International Journal of Obesity (2006) 30, 1565–1573
& 2006 Nature Publishing Group All rights reserved 0307-0565/06 $30.00
www.nature.com/ijo
ORIGINAL ARTICLE
Mail and phone interventions for weight loss in a
managed-care setting: weigh-to-be 2-year outcomes
NE Sherwood1, RW Jeffery2, NP Pronk1,3, JL Boucher3, A Hanson1, R Boyle1, K Brelje2, K Hase1 and
V Chen2
1
HealthPartners Research Foundation, Minneapolis, MN, USA; 2Division of Epidemiology, University of Minnesota School of
Public Health, Minneapolis, MN, USA and 3HealthPartners Center for Health Promotion, Minneapolis, MN, USA
Objective: Evaluate effectiveness of weight-loss interventions in a managed care setting.
Methods: Three-arm randomized clinical trial: usual care, mail, and phone intervention. Participants were 1801 overweight
managed care organization (MCO) members. Measures included baseline height, weight at baseline and 24 months, selfreported weight at 18 months. Intervention and participation in other weight-related programs was monitored across 24
months.
Results: Weight losses were 2.2, 2.4, and 1.9 kg at 18 months in the mail, phone, and usual care groups, respectively. Mail and
phone group weight changes were not significantly different from usual care (Po0.35).Weight losses at 24 months did not differ
by condition (0.7 kg mail, 1.0 kg phone, and 0.6 kg usual care, P ¼ 0.55). Despite treatment availability over 24 months,
participation diminished after 6 months. Participation was a significant predictor of outcomes in the mail and phone groups at
18 months and the mail group at 24 months. Cost-effectiveness of phone counseling was $132 per 1 kg of weight loss with mail
and usual care achieving similar cost-efficiency of $72 per 1 kg of weight loss.
Conclusion: Although mail- and phone-based weight-loss programs are a reasonably efficient way to deliver weight-loss
services, additional work is needed to enhance their short- and long-term efficacy.
International Journal of Obesity (2006) 30, 1565–1573. doi:10.1038/sj.ijo.0803295; published online 21 March 2006
Keywords: weight loss; managed care; phone; mail
Introduction
Obesity is emerging as an increasingly important public
health problem.1–4 Current estimates from the National
Center for Health Statistics suggest that more than a third of
adults are obese, with an additional 30% classified as overweight.4 Overweight and obesity increase the risk of many
serious health conditions, including hypertension, hypercholesterolemia, diabetes, coronary heart disease, and some
forms of cancer.5–8 Obesity is fast approaching tobacco as the
leading contributor to premature death in the United States.9
Despite the major health significance of obesity, health
care providers currently have few obesity management
options. Current pharmacological treatments are only
modestly effective relative to cost10,11 and bariatric surgery
is not appropriate for many patients and prohibitively
Correspondence: Dr NE Sherwood, HealthPartners Research Foundation, PO
Box 1524, MS 21111R, Minneapolis, MN 55440-1524, USA.
E-mail: [email protected]
Received 29 June 2005; revised 16 December 2005; accepted 9 January 2006;
published online 21 March 2006
expensive as a public health strategy.12 Other treatments,
including intensive behavioral treatments requiring multiple
in-person contacts, are also too expensive for widespread use
and only modestly effective.13 Participation in these treatment programs is also relatively low. Less than 50% of
overweight women and 20% of overweight men have ever
participated in a formal weight-control program,14,15 and
about two-thirds of these are low-cost commercial programs
with very high dropout rates.16,17 Cost-effective approaches
that can reach a broad population of individuals are needed.
Mail and/or phone-based counseling approaches may be
viable alternatives to expensive in-person clinic visits.18–24
However, the efficacy of these types of lower-intensity
interventions needs to be evaluated.25
The present report describes 2-year results of the Weigh-ToBe study (WTB), the first ever large-scale randomized trial
examining the efficacy of nonclinic-based weight-loss interventions in a health care delivery system. WTB compared
mail and phone-based weight loss interventions to usual care
in a managed care setting. This report compares the three
treatment groups in terms of level of member participation,
weight loss efficacy, and cost. Additionally, associations
Mail and phone interventions for weight loss in a managed-care setting
NE Sherwood et al
1566
between program participation and weight outcomes are
examined.
Materials and methods
This project was a collaboration among the University of
Minnesota School of Public Health, the HealthPartners
Research Foundation, and the HealthPartners Center for
Health Promotion (CHP). Study recruitment began in August
1999 and data collection was completed for all study
participants in October 2002. HealthPartners is a mixedmodel managed care organization (MCO) that provides
medical care service for approximately 700 000 members in
the Minneapolis/St Paul, Minnesota metropolitan area. The
study was a randomized trial evaluating the effectiveness of
phone-based and mail-based interventions for weight reduction offered to members in comparison to usual care over
2-years. Recruitment targeted overweight MCO members
from four clinics. Two clinics served members in St Paul and
Minneapolis and two served individuals in adjacent suburbs.
