the pediatric primary care clinic, fol-
lowed by 4 months of telephone- and mail-based behavioral
counseling. It was hypothesized that adolescents receiving
this intervention would have better weight outcomes than
adolescents receiving single-session physician counseling
(typical care [TC]) by the end of the behavioral intervention
time period and at the follow-up 3 months after treatment
cessation. It was further hypothesized that intervention ad-
olescents would have improved weight control behaviors
(e.g., more physical activity) and more behavioral skills use
for weight control relative to TC adolescents. It was pro-
posed that neither the behavioral intervention nor the TC
approach would result in increased problematic eating- or
weight-related behaviors or beliefs, given both conditions
involved health professional contact.
Research Methods and Procedures
Adolescents were recruited over a 5-month period from
two pediatric primary care clinics in southern California.
Participants were recruited from waiting room flyers and by
encouraging pediatricians in the clinics to discuss possible
study participation with seemingly eligible adolescents.
Fifty-nine interested participants were encouraged to sched-
ule an appointment at one of the clinics for baseline assess-
ment (see Figure 1 for participant flow). Only gender and
age information were collected before baseline assessment.
There was no significant difference in gender distribution or
age between those individuals completing (n ? 47) and not
completing baseline assessment (n ? 12).
Inclusion/exclusion criteria were being between 12 and
16 years old, 20% to 100% above the median (50th percen-
tile) for body mass index (BMI) for sex and age (21),
interested in weight control, but not currently engaged in
another weight control program, and otherwise healthy as
determined by a pediatrician. Randomized participants were
on average 14.2 years old (SD ? 1.2) with a BMI of 30.7
kg/m2(SD ? 3.1), 59.1% were boys (26/44), and self-
identified as 70.5% white, 15.9% Hispanic, 4.5% African
American, 2.3% Asian, and 6.8% multi-ethnic. All partici-
pants had a BMI above the 89th percentile for their respec-
tive age and gender. The parents of participants reported a
median household income of $60,000 to $69,000, with 24%
Figure 1: Participant flow chart representing recruitment, random-
ization, and retention of cohort of the Healthy Habits (HH) inter-
vention and typical care (TC) participants. *One TC adolescent
completed follow-up measurement of height and weight but did
not complete follow-up measures of secondary outcomes (e.g.,
Weight Control for Adolescents, Saelens et al.
OBESITY RESEARCH Vol. 10 No. 1 January 2002 23
of adolescents living in single-parent homes. Parental con-
sent and adolescent assent were obtained. Study procedures
were approved by the San Diego State University Commit-
tee on the Protection of Human Subjects and the Institu-
tional Review Boards of each of the clinics.
Sample size determination was based on moderate to
large short-term differences found between previous com-
prehensive weight control intervention and nonspecific
treatment for childhood obesity (22) and an interest in
pursuing novel adolescent obesity interventions (e.g., tele-
phone-based) only if adequate intervention potency could
be demonstrated relative to TC. With large effect size esti-
mates of f ? 0.40 (23), ?21 participants per condition
would provide adequate power (?0.80) to detect post-treat-
ment condition differences at p ? 0.05.
Three interested adolescents did not meet BMI inclusion
criteria, leaving 44 adolescents to be randomized (Figure 1).
After baseline assessment, eligible adolescents were strati-
fied by sex and level of percent overweight (low, 20% to
40%; moderate, 41% to 60%; or high, 61% to 100%) and
randomly assigned to the intervention (Healthy Habits
[HH]) or TC condition. Randomization occurred by selec-
tion among opaque envelopes labeled with levels of percent
overweight, each envelope containing an HH or TC card.
Baseline assessments were conducted at the pediatric clinic
before the condition-specific physician counseling. Post-
treatment (median ? 4.1 months after clinic visit) and
follow-up (actual median ? 7.2 months after pediatric
clinic visit) assessments occurred at a university-based
weight control clinic.
HH Intervention. The newly developed HH intervention
included various delivery formats. Formats, particularly
computer and telephone, were selected given their appeal to
adolescents, and materials and procedures were develop-
mentally tailored (Table 1).
Immediately after baseline assessment in the clinic, HH
adolescents engaged in a computer program adapted from
PACE? (Patient-Centered Assessment and Counseling for
Exercise plus Nutrition) software designed for adolescents
(24) and modified for overweight adolescents. The com-
puter program assessed eating, physical activity, and sed-
entary behavior and guided adolescents through individual-
ized plans generated to increase physical activity or
decrease sedentary behavior and decrease dietary fat or
increase fruits/vegetables or decrease overeating/snacking.
Plan generation included identifying benefits, barriers, and
specific strategies to achieve goals. The program generated
printed action plan summaries and produced a provider
summary form that helped identify for pediatricians which
behaviors were targeted by the adolescent and whether
problematic eating behaviors were reported (e.g., taking
laxatives or other purging). HH adolescents then met with a
pediatrician to discuss and finalize their individualized ac-
tion plans. This tailored physician counseling was based on
the computer responses of the adolescents.