Participant recruitment was conducted over 12 months
using several methods. Direct mail announcements were
sent to the households of 31 000 clinic members. Posters and
flyers were displayed in each of the four clinics. Referrals
from clinic physicians and nurses were encouraged by e-mail
announcement and presentations at staff meetings. The
study was also advertised on the MCO web site. Prospective
study participants were instructed to call a study phone
number for information about the project.26
Eligibility criteria were intentionally broad. Inclusion
requirements included age 18 years or older and body mass
index (BMI) greater than 27.0 based on reported height and
weight. Age and weight eligibility were ascertained in the
informational phone call. A clinic visit at one of the participating clinics was also scheduled at that time. Individuals
attending this visit had the study described to them and
signed an informed consent statement approved by the
Human Subjects Review Committees of both the University
of Minnesota and HealthPartners Research Foundation.
After consent, weight and height were measured using a
calibrated balance-beam scale and a wall-mounted ruler.
Questionnaires, described further below, were also completed. Overall, 3294 individuals requested information
about the study, 2205 met the eligibility criteria and of
those, 1801 completed the baseline consent process, and
enrolled in the study. Seventy-five percent of enrolling
participants indicated that they learned of the study through
the mailed announcements, 14% through clinic posters, 4%
from physician referral, 4% by word-of-mouth, 1% from the
web site, and 2% other source.
Following baseline, the Project Manager randomized
participants using an automated computer system to one
of three conditions: mail intervention, phone intervention,
and usual care. The randomization scheme consisted of
International Journal of Obesity
blocks of 15 with the numbers 1–3 to indicate treatment
group (phone, mail and usual care); the randomization
sequence was concealed until after interventions were
assigned. After randomization, participants were sent a letter
informing them of their assignment. To measure relative
interest in the two treatment conditions, participants were
asked to notify the study when they wished to begin their
program. Mail intervention individuals were asked to
indicate their readiness by sending a postcard to the study
office. Phone treatment individuals were given a phone
number to call to activate treatment.
Once activated, the two weight loss interventions proceeded in parallel formats. Both were comprised of 10
interactive lessons designed to be completed in sequence
with feedback between each lesson from a health counselor.
Each lesson included instructional material describing a
rationale for a specific behavior change strategy, behavior
change goals related to that strategy, and homework to be
completed before beginning the next lesson. Lesson topics
included nutrition, physical activity, and behavior management techniques (e.g., behavioral assessment, goal setting,
stimulus control, social support, and self-motivation). The
primary homework assignment was to keep a food and
exercise log.
Weight management lessons were designed to be completed as rapidly as one lesson per week. However, study
participants were encouraged to proceed at a pace comfortable for them. For phone intervention individuals, all 10
lessons and homework assignment materials were mailed at
the beginning of the program. A series of calls was scheduled
between the participant and a phone counselor to provide
guidance through each lesson and feedback about progress.
Phone counselors were staff members of the CHP and were
trained nutritionists and/or exercise specialists.27 During an
introductory telephone call, program format and expectations were explained and subsequent calls were scheduled.
These calls were comprised of discussion of behavioral
strategies tried since the last session, discussion of content
and activities for the lesson, counselor advice about how to
improve/maintain lifestyle behaviors, goal setting, and
counselor description of the rationale and behavioral assignment for the next lesson. The average length of calls was
19 min.
Mail intervention used the same 10 written lessons,
behavioral assignments, and counseling protocol and staff.
However, interactions between counseling staff and participants were entirely by mail. Participants were first mailed a
course manual with two lessons and two feedback forms and
were instructed to complete the first lesson and return a
progress report. Progress report information included behavior change goals, perceived progress, and action steps taken
to achieve goals. When this progress report was received by
the counselor, she reviewed it and made comments in
writing, which were forwarded, along with the next session,
by return mail. This sequence was repeated for each lesson
until the course was completed.
Mail and phone interventions for weight loss in a managed-care setting
NE Sherwood et al
1567
Follow-up intervention options were available to both the
phone and the mail groups after completion of the 10-lesson
course. These were comprised of individual follow-up on
topics of the participant’s choosing. Resources available to
the counselor included a wide range of educational resources
on lifestyle topics related to weight management maintained
by the CHP. Participants could also enroll in other CHP
health-related courses. Additionally, participants could repeat all or any part of the WTB intervention. Participants
who discontinued contact with their counselor prior to
course completion were contacted at 1-, 2-, and then 6month intervals for up to 2 years to encourage intervention
resumption. Individuals who did not activate their assigned
intervention were also contacted at 6-month intervals to
encourage engagement.
Usual care participants had access only to weight management services generally available to members of HealthPartners. After randomization, they were sent a resource sheet
detailing MCO and community weight management options
including free general phone counseling, a structured weight
management phone course, or a group class offered at several
MCO clinics. The phone course and group classes required a
modest fee of $25. Similar to participants in the treatment
groups, usual care participants could enroll in other CHP
health-related courses.
Study measures
The WTB project encompassed 24 months for each study
participant. At baseline and 24 months, clinic visits were
held at which body weight was measured and self-report
measures were completed. Measurement staff were blind to
study condition. Between these time points, follow-up
assessments were conducted by mailed questionnaire at 6,
12, and 18 months. Self-reported weight was obtained at
these time points. 6 and 12-months outcomes have been
reported previously28 and 18 and 24-month outcomes are
reported here. The following information is included in this
report.
Demographic characteristics: Measures at baseline included
gender, age (year), education (phigh school, 4high school
education but ocollege degree, Xcollege degree or greater),
ethnicity (white vs other), and marital status (currently
married; separated, divorced, or widowed; never married).
Smoking status: Currently smoking (yes/no), measured
every 6 months.
Weight status: BMI was computed from weight and height
(kg/m2) for each time point. A categorical weight status
variable was also created: overweight (BMI 25–29.9); Class I
obesity (BMI 30–34.9), Class II obesity (BMI 35–39.9), and
Class III obesity (BMIX40). Measured weight was obtained at
the baseline and 24-month follow-up clinic visits; selfreported weight was obtained at 18-months.
Weight loss history: Ever dieted (yes/no) and participated in
a formal weight loss program in the last 2 years (yes/no),
measured at baseline.
Current medication use: Diabetes (yes/no), depression (yes/
no), and cardiovascular disease (yes/no), measured at baseline.
Participation measures: Weight control activity participation was assessed in two ways using the tracking systems that
are part of the CHP delivery platform.27 These records
document the dates and types of all contacts between CHP
staff and members, both for the WTB program and other
CHP programs. Analysis variables for mail and phone group
participants included enrollment status (yes/no) and number
of WTB course sessions completed (0–10). Additionally, the
total number of weight-related encounters outside of the
WTB protocols were examined. The operational definition of
an ‘encounter’ was an educational interaction that focused
on the topics of weight, diet, and/or physical activity
between CHP staff and a participant. This information was
available for all three study conditions.
Analysis
The primary outcomes examined in this study are changes in
body weight from baseline to 18 and 24 months. A required
sample size of 500 participants was determined using
calculations to have 90% power (a ¼ 0.05, two-tailed) to
detect a small effect size for intent-to-treat analyses. Data
analyses were performed using the general linear models
programs of SAS (Version 8.7). First, group differences in
weight change at 18 and 24 months were examined
controlling for baseline weight. Second, group differences
in weight change at 18 and 24 months were examined
controlling for baseline body weight and individual-level
factors significantly associated with weight change. Third,
treatment participation was compared between the mail and
phone groups and outcomes were examined as a function of
enrollment status and participation, categorized into three
groups: no lessons completed (included individuals not
active in treatment), 1–9 lessons, and 10 lessons completed.
Finally, the relationship between number of weight-related
encounters and weight loss outcomes was examined across
all three groups. These analyses used an intent-to-treat
approach in which baseline values for body weight (0 weight
loss) were used for individuals who did not complete followup surveys.
A cost-effectiveness analysis was also completed. Cost data
were compiled on counseling, program development, materials/supplies and overhead. Counseling cost was derived
from time spent in encounters between the participants and
the counselors and the salary (including fringe benefits) of
an average CHP counselor employed. The times of 30 actual
intervention calls and letters were recorded and an average
counseling time for each intervention type was computed. A
time estimate was attached to each encounter call and total
counseling time per participant over the 24 months was
computed. Development cost was based on staff time and
material for intervention development. The usual care group
was assigned zero on this component since the counseling
International Journal of Obesity
Mail and phone interventions for weight loss in a managed-care setting
NE Sherwood et al
1568
programs utilized by this group were previously developed
by CHP. Overhead costs included the operating cost of the
computer system, office space rental, and phone charges. A
cost-effectiveness ratio was computed for each group to
provide a measure of how efficiently the interventions
produced a kilogram of weight loss.