Approximately 1 week after the clinic visit, each HH
adolescent and his/her parent met in-person with the first
author (B.E.S.) to discuss upcoming mail and phone con-
tacts and to learn food self-monitoring. Calls from a phone
counselor began 1 week after this meeting. Telephone coun-
Table 1. Brief descriptions of the behavioral skills and developmental tailoring of the Healthy Habits telephone
and mail contacts
Self-monitoring: writing down specific foods, amounts, calories, and each food’s category (GREEN, RED, no color)
daily; weighing weekly; writing down type and amount of physical activity
Goal setting: weekly establishment of goals for calories, GREEN and RED foods, and amount of physical activity
Problem solving: identifying specific problems/barriers that prevent goal attainment; brainstorming solutions and ways
to implement modified plan
Stimulus control: modifying the food and physical activity environment to make healthful choices more available and
less healthful choices more difficult to obtain
Self-reward: rewarding self for reaching goals
Preplanning: establishing plans for high-risk situations (e.g., parties) to decrease likelihood of unhealthful eating or lack
of physical activity
11-module illustrated manual detailing behavioral skills, written at the 7th to 9th grade reading level
Telephone contact only with adolescent and infrequent mail-only contact with parent(s) to promote autonomy
Lottery reward system for goal achievement provided by study staff rather than parent
Encouragement to have adolescents ask others (e.g., parents) for help to meet eating and physical activity goals
Weight Control for Adolescents, Saelens et al.
24 OBESITY RESEARCH Vol. 10 No. 1 January 2002
selors had at least a bachelor’s degree in psychology or
nutrition and received weekly supervision in the provision
of behavioral weight control treatment by the first author.
Calls were intermittently monitored for compliance with
call scripts. For each HH adolescent, the telephone coun-
selor remained constant throughout intervention. Telephone
contact was structured to last 10 to 20 minutes and sched-
uled weekly for the first eight calls and biweekly for the last
three calls, thus, lasting a total of 14 to 16 weeks. Telephone
counselors used detailed telephone scripts to address ado-
lescents’ weight change since the last call (adolescents were
encouraged to weigh once weekly), the link between weight
change and eating and physical activity behaviors, instruc-
tion and feedback on previous self-monitoring, eating and
physical activity goals, and the use of behavioral skills
relevant to goal achievement. Behavioral skills examples
are provided in Table 1. A participant manual designed to
help adolescents acquire various behavioral skills for weight
control was developed and distributed to HH adolescents.
Initial manual sections were provided at the in-person meet-
ing with the first author, with subsequent sections mailed
after the 5th, 8th, and 10th calls. Counselors referred ado-
lescents to the computer printouts generated in clinic and
relevant manual sections to help adolescents formulate strat-
egies for meeting goals.
HH adolescents were encouraged to self-monitor all food
and beverage intake and the amount and calories consumed.
Adolescents were given self-monitoring booklets, one
booklet for each week, to record daily: time foods/beverages
consumed, description of foods/beverages, food/beverage
amounts, calories, and whether a food/beverage was a RED
or GREEN food (described below). Stamped envelopes
were provided with each self-monitoring booklet, so ado-
lescents could mail booklets to their telephone counselor
after the call in which the self-monitoring was reviewed.
Calories were estimated using The Fat Counter (25), which
was provided. Telephone counselors helped HH adolescents
gradually reduce calories from baseline levels to ?1200 to
1500 kcal/d, although this goal was flexible upward depen-
dent on initial weight. Adapted from Epstein and Squires’
Stoplight Diet (26), foods were also categorized into
GREEN, RED, or no color foods. GREEN foods were
defined as having 1 or fewer fat grams per serving, ?150
calories per serving, and providing a good source of one or
more valuable dietary components (e.g., calcium, fiber, or
protein). RED foods were defined as having 5 or more
grams of fat per serving or were diet versions of high-fat
foods. The eventual GREEN food goal was 40 GREEN food
servings (based on standard serving sizes) or more per
week, and the RED food goal was ?15 RED food servings
per week. Telephone counselors did not encourage or dis-
courage eating any prescribed foods but encouraged reduc-
tion in food quantity and more healthful eating, within
established eating preferences and food availability. For
example, if an adolescent liked eating tacos, he/she was not
discouraged from eating tacos, but rather encouraged to eat
fewer tacos and to reduce or eliminate the high-fat high-
calorie food items often in tacos (e.g., fried tortilla, sour
cream, and guacamole). This procedure allowed for flexi-
bility in addressing adolescents’ individual food preferences
within a given adolescent’s family and cultural context.
Telephone counselors provided calorie and macronutrient
information for foods not listed in The Fat Counter (e.g.,
Beginning at the fifth call, HH adolescents were encour-
aged to self-monitor physical activity daily. The physical
activity goal was a minimum of 60 minutes of at least
moderate intensity physical activity on 5 days per week
(27), with gradual increases from baseline levels. Telephone
counselors encouraged adolescents to increase time in pre-
ferred physical activities and to add new types of physical
activity. Conversely, adolescents were encouraged to de-
crease time spent in least preferred sedentary behaviors and
to reallocate that time to being more active.
Adolescents were awarded 1 point each for meeting self-
monitoring, physical activity, calorie, GREEN food, and
RED food goals each week (total maximum of 5 points/wk),
based solely on the counselor’s review of self-monitoring
booklets. Points were accumulated for tickets for a study-
based lottery (1 point ? 1 ticket). The lottery occurred after
all adolescents had completed post-treatment assessment,
and the adolescent with the randomly selected ticket re-
Parents of HH adolescents were sent information sheets
when adolescents were sent manual sections. Parent infor-
mation sheets highlighted ways parents could be most help-
ful with their adolescents’ behavior change. Recommended
parental skills included stimulus/environmental control,
positive reinforcement, and preplanning. There was no tele-
phone contact with parents during the HH intervention.