Figure 2 provides an overview of weight change over time
by treatment group status, including 6 and 12 months
outcomes presented in previously published work.28 Table 2
presents weight-change outcomes at 18- and 24-month
follow-up. Although on average participants across the study
conditions lost weight (i.e., weight loss in each group was
significantly different from 0), no statistically significant
group differences in weight change were observed at 18
months (Cohen’s d, Mail vs UC ¼ 0.03; Phone vs UC ¼ 0.04) –
or 24-months (Cohen’s d, Mail vs UC ¼ 0.01; Phone vs
UC ¼ 0.03). At 18-months, about 24% of participants lost 5%
of their body weight, with almost 10% losing 10% of their
body weight. At 24-months, about 13% of participants lost
5% of their body weight, with about 5% losing 10% of their
body weight. No adverse events and/or side effects were
observed as a result of intervention participation.
Table 3 shows the proportion of individuals in the mail
and phone groups who activated treatment. Activation was
higher in the mail (88%) compared to the phone group
(69%; w2 ¼ 62.97, Po0.001). Among treatment activators,
participation was higher in the phone compared to the mail
condition. The average number of treatment lessons completed was 7.2 in the phone group and 2.3 in the mail group
through 24 months (t ¼ 20.60, Po0.001). Over half of
activated phone participants completed the program,
whereas just over 10% in the mail group completed the
Results
A flowchart of study participation according to CONSORT
guidelines is shown in Figure 129,30 and study sample
descriptive information is shown in Table 1.28 The majority
of participants were Caucasian (n ¼ 1639, 91%), welleducated (n ¼ 899, 50% college or graduate degree), female
(n ¼ 1293, 72%), middle-aged (mean age ¼ 50 years, s.d. ¼ 12)
adults. Average BMI was about 33.5 kg/m2. About 30% met
the prevailing definition for overweight, 40% for Class I
obesity, 20% for Class II obesity, and 10% for Class III
obesity. A substantial number of participants reported taking
medication for diabetes (5.9%), depression (13.9%), or
cardiovascular disease (CVD) related problems (e.g., high
blood pressure) (27.3%). Treatment groups differed significantly on one baseline variable. Phone group participants
were more likely to report taking depression medication
than those in the other groups (Po0.013).
Assessed for eligibility
(n=3294)
Enrollment
Excluded (n=1493)
Did not meet inclusion criteria (n=1091)
Refused to participate (n=60)
Other reasons (e.g., clinic visit no show) (n=342)
Randomization
Mail (n=600)
Weigh-To-Be Course
Activated intervention (n=528)
Activated, 0 sessions completed
(n=260)
Completed 1-9 sessions (n=206)
Completed all 10 sessions (n=62)
Did not activate intervention (n=72)
CHP Weight-Related Encounters
0-1 Counseling Encounters (n=430)
2-9 Counseling Encounters (n=115)
10+ Counseling Encounters (n=55)
Phone (n=601)
Weigh-To-Be Course
Activated intervention (n=416)
Activated, 0 sessions completed (n=24)
Completed 1-9 sessions (n=165)
Completed all 10 sessions (n=227)
Did not activate intervention (n=185)
CHP Weight-Related Encounters
0-1 Counseling Encounters (n=218)
2-9 Counseling Encounters (n=138)
10+ Counseling Encounters (n=245)
Usual Care (n=600)
Weigh-To-Be Course
Activated intervention NA
Activated, 0 sessions completed NA
Completed 1-9 sessions NA
Completed all 10 sessions NA
Did not activate intervention NA
CHP Weight-Related Encounters
0-2 Counseling Encounters (n=486)
2-9 Counseling Encounters (n=54)
10+ Counseling Encounters (n=60)
Follow-Up
Mail
Lost to follow-up
Quit study prior to 24 mos (n=85)
Lost to follow-up at 24 mos (n=134)
24 Month data available for analysis
Measured weight & survey data (n=325)
Survey data only (n=56)
Figure 1 Weigh-To-Be (WTB) CONSORT guideline flowchart.
International Journal of Obesity
Phone
Lost to follow-up
Quit study prior to 24 mos (n=68)
Lost to follow-up at 24 mos (n=128)
24 Month data available for analysis
Measured weight & survey data (n=341)
Survey data only (n=63)
Usual Care
Lost to follow-up
Quit study prior to 24 mos (n=84)
Lost to follow-up at 24 mos (n=106)
24 Month data available for analysis
Measured weight & survey data (n=337)
Survey data only (n=73)
Mail and phone interventions for weight loss in a managed-care setting
NE Sherwood et al
1569
Table 1
Description of study population at baseline
Mail (n ¼ 600)
%/mean (s.e.)
Phone (n ¼ 601)
%/mean (s.e.)
Usual care (n ¼ 600)
%/mean (s.e.)