TC Intervention. Immediately after baseline assessment,
TC adolescents met with a pediatrician. Based on expert
committee recommendations regarding pediatric obesity
that were given to pediatricians (28), pediatricians were
instructed to assess/encourage adolescent’s motivation for
weight-related behavior change, provide information about
short- and long-term health consequences of high weight
status and benefits of better weight control, make recom-
mendations for healthful eating consistent with the Food
Guide Pyramid (29), review physical activity recommenda-
tions for adolescents (60 min/d of at least moderate intensity
physical activity) (27), and encourage consistency and per-
sistence with health behavior changes. Pediatricians used a
worksheet outlining these topics to facilitate thorough dis-
cussion with each TC adolescent. TC adolescents were
encouraged to implement recommended behavior changes
on their own and with the help of their family. After this
non-tailored physician-counseling session, TC adolescents
Weight Control for Adolescents, Saelens et al.
OBESITY RESEARCH Vol. 10 No. 1 January 2002 25
were not contacted again until scheduling for the post-
treatment assessment ?4 months later.
The same pediatricians provided counseling for HH and
TC adolescents. Pediatricians had participated in previous
intervention studies for dietary and physical activity change
among non-overweight adolescents (24). Pediatricians met
with the first author (B.E.S.) once before the study begin-
ning to review the training and procedural materials.
After completing post-treatment assessment, neither HH
nor TC adolescents were contacted until scheduling for
follow-up assessment. Adolescents received $25 for post-
treatment and $25 for follow-up assessment.
Measures were obtained at all assessment time-points
unless otherwise noted.
Weight and Height. Weight was measured at baseline in
the pediatric clinic using a calibrated standard digital scale.
Weight was measured at post-treatment and follow-up on a
calibrated balance beam scale. Height was measured by
stadiometer at all assessments. BMI was calculated as kilo-
grams per square meters. For purposes of determining study
inclusion, population data from Rosner and colleagues (21)
were used to calculate percent overweight. Updated national
norms published during the course of this study were used
to calculate the BMI z scores and percentage of overweight
presented and used in data analysis (30). BMI z scores were
calculated using age- (to the nearest month) and sex-specific
median, SD, and power of the Box-Cox transformation (30).
Dietary Intake. Dietary intake was assessed by the 2-day
dietary recall interview. Among youth, the recall procedure
has evidenced high levels of between-interviewer reliability
for total calories recalled (coefficient of variation ? 17%)
(31), and moderate correlations between observed and re-
called total calories consumed (0.47 to 0.57) (32,33), with
similar associations for calories from fat (33). Foods and
beverages were entered into Nutritionist-V software (34) to
determine total average daily energy intake and percentage
of calories from dietary fat. Consistency estimates across
the two recalled days were similar at different assessment
time-points for calories (0.59 to 0.65) and the percentage of
dietary fat (0.31 to 0.51).
Physical Activity. Physical activity was assessed by the
Seven-Day Physical Activity Recall (PAR) interview.
1-week test–retest reliability for total physical activity esti-
mates obtained by PAR are 0.47 to 0.59 for adolescent
samples, with correlations of 0.44 to 0.53 between PAR
physical activity estimates and heart rate monitoring (35)
and similar correlations between PAR and accelerometer
estimates of physical activity among adults (36). Standard-
ized scoring procedures (37) were used to estimate daily
physical activity-related energy expenditure, independent of
body weight (kcal/kg per day), that was at least moderate in
intensity (i.e., energy expenditure by sleeping and light
activity was not included).
Sedentary Behavior. Sedentary behavior was measured
through self-report questionnaire that queried participation
in various discretionary sedentary behaviors (e.g., television
watching, but not homework) during the past 7 days. Ado-
lescents indicated the number of days engaged in and
amount of time typically spent in each sedentary behavior
over the past week. This methodology is similar to that used
to assess television watching (38) but to date has unknown
reliability and validity. Mean daily time in individual sed-
entary behaviors was calculated by multiplying the number
of days by the typical time spent doing that sedentary
behavior and then dividing by seven. Total daily sedentary
behavior was the sum of these daily individual sedentary
Problematic Eating and Weight-Related Behaviors and
Beliefs. Cognitive dietary restraint and eating disinhibition
were assessed by the Three-Factor Eating Questionnaire
(39). The restraint and disinhibition subscales have demon-
strated internal consistency ?0.79 across adult dieters and
free eaters (39) and were 0.81 to 0.87 for restraint and 0.63
to 0.75 for disinhibition at different assessment time-points
in this study. The total score from the Children’s Eating
Attitude Test (26-item CHEAT) was used to obtain a con-
tinuous measure of eating disorder psychopathology. The
CHEAT has previous internal consistency coefficients be-
tween 0.76 and 0.87 (40,41) and current study internal
consistency of 0.70 to 0.81 at different assessment time-
points, a previous test–retest estimate of 0.81 (40), and
seems related to body dissatisfaction and problematic
weight management behaviors (41). The Killen Weight
Concerns scale (42), with present internal consistency of
0.75 to 0.79 at different assessment time-points, assessed
concern about weight and weight change and is related to
the development of eating disorders (42,43).
Physician Counseling, Behavioral Skills Use, and Par-
ticipant Satisfaction. After physician counseling, adoles-
cents and physicians rated the perceived efficacy, specific-
ity, and the extent of tailoring of the physician’s counseling
using 1- to 7-point Likert scale items (higher ratings indi-
cating more). Adolescents and physicians also reported
length of physician counseling. At post-treatment and fol-
low-up, using Likert scales (1 ? never to 5 ? very often),
adolescents and their parents rated the frequency of adoles-
cents’ behavioral skills use (e.g., self-monitoring and stim-
ulus control). Skills use was asked separately for eating and
physical activity/sedentary behavior. Only HH adolescents
rated satisfaction for intervention components (computer
program, physician counseling, manual and other written
materials, and telephone counseling) at post-treatment, us-
ing Likert scale items (1 ? not at all to 5 ? very much) that
measured helpfulness, perceived satisfaction, perceived im-
pact on weight-related behaviors, and overall appeal.
Weight Control for Adolescents, Saelens et al.
26OBESITY RESEARCH Vol. 10 No. 1 January 2002
Demographics. Parents self-reported level of house-
hold income and adolescents self-reported birth date,
gender/sex, and ethnicity.
Data were screened for normality, with sedentary be-
havior time requiring logarithmic transformation. To test
for baseline condition differences, ?2analyses were used
for dichotomous variables (e.g., ethnicity) and one-way
ANOVAs for continuous variables (e.g., BMI z score). BMI
z score was considered the primary outcome measure so that
comparisons could be made among adolescents across age,
gender, and over time. To test for site and sex effects,
two-way ANOVAs examined condition by clinic site and
condition by sex interactions at baseline and repeated mea-
sures three-way ANOVAs tested corresponding interactions
with time (baseline, post-treatment, and follow-up). For
main analyses, repeated measures ANOVA were used to
assess change over time among primary (BMI z scores) and
secondary outcomes (e.g., physical activity), with condition
as the between-subjects factor and time as the within-sub-
jects factors. Analyses were conducted separately for base-
line to post-treatment and for baseline to post-treatment to
follow-up. Completers were considered those adolescents
who completed assessments (e.g., regardless of amount of
phone contact completed among HH adolescents). Linear
contrasts allowed for post-treatment to follow-up and base-
line to follow-up comparisons. Conservative intent-to-treat
analyses on the primary outcome were conducted by replac-
ing missing values of HH adolescents at post-treatment
(N ? 3) and follow-up (N ? 5) with the mean change of
the TC condition from the baseline to post-treatment and
post-treatment to follow-up, respectively. Mean change in
the TC condition also replaced missing TC participant data
(N ? 2). ?2analyses were used to test the frequency of
increases vs. decreases in BMI z scores from baseline to
post-treatment and follow-up. One-way ANOVAs were
used to test condition differences in physician counseling
characteristics and behavioral skills use and paired t tests
allowed for comparison of satisfaction among different
HH intervention components. Exploratory bivariate corre-
lations were used to examine relations between process
variables (e.g., counseling characteristics, skills use) and
post-treatment BMI z score, after partialing out baseline
BMI z score. Statistical significance was set at p ? 0.05 and
all tests were two-tailed.
There were no significant differences by condition on
demographic variables or baseline weight status variables
(e.g., BMI z-scores), physical activity, dietary intake, or
problematic eating and weight-related behaviors or be-
liefs (Table 2). In addition, there were no significant site
or sex by condition effects or three-way interactions with
time on BMI z scores, so data were collapsed across site
and sex for all analyses.
There was a significant group by time interaction from
baseline to post-treatment for BMI z scores among post-
treatment completers (F(1,37)? 6.04, p ? 0.02; effect size
f ? 0.40). As seen in Figure 2, mean BMI z scores signif-
icantly increased among TC adolescents compared with the
slight decrease of BMI z scores among HH adolescents
during the intervention period. Intent-to-treat analyses did
not markedly alter these results (F(1,42)? 5.59, p ? 0.03).
More HH adolescents had reduced their BMI z score by
post-treatment than TC adolescents (40.0% vs. 10.5%, re-
weight status outcomes, there were no significant differen-
tial changes by condition from baseline to post-treatment or
main effects of time in the secondary outcomes of total
energy or dietary fat intake, physical activity, sedentary
behavior, or problematic eating- and weight-related behav-
iors or beliefs (Table 2).
With the inclusion of the follow-up assessment, the over-
all condition by time interaction for BMI z scores remained
statistically significant (F(2,70)? 4.08, p ? 0.03, effect size
f ? 0.35; Figure 2). However, the baseline to follow-up
contrast for BMI z scores only approached statistical signif-
icance (F(1,35)? 3.50, p ? 0.070, effect size f ? 0.32).
Linear contrasts suggested no differential change in BMI z
scores by condition from post-treatment to follow-up, with
mean BMI z scores remaining generally stable from post-
treatment to follow-up in both conditions. Intent-to-treat
analyses only slightly attenuated the overall follow-up re-
sults (F(2,84)? 3.60, p ? 0.04) and the baseline to follow-up
contrast (F(1,42)? 3.11, p ? 0.09). From baseline to follow-
up, more HH adolescents had decreased BMI z score from
baseline values than TC adolescents (55.6% vs. 15.8%; ?2
? 6.41, p ? 0.02). Again, there were no significant inter-
actions of condition by time or main effects of time for any
secondary outcomes from baseline to post-treatment to fol-
low-up (Table 2).
The HH and TC condition did not significantly differ in
the amount of time adolescents (3.6 vs. 4.9 minutes, respec-
tively) or pediatricians (6.3 vs. 8.1 minutes, respectively)
perceived the physician counseling session lasted (both
F(1,44)? 4.0, p ? 0.05). There were no significant condition
differences in how adolescents and physicians perceived the
effectiveness, specificity, or extent of tailoring of the phy-
sician counseling. At post-treatment, HH adolescents re-
ported higher rates of total and eating- and physical activity-
specific behavioral skills use than TC adolescents (all F(1,37)
? 5.17, p ? 0.03). Parents of HH adolescents also reported
that their adolescents used more overall and specifically
eating-related behavioral skills than parents of TC adoles-
cents (both F(1,37)? 4.50, p ? 0.04). HH adolescents
continued to report higher overall and eating-related behav-
(1)? 4.44, p ? 0.04). Despite differential
Weight Control for Adolescents, Saelens et al.