P-value
69.0
73.5
72.8
0.170
50.6 (0.5)
50.7 (0.5)
50.8 (0.5)
0.943
Education:
High school or less
Some college/vocational training
College or grad degree
12.2
39.6
48.3
11.7
37.3
51.1
13.2
36.3
50.5
Ethnicity (% white)
90.0
92.2
91.0
Marital status:
Married
Separated, divorced, or widowed
Never married
70.8
20.8
8.3
68.2
20.0
11.8
71.8
16.5
11.7
9.0
9.0
7.8
0.708
34.1 (0.2)
33.5 (0.2)
34.0 (0.2)
0.435
Weight status (BMI)
Overweight (25–29.9)
Obese class I (30–34.9)
Obese class II (35–39.9)
Obese class III (X40)
Ever dieted (% yes)
25.3
39.0
21.3
14.3
85.8
27.8
35.8
19.0
17.3
87.0
27.4
39.1
18.5
15.0
84.7
0.525
Formal weight-loss program in past 2 years (% yes)
26.2
30.8
30.8
0.125
Current medication (% yes)
Diabetes
Depression
CVD-related
4.7
12.5
26.0
6.5
18.0
27.6
5.3
11.2
28.3
0.374
0.002
0.649
Gender (% female)
Age (year)
0.737
0.214
Current smoker (% yes)
2
BMI (kg/m )
0.525
0
-0.5
Weight change
-1
-1.5
-2
-2.5
m
os
24
m
os
18
m
os
12
os
m
6
el
in
e
-3
Ba
s
0.900
Assessment T i me
Mail
Phone
Usual Care
Figure 2 Weight change over time among Weigh-To-Be (WTB) participants
by treatment group.
program (w2 ¼ 294.21, Po0.001). The phone course took
fewer days to complete (phone average, 158 days; mail
average, 263). Table 3 also shows the predictive value of
study activation and level of program completion with
respect to 18- and 24-month weight loss. Study activation
alone was not a strong predictor of weight-loss. Difference in
weight loss by this measure was significant only at 18months in the phone intervention (Cohen’s d ¼ 0.08).
Number of completed lessons was a better predictor of
outcomes than treatment activation. Observed weight losses
were consistently higher in those who completed more
lessons in the mail condition at both 18 months (Cohen’s d,
1–9 vs 0 lessons ¼ 0.11; 10 vs 0 lessons ¼ 0.34) and 24 months
(Cohen’s d, 1–9 vs 0 lessons ¼ 0.04; 10 vs 0 lessons ¼ 0.28).
Participation was associated significantly with weight loss at
18-months in the phone condition (Cohen’s d, 1–9 vs 0
lessons ¼ 0.06; 10 vs 0 lessons ¼ 0.18); however, no significant weight change differences by level of lesson completion
were observed at 24-months (Cohen’s d, 1–9 vs
lessons ¼ 0.02; 10 vs 0 lessons ¼ 0.03).
In addition to the WTB course, participants across the
three study conditions had the opportunity to participate in
CHP weight management and related programs. The average
number of classes taken by participants in the mail, phone,
and usual care conditions were 1.1 (s.d. ¼ 0.6), 1.2
(s.d. ¼ 1.1), and 0.5 (s.d. ¼ 0.7), respectively. About a third
of usual care participants (34.3%, n ¼ 394) were active in
International Journal of Obesity
Mail and phone interventions for weight loss in a managed-care setting
NE Sherwood et al
1570
Table 2
Weight-change outcomes at 18 and 24 months by treatment group
Mail
%/mean (s.e.)
Phone
%/mean (s.e.)
Usual care
%/mean (s.e.)
P-value
18-month weight outcomesa
Weight changeb (kg)
Adjusted weight changec (kg)
5% weight lossb (% yes)
10% weight lossb (% yes)
n ¼ 600
2.27 (0.24)
2.24 (0.24)
23.67
9.67
n ¼ 601
2.35 (0.24)
2.39 (0.24)
25.33
9.50
n ¼ 600
1.91 (0.24)
1.90 (0.24)
22.67
8.17
0.390
0.348
0.549
0.613
24-month weight outcomes
Weight changeb (kg)
Adjusted weight changec (kg)
5% weight lossb (% yes)
10% weight lossb (% yes)
n ¼ 600
0.73 (0.22)
0.70 (0.22)
13.33
4.83
n ¼ 601
0.93 (0.22)
0.96 (0.22)
14.50
5.33
n ¼ 600
0.59 (0.22)
0.59(0.22)
12.17
4.33
0.552
0.475
0.493
0.722
Note: Means with unshared superscripts were significantly different in pair-wise post hoc comparisons (Pp0.05). a18-month weight data is based on self-report by
participants. bAnalysis controlled for baseline body weight (kg). cAnalysis controlled for baseline weight and significant predictors of 18- and 24-month weight
change: sex, age, formal weight-loss program participation, and depression medication.
Table 3
Program participation and weight loss by treatment group
Mail condition
%/mean (s.d.)