OBESITY RESEARCH Vol. 10 No. 1 January 200227
ioral skills use at the follow-up assessment compared with TC
adolescents (both F(1,33)? 7.88, p ? 0.01), but parent reports
of adolescent behavioral skills use were no longer significantly
different between HH and TC conditions at follow-up.
The median number of HH intervention calls completed
was 9.0 of the planned 11 calls among all adolescents
randomized to HH, with calls lasting on average 16.4 min-
utes (SD ? 4.6). Approximately 70% of the HH adolescents
completed 9 or more calls. HH adolescents were signifi-
cantly more satisfied with the telephone counseling compo-
nent than all other intervention components (mean of 4.05
of 5, SD ? 0.87; all t(18)? 3.33, p ? 0.01). HH adolescents
reported significantly greater satisfaction for mailed mate-
rials/manual than the computer interaction (t(18)? 3.01, p ?
0.01) but indicated similar levels of satisfaction between the
mailed materials/manual (mean ? 3.57, SD ? 1.13) and the
physician counseling (mean ? 3.39, SD ? 0.92), and be-
tween physician counseling and the computer interaction
(mean ? 2.98, SD ? 1.06).
Table 2. Weight status and secondary outcomes for post-treatment and follow-up assessment completers,
(n ? 20)
(n ? 19)
(n ? 20)
(n ? 19)
(n ? 18)
(n ? 19)‡)
Body mass index (BMI)
Percentage of overweight§
Percentage of calories
33.9 (8.8) 34.3 (4.9) 32.8 (10.0)34.2 (7.5)32.9 (9.4) 35.6 (6.4)
6.7 (5.6) 5.6 (5.1)7.8 (5.0)6.4 (6.1) 6.3 (3.5)6.9 (3.6)
* HH ? Healthy Habits intervention participants.
† TC ? typical care participants.
‡ 1 TC adolescent completed follow-up measurement of height and weight but refused completion of follow-up measures of secondary
outcome (e.g., physical activity).
§ Average percentage above the 50th-percentile BMI, based on National Center for Health Statistics/Centers for Disease Control and
Prevention 2000 growth curves (30).
¶ TFEQ ? Three-Factor Eating Questionnaire.
** CHEAT ? Children’s Eating Attitude Test.
Figure 2: Mean body mass index (BMI) z scores (?SEs) for the
Healthy Habits intervention and typical care conditions among
Weight Control for Adolescents, Saelens et al.
28 OBESITY RESEARCH Vol. 10 No. 1 January 2002
Bivariate partial correlations revealed a significant nega-
tive association between adolescent behavioral skills use
and post-treatment BMI z scores (r(36)? ?0.48, p ? 0.003;
more skill use related to lower post-treatment BMI z scores).
Neither the length of physician counseling (perceived by
adolescents or physicians) nor the amount of telephone
contact (number of calls and average call length) was re-
lated to weight outcomes.
A 4-month behavioral intervention initiated in primary
care with continuing telephone and mail contact led to a
modest decrease in weight status among overweight adoles-
cents (approximately ?0.05 in BMI z score). In contrast,
adolescents provided TC of a single provider counseling
session had an increase in BMI z score of 0.06, suggesting
weight acceleration among these already overweight ado-
lescents above the normative positive BMI trajectories of
adolescence. Although it is necessary to interpret cautiously
across age and gender among adolescents, these BMI z
score differences were associated with a ?0.2 BMI decrease
for the multiple component intervention and a BMI increase
of ?1.1 among adolescents provided TC. The behavioral
intervention did not produce absolute average weight loss,
but intervention adolescents were more likely than TC ad-
olescents to have improved their relative weight status, with
40% reducing BMI z scores by post-treatment. There was no
evidence that the intervention further favorably affected
weight from treatment cessation to follow-up. By the fol-
low-up, intervention adolescents continued to be more
likely to have decreased their BMI z scores and did not fully
return to their average baseline BMI z scores level or reach
those of TC adolescents.
Adolescents reported high levels of satisfaction with pro-
vider interaction around weight-related issues, regardless of
whether they received intervention after physician contact.
The multi-component intervention, and especially the tele-
phone component, was also rated highly. The telephone and
mail components seemed feasible and effective in promot-
ing behavioral skills use for weight control. Both adolescent
and parent reports indicated greater behavioral skills use
among intervention adolescents relative to typical care, at
least initially. Greater use of behavioral skills could have
served as a mechanism for better outcome among interven-
tion adolescents, because adolescents’ report of greater be-
havioral skills use was related to more positive weight
outcomes. The association between skills use and outcome
has been documented in other health behavior interventions
(44), and weight control interventions for younger children
(45) and adults (46,47).
Despite differential skills use and weight change by con-
dition, there were no observed differential changes by con-
dition in diet or physical activity behaviors. This disparity
has been found among some recent childhood obesity treat-
ment (48) and prevention (49) studies. It could be that
measures of weight-related behaviors in this study were not
sensitive enough to detect actual behavior change because
small daily changes in energy balance could explain weight
changes. Self-report methodology, consisting of short sam-
pling duration among constructs known to exhibit variation
across days (e.g., eating) (50), and modest measurement
reliability and validity, could have affected the results.