Participation, 24 months:
Activating treatment (%)
Number of sessions completed
Completed 10 sessions (% yes)
(Of those enrolled % yes)
Weight loss, 18 months:
Not activated
Activated
Weight loss, 24 months:
Not activated
Activated
Weight loss, 18 months:
Not activating or 0 lessons
Completed 1–9 lessons
Completed 10 lessons
Weight loss, 24 months:
Not activating or 0 lessons
Completed 1–9 lessons
Completed 10 lessons
88.0 (n ¼ 528)
2.3 (3.5)
10.3 ( n ¼ 62)
11.7
Phone condition
%/mean (s.d.)
69.2 (n ¼ 416)
7.2 (3.7)
38.4 (n ¼ 231)
55.5
1.5 (0.7)(n ¼ 72) 1.6 (0.5)(n ¼ 187)
2.4 (0.2)(n ¼ 528) 2.7 (0.3)(n ¼ 414)
Po0.178
Po0.030
0.6 (0.6)(n ¼ 72) 0.9 (0.4)(n ¼ 186)
0.8 (0.2)(n ¼ 528) 1.0 (0.3)(n ¼ 415)
Po0.818
Po0.900
1.5a (0.3)(n ¼ 332) 1.4a (0.4)(n ¼ 209)
2.7b (0.4)(n ¼ 206) 2.1a (0.5)(n ¼ 165)
5.5c (0.7)(n ¼ 62) 3.5b (0.4)(n ¼ 227)
Po0.001
Po0.002
0.3a (0.3)(n ¼ 332) 0.8 (0.4)(n ¼ 208)
0.7a (0.3)(n ¼ 206) 1.0 (0.4)(n ¼ 163)
3.3b (0.6)(n ¼ 62) 1.1 (0.4)(n ¼ 230)
Po0.001
Po0.845
ences were observed between usual care group participants
by participation status.
Data were also analyzed by the number of weight-related
encounters with CHP. As shown in Table 4, number of CHP
encounters was significantly associated with 18-month
weight change in the mail and phone conditions. Mail
participants with ten or more encounters lost more weight
than those with fewer encounters (Cohen’s d, 2–9 vs 0-1
encounters ¼ 0.07; 10 þ vs 0–1 encounters ¼ 0.23). Phone
participants with 10 or more encounters lost more weight
than those with one or fewer encounters (Cohen’s d, 2–9 vs
0–1 encounters ¼ 0.08; 10 þ vs 0–1 encounters ¼ 0.16).
Encounter number was a significant predictor of weight
change at 24-months among mail participants only. Mail
participants with two or more encounters lost more weight
than those with one or fewer encounters (Cohen’s d, 2–9 vs
0–1 encounters ¼ 0.14; 10 þ vs 0–1 encounters ¼ 0.23).
Table 5 presents data from the cost-effectiveness analyses
and show that counseling time was the largest cost
component, accounting for 54.9, 73.2 and 84.4% of total
expenditure in mail, phone and usual care groups, respectively. Program development costs were the second largest
expenditure. The cost-effectiveness ratios show that phone
counseling appeared to be least efficient at a price tag of $132
in producing 1 kg of weight loss while mail and usual care
group achieved similar efficiency of $72 per 1 kg weight loss.
Note: Means with unshared superscripts were significantly different in pairwise post hoc comparisons (Pp0.05).
Discussion
CHP classes; among those who were active in classes, the
average number of classes taken was 1.33 (s.d. ¼ 0.04). Usual
care group participants active in classes lost more weight at
18-months compared to inactive participants (Active ¼ 2.6
(s.d. ¼ 7.9); Not Active ¼ 1.5 (s.d. ¼ 4.3), Po0.05, Cohen’s
d ¼ 0.09). No significant 24-month weight change differInternational Journal of Obesity
This study was the first large scale randomized trial
evaluating weight loss interventions on a large scale in a
real world health care delivery setting. We deliberately chose
to test mail and phone-based programs that might be
affordable in such a setting and overall we conclude that
Mail and phone interventions for weight loss in a managed-care setting
NE Sherwood et al
1571
Table 4
Center for Health Promotion encounters and weight loss by treatment group – participation, 18 and 24 months
Mail
Phone
Usual care
All weight-related encounters (including WTB encounters), 18 months
0–1
1.89a (0.26)(n ¼ 430)
2–9
2.68a (0.51)(n ¼ 115)
10+
4.48b (0.74)(n ¼ 55)
Po0.003
1.36a (0.42)(n ¼ 218)
2.35a,b (0.52)(n ¼ 138)
3.30b (0.39)(n ¼ 245)
Po0.003
1.72 (0.28)(n ¼ 486)
1.98 (0.84)(n ¼ 54)
2.96 (0.80)(n ¼ 60)
Po0.347
All weight-related encounters (including WTB encounters), 24 months
0–1
0.27a (0.24)(n ¼ 430)
2–9
1.69b (0.47)(n ¼ 115)
10+
2.34b (0.49)(n ¼ 55)
Po0.001
0.69 (0.37)(n ¼ 218)
0.86 (0.46)(n ¼ 138)
1.23 (0.34)(n ¼ 245)
Po0.553
0.47 (0.33)(n ¼ 486)
0.18 (0.47)(n ¼ 54)
1.65 (0.69)(n ¼ 60)
Po0.275
Note: Means with unshared superscripts were significantly different in pair-wise post hoc comparisons (Pp0.05).
Table 5
Costs, effects, cost-effectiveness ratios and incremental cost-effectiveness ratios for the weigh-to-be study
Mail (n ¼ 600)
Phone (n ¼ 601)
Usual care (n ¼ 600)
Total (n ¼ 1801)
Costs
Counseling/subject
Program development/subject
Materials and supplies/subject
Overhead/subject
Total cost/subject
$27.71
$11.98
$5.65
$5.11
$50.45
$93.28
$11.98
$4.92
$17.21
$127.39
$35.61
$0.00
$0.00
$6.57
$42.18
$52.23
$7.99
$3.53
$9.64
$73.38
Effects
24-month weight outcomes
Adjusted weight change (kg)
0.70 (1.13, 0.27)a
0.96 (1.39,0.53)
0.59 (1.02,0.16)
$72.08
$132.70
$71.50
Cost-effectiveness ratios
Cost/weight loss of 1 kg
a
The numbers in parentheses represent 95% confidence interval of the estimated parameters.
delivering mail and phone interventions are potentially
feasible in this regard. Recruitment and participation rates
were acceptable and program implementation costs under
$200.00 per person seem reasonable, particularly in comparison to recent estimates of the economic costs of obesity.31
Examination of relationships between BMI and health care
costs has shown that annual medical charges are about
$214.00 greater for individuals with a BMI between 30 and
34 kg/m2 compared to individuals with a BMI in the normal
range (o25 kg/m2). Data also suggest, not surprisingly, that
outcome is dependent on engagement level. At 18 months,
participants who were more active in both the phone and
mail conditions lost more weight compared to less active
participants. Interestingly, at 24 months, mail participants
who were more active in treatment were more successful,
whereas there was no similar effect for phone participants.
These data suggest that mail completers may have been more
self-motivated for weight loss.
Despite room for optimism, our data clearly show that
weight-loss efficacy needs improvement. At best, the 24
month results show that the interventions, including those
that usual care participants took part in, served as effective
weight gain prevention as opposed to weight loss programs.
Weight loss results were more promising at 6 months than
later.28 In examining the timing of the WTB course and other
weight-related encounters, we observed that the majority of
counseling contacts occur during the first 6–12 months of
the study, so failure to detect significant group differences at
the later time points is not completely surprising. An
additional factor in the interpretation of these findings is
the fact that the ‘usual care’ was unusually potent in this
study. The CHP is unique in its offering of relatively low cost
weight management services to members. Many members,
however, are probably not aware of these services and thus
don’t use them. Usual care participants in this study were
explicitly made aware of these member services and
participated in them at relatively high rates, about 1 person
in 3. As a result, significant weight loss observed in our
‘control’ group may have lessened our ability to detect
effects in our active treatments.
Considering our intervention experience overall, we
believe that the interventions could be considerably
strengthened by implementation of stronger intervention
messages and engagement strategies for keeping people
International Journal of Obesity
Mail and phone interventions for weight loss in a managed-care setting
NE Sherwood et al
1572
involved in weight loss over longer periods of time. The
Diabetes Prevention Program (DPP) provides an excellent
model for both improving the potency of the intervention
messages and engaging participants long term.32 Examination of these results and lessons learned from the DPP
suggest several strategies to improve outcomes. First, in this
trial, participants in the phone and mail conditions selfinitiated treatment so that the attractiveness of each format
could be evaluated. Future interventions should include
more active outreach efforts to engage participants. A second
strategy is the incorporation of stronger behavioral messages
and strategies. Our previous work and the work of others,
shows that participants who more actively self-monitor their
behavior and weight are more successful.33 Weight selfmonitoring was not strongly encouraged in WTB, but future
interventions should promote this behavior. Dietary and
activity goal-setting should also be strengthened. As suggested by the DPP, strategies such as motivational campaigns
and a ‘tool box’ should be used to provide extra support to
struggling participants and novelty to keep participants
engaged longer.28 Additionally, the use of incentives for
program participation should be explored (e.g., reimbursement of program costs and/or lower insurance premium for
program participation).