Given the intervention emphasis on self-monitoring, the
accuracy of self-reporting of diet and physical activity may
have improved among intervention adolescents, thus, in-
creasing the ratio of calories reported to actual calories
eaten, and perhaps resulting in the lack of condition differ-
ences. Furthermore, the social desirability of reporting
more physical activity and less overall and dietary fat cal-
ories within the context of weight control study assess-
ments could decrease the likelihood of condition differ-
ences. Participants could also have temporarily improved
eating and physical activity behaviors immediately before
assessments, which at most assessed the past week (e.g.,
PAR). More frequent and objective diet and physical activ-
ity assessment in the future would alleviate many of these
The lack of change in problematic eating- and weight-
related behaviors and beliefs suggests that neither the multi-
component intervention nor physician counseling alone
promotes increased problematic eating or weight-related
behaviors and beliefs among overweight youth. Present
mean scores on these measures were similar to those found
among normative adult and youth samples (39,41). The lack
of disturbed eating-related effects is consistent with results
from studies involving younger overweight children pro-
vided professionally led weight-loss treatment (51).
The magnitude of weight status change among interven-
tion adolescents was generally lower than results obtained
by the best clinic-based weight control treatments for over-
weight youth (7). Possible reasons for this difference could
include intervention intensity. The present 4-month inter-
vention was shorter than most typical adult (10) and pedi-
atric obesity interventions (8). Duration of each intervention
contact was shorter than most in-person treatment contacts.
In addition, the reinforcement system assigned points
weekly, but rewards were distal (only delivered at program
end), not guaranteed (lottery-based), and were program-
administered and not parent-based. Whereas phone coun-
selors praised healthful changes in eating and physical ac-
tivity, this motivational encouragement was infrequent
(weekly at best) and nonexistent for adolescents who did not
adhere well to the contact schedule. The level and type of
optimal parental involvement in adolescent weight control
remains unknown (7,52), but perhaps involving parents and
other caregivers in the praise and other positive reinforce-
ment of adolescents’ weight control behaviors, as is done
with younger children, would improve efficacy.
Weight Control for Adolescents, Saelens et al.
OBESITY RESEARCH Vol. 10 No. 1 January 200229
Reduced intensity may reduce efficacy but may expand
the number of adolescents who can participate and benefit
from more flexible intervention delivery format. Present
results compare favorably with other telephone- and corre-
spondence-based programs for adult weight-loss and
weight-gain prevention (11). Adolescents in the present
study may also differ from those included in previous youth
weight-control research. The telephone- and mail-based in-
tervention benefits from greater generalization, but the less
stringent inclusion criteria (e.g., no exclusion for psychiatric
comorbidity) and higher inclusion rate in the present study
relative to many previous randomized trials (8) could have
resulted in lower average short-term weight control or main-
tenance. Initially interested participants (?75%) were ran-
domized in this study (Figure 1), whereas recent clinic-
based studies have generally randomized fewer than 50% of
initially interested individuals (53,54). The factors that re-
duce likelihood of engaging in clinic-based interventions
(e.g., chaotic lifestyle) may be those related to poorer be-
havioral compliance and outcome. Boys and adolescents
from across ethnic groups were also well-represented in the
current study, unlike many previous evaluations of adoles-
cent weight-control treatment. However, as with previous
pediatric obesity treatment research, this sample was not
generally from low socioeconomic strata, perhaps due to the
bias of recruiting overweight adolescents from primary care
clinics and only those with an interest in weight control.
That single-episode provider counseling did not result in
positive weight change among overweight adolescents con-
verges with effects of other minimal and no-treatment con-
trol conditions on overweight children’s weight change
(55). There is some evidence that untreated overweight
youth gradually increase their weight status above expected
increases in BMI associated with age (56), although more
comprehensive data on weight-gain trajectories of over-
weight youth are generally lacking. Naturalistic or self-
initiated weight control efforts by youth may not only be
ineffective, but also perhaps paradoxically facilitating
weight gain (57). Such weight gain among the typical care
adolescents is consistent with the higher BMIs of over-
weight adolescents in the population compared with previ-
ous decades (2). However, it is also possible that an inad-
equate level of guidance in the TC condition somehow
promoted weight gain. Given high youth obesity prevalence
and high rates of weight dissatisfaction and reported weight-
control attempts among adolescents (58), more information
is needed about the effects of self-initiated weight control
practices among overweight youth.
This study provides preliminary evidence for the accept-
ability and short-term efficacy of a multi-component inter-
vention for adolescent weight control beginning in primary
care. The small sample size and response variability limited
power to detect statistically significant condition differences
through follow-up, because the effect size of condition
differences in BMI z scores to follow-up were moderate in
size (23). Studies including larger sample sizes, longer
follow-up, and more thorough process evaluation would
inform more comprehensive examination of efficacy and
identify the most potent components of this multi-compo-
nent approach. This could help formulate stepped-care treat-
ment models for intervening with overweight adolescents.
Innovative adolescent interventions that do not involve
weekly clinic-based visits have the potential to decrease
provider and participant cost, increase the number and di-
versity of treated overweight youth, and perhaps increase
participants’ acceptability of greater length of therapeutic
contact. This may help treatment providers capitalize on the
positive relationship between length of provider contact and
weight loss/maintenance success (59), consistent with a
continuous care model of obesity intervention (60).
This research was supported in part by a Young Investi-
gator’s Grant awarded by the North American Association
for the Study of Obesity to the first author. Appreciation is
expressed to Richard I. Stein, Danielle Kukene, and Beat-
rice Schmid for their invaluable assistance with the health
counseling and data collection.
1. Troiano RP, Flegal KM. Overweight children and adoles-
cents: description, epidemiology, and demographics. Pediat-
2. Flegal KM, Troiano RP. Changes in the distribution of body
mass index of adults and children in the U.S. population. Int J
Obes Relat Metab Disord. 2000;24:807–18.
3. Must A, Strauss RS. Risks and consequences of childhood
and adolescent obesity. Int J Obes Relat Metab Disord. 1999;
4. Must A, Jacques PF, Dallal GE, Bajema CJ, Dietz WH.
Long-term morbidity and mortality of overweight adolescents:
a follow-up of the Harvard Growth Study of 1922 to 1935.
N Engl J Med. 1992;327:1350–5.
5. Whitaker RC, Wright JA, Pepe MS, Seidel KD, Dietz WH.
Predicting obesity in young adulthood from childhood and
parental obesity. N Engl J Med. 1997;337:869–73.
6. U.S. Department of Health and Human Services. Healthy
People 2010. Washington, DC: U.S. Department of Health
and Human Services; 2000.
7. Brownell KD, Kelman JH, Stunkard AJ. Treatment of
obese children with and without their mothers: changes in
weight and blood pressure. Pediatrics. 1983;71:515–23.
8. Epstein LH, Myers MD, Raynor HA, Saelens BE. Treat-
ment of pediatric obesity. Pediatrics. 1998;101:554–70.
9. Jelalian E, Saelens BE. Empirically supported treatments in
pediatric psychology: pediatric obesity. J Pediatr Psychol.
10. National Institutes of Health, National Heart Lung and
Blood Institute. Clinical guidelines on the identification,
evaluation, and treatment of overweight and obesity in
adults—the evidence report. Obes Res. 1998;6:51S–129S.
Weight Control for Adolescents, Saelens et al.
30OBESITY RESEARCH Vol. 10 No. 1 January 2002
11. Leermakers EA, Jakicic JM, Viteri J, Wing RR. Clinic-
based vs. home-based interventions for preventing weight gain
in men. Obes Res. 1998;6:346–52.
12. Tate DF, Wing RR, Winett RA. Using Internet technology to
deliver a behavioral weight loss program. JAMA. 2001;285:
13. Marks A, Malizio J, Hoch J, Brody R, Fisher M. Assess-
ment of health needs and willingness to utilize health care
resources of adolescents in a suburban population. J Pediatr.
14. Gans JE, McManus MA, Newacheck PW. Adolescent
Health Care: Use, Costs, and Problems of Access. Chicago,
IL: American Medical Association; 1991.
15. Gauthier BM, Hickner JM, Ornstein S. High prevalence of
overweight children and adolescents in the Practice Partner
Research Network. Arch Pediatr Adolesc Med. 2000;154:
16. Nawaz H, Adams ML, Katz DL. Weight loss counseling by
health care providers. Am J Public Health. 1999;89:764–7.
17. Hodgson C, Feldman W, Corber S, Quinn A. Adolescent
health needs: utilization of health care by adolescents. Ado-
18. Sciamanna CN, Tate DF, Lang W, Wing RR. Who reports
receiving advice to lose weight?: results from a multistate
survey. Arch Intern Med. 2000;160:2334–9.
19. Simkin-Silverman LR, Wing RR. Management of obesity in
primary care. Obes Res. 1997;5:603–12.
20. Fairburn CG, Doll HA, Welch SL, et al. Risk factors for
binge eating disorder: a community-based case-control study.
Arch Gen Psychiatry. 1998;55:425–32.
21. Rosner B, Prineas R, Loggie J, Daniels SR. Percentiles for
body mass index in U.S. children 5 to 17 years of age.
J Pediatr. 1998;132:211–22.
22. Haddock CK, Shadish WR, Klesges RC, Stein RJ. Treat-
ments for childhood and adolescent obesity. Ann Behav Med.
23. Cohen J. Statistical Power Analysis for the Behavioral Sci-
ences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates;
24. Prochaska JJ, Zabinski MF, Calfas KJ, Sallis JF, Patrick
K. PACE?: interactive communication technology for behav-
ior change in clinical settings. Am J Prev Med. 2000;19:
25. Natow AB, Heslin J. The Fat Counter. 4th ed. New York,
NY: Pocket Books; 1998.
26. Epstein LH, Squires S. The Stoplight Diet for Children.
Boston, MA: Little, Brown and Co.; 1988.
27. Pate R, Trost S, Williams C. Critique of existing guidelines
for physical activity in young people. In: Biddle S, Sallis J,
Cavill N, eds. Young and Active? Young People and Health-
Enhancing Physical Activity—Evidence and Implications.
London, UK: Health Education Authority; 1998, pp. 162–73.
28. Barlow SE, Dietz WH. Obesity evaluation and treatment: expert
committee recommendations. Pediatrics. 1998;102:E29.
29. U.S. Department of Agriculture, U.S. Department of
Health and Human Services. Dietary Guidelines for Amer-
icans 2000. 5th ed. Washington, DC: U.S. Department of
Agriculture, U.S. Department of Health and Human Services;
30. Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, et al.
CDC Growth Charts: United States Advance Data from Vital
and Health Statistics. Hyattsville, MD: National Center for
Health Statistics; 2000, no. 314.
31. Frank GC, Berenson GS, Schilling PE, Moore MC. Adapt-
ing the 24-hr recall for epidemiologic studies of school chil-
dren. J Am Diet Assoc. 1977;71:26–31.
32. Lytle LA, Murray DM, Perry CL, Eldridge AL. Validating
fourth-grade students’ self-report of dietary intake: results
from the 5 a Day Power Plus Program. J Am Diet Assoc.
33. Crawford PB, Obarzanek E, Morrison J, Sabry ZI. Com-
parative advantage of 3-day food records over 24-hour recall
and 5-day food frequency validated by observation of 9- to
10-year-old girls. J Am Diet Assoc. 1994;94:626–30.
34. First DataBank I. Nutritionist Five User’s Guide, Version
1.6. San Bruno, CA: First DataBank I; 1998.
35. Sallis JF, Buono MJ, Roby JJ, Micale FG, Nelson JA.
Seven-day recall and other physical activity self-reports in
children and adolescents. Med Sci Sports Exerc. 1993;25:
36. Sallis JF, Saelens BE. Assessment of physical activity by
self-report: status, limitations, and future directions. Res Q
Exerc Sport. 2000;71:1–14.
37. Sarkin JA, Nichols JF, Sallis JF, Calfas KJ. Self-report
measures and scoring protocols affect prevalence estimates of
meeting physical activity guidelines. Med Sci Sports Exerc.
38. Guillaume M, Lapidus L, Bjo ¨rntorp P, Lambert A. Phys-
ical activity, obesity, and cardiovascular risk factors in chil-
dren: the Belgian Luxembourg Child Study II. Obes Res.
39. Stunkard AJ, Messick S. The three-factor eating question-
naire to measure dietary restraint, disinhibition and hunger.
J Psychosom Res. 1985;29:71–83.
40. Maloney MJ, McGuire JB, Daniels SR. Reliability testing of
a children’s version of the eating attitude test. J Am Acad
Child Adolesc Psychiatry. 1988;27:541–3.
41. Smolak L, Levine MP. Psychometric properties of the chil-
dren’s eating attitudes test. Int J Eat Disord. 1994;16:275–82.
42. Killen JD, Taylor CB, Hayward C, et al. Pursuit of thinness
and onset of eating disorder symptoms in a community sample
of adolescent girls: a three-year prospective analysis. Int J Eat
43. Killen JD, Taylor CB, Hayward C, et al. Weight concerns
influence the development of eating disorders: a 4-year pro-
spective study. J Consult Clin Psychol. 1996;64:936–40.
44. Saelens BE, Gehrman CA, Sallis JF, et al. Use of self-
management strategies in a two-year cognitive-behavioral in-
tervention to promote physical activity. Behav Ther. 2000;31:
45. Israel AC, Silverman WK, Solotar LC. The relationship
between adherence and weight loss in a behavioral treatment
program for overweight children. Behav Ther. 1988;19:25–33.
46. Boutelle KN, Kirschenbaum DS. Further support for consis-
tent self-monitoring as a vital component for successful
weight control. Obes Res. 1998;6:219–24.
Weight Control for Adolescents, Saelens et al.
OBESITY RESEARCH Vol. 10 No. 1 January 200231
47. Baker RC, Kirschenbaum DS. Self-monitoring may be nec-
essary for successful weight control. Behav Ther. 1993;24:
48. Epstein LH, Valoski AM, Vara LS, et al. Effects of
decreasing sedentary behavior and increasing activity on
weight change in obese children. Health Psychol. 1995;14:
49. Robinson TN. Reducing children’s television viewing to pre-
vent obesity: a randomized controlled trial. JAMA. 1999;282:
50. de Castro JM. Prior day’s intake has macronutrient-specific
delayed negative feedback effects on the spontaneous food
intake of free-living humans. J Nutr. 1998;128:61–7.
51. Epstein LH, Paluch RA, Saelens BE, Ernst MM, Wilfley
DE. Changes in eating disorder symptoms with pediatric
obesity treatment. J Pediatr. 2001;139:58–65.
52. Wadden TA, Stunkard AJ, Rich L, et al. Obesity in black
adolescent girls: a controlled clinical trial of treatment by diet,
behavior modification, and parental support. Pediatrics. 1990;
53. Golan M, Weizman A, Apter A, Fainaru M. Parents as the
exclusive agents of change in the treatment of childhood
obesity. Am J Clin Nutr. 1998;67:1130–5.
54. Epstein LH, Paluch RA, Gordy CC, Saelens BE, Ernst
MM. Problem solving in the treatment of childhood obesity. J
Consult Clin Psychol. 2000;68:717–21.
55. Lansky D, Vance MA. School-based intervention for adoles-
cent obesity: analysis of treatment, randomly selected control,
and self-selected control subjects. J Consult Clin Psychol.
56. Braet C, Van Winckel M, Van Leeuwen K. Follow-up
results of different treatment programs for obese children.
Acta Paediatr. 1997;86:397–402.
57. Stice E, Cameron RP, Killen JD, Hayward C, Taylor CB.
Naturalistic weight-reduction efforts prospectively predict
growth in relative weight and onset of obesity among female
adolescents. J Consult Clin Psychol. 1999;67:967–74.
58. Neumark-Sztainer D, Hannan PJ. Weight-related behaviors
among adolescent girls and boys: results from a national
survey. Arch Pediatr Adolesc Med. 2000;154:569–77.
59. Perri MG, Nezu AM, Patti ET, McCann KL. Effect of
length of treatment on weight loss. J Consult Clin Psychol.
60. Latner JD, Stunkard AJ, Wilson GT, et al. Effective long-
term treatment of obesity: a continuing care model. Int J Obes
Rel Metab Disord. 2000;24:893–8.
Weight Control for Adolescents, Saelens et al.
32 OBESITY RESEARCH Vol. 10 No. 1 January 2002