To summarize, cost-effective approaches to weight loss and
maintenance are needed for use in health care settings.
Phone and mail-based options are viable options in terms of
cost and logistics, but weight loss outcomes were less than
desired. Strengthened behavioral messages and engagement
strategies are needed to increase the magnitude of weight
losses and their maintenance over time.
References
1 Flegal KM, Carroll MD, Kuczmarski RJ, Johnson CL. Overweight
and obesity in the United States: prevalence and trends, 1960–
1994. Int J Obes Relat Metab Disord 1998; 22: 39–47.
2 Mokdad AH, Bowman BA, Ford ES, Vinicor F, Marks JS, Koplan JP.
The continuing epidemics of obesity and diabetes in the United
States. JAMA 2001; 286: 1195–1200.
3 Ogden CL, Flegal KM, Carroll MD, Johnson CL. Prevalence and
trends in overweight among US children and adolescents,
1999–2000. JAMA 2002; 288: 1728–1732.
4 Flegal KM, Carroll MD, Ogden CL, Johnson CL. Prevalence and
trends in obesity among US adults, 1999–2000. JAMA 2002; 288:
1723–1727.
5 Manson JE, Bassuk SS. Obesity in the United States: a fresh look at
its high toll. JAMA 2003; 289: 229–230.
6 Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW. Bodymass index and mortality in a prospective cohort of US adults.
N Engl J Med 1999; 341: 1097–1105.
7 Paffenbarger R, Hyde R, Wing A, Lee I-M, Jung D, Kampert J. The
association of changes in physical-activity level and other
lifestyle characteristics with mortality among men. N Engl J Med
1993; 328: 538–545.
8 Wilson PWF, D’Agostino RB, Sullivan L, Parise H, Kannel WB.
Overweight and obesity as determinants of cardiovascular
risk: the framingham experience. Arch Intern Med 2002; 162:
1867–1872.
International Journal of Obesity
9 Mokdad AH, Marks JS, Stroup DF, Gerberding JL. Actual causes of
death in the United States, 2000. JAMA 2004; 291: 1238–1245.
10 O’Meara S, Riemsma R, Shirran L, Mather L, ter Riet G. A
systematic review of the clinical effectiveness of orlistat used for
the management of obesity. Obes Rev 2004; 5: 51–68.
11 Arterburn DE, Crane PK, Veenstra DL. The efficacy and safety of
sibutramine for weight loss: a systematic review. Arch Intern Med
2004; 164: 994–1003.
12 NHLBI. Clinical Guidelines on the Identification, Evaluation, and
Treatment of Overweight and Obesity in Adults. NIH, NHLBI:
Bethesda, MD, 1999.
13 Jeffery RW, Drewnowski A, Epstein LH, Stunkard AJ, Wilson GT,
Wing RR et al. Long-term maintenance of weight loss: current
status. Health Psychol 2000; 19 (Suppl 1): 5–16.
14 French SA, Jeffery RW. Consequences of dieting to lose weight:
effects on physical and mental health. Health Psychol 1994; 13:
195–212.
15 French SA, Jeffery RW, Murray D. Is dieting good for you?
Prevalence, duration and associated weight and behaviour
changes for specific weight loss strategies over four years in US
adults. Int J Obes Relat Metab Disord 1999; 23: 320–327.
16 Jeffery RW, Folsom AR, Luepker RV, Jacobs Jr DR, Gillum RF,
Taylor HL et al. Prevalence of overweight and weight loss
behavior in a metropolitan adult population: the Minnesota
Heart Survey experience. Am J Public Health 1984; 74: 349–352.
17 Volkmar FR, Stunkard AJ, Woolston J, Bailey RA. High attrition
rates in commercial weight reduction programs. Arch Intern Med
1981; 141: 426–428.
18 Hellerstedt WL, Jeffery RW. The effects of a telephone-based
intervention on weight loss. Am J Health Promot 1997; 11:
177–182.
19 Cameron R, MacDonald MA, Schlegal RP, Young CI, Fisher SE,
Killen JD et al. Toward the development of self-help health
behaviour change programs: weight loss by correspondence. Can
J Public Health 1990; 81: 275–279.
20 Jeffery RW, Danaher BG, Killen J, Farquhar JW, Kinnier R. Selfadministered programs for health behavior change: smoking
cessation and weight reduction by mail. Addict Behav 1982; 7:
57–63.
21 Jeffery RW, Gerber WM. Group and correspondence treatments
for weight reduction used in the Mulitple Risk Factor Intervention Trial. Behav Ther 1982; 13: 24–30.
22 Black DR, Threlfall WE. A stepped approach to weight control: a
minimal intervention and a bibliotherapy problem-solving
program. Behav Ther 1986; 17: 144–157.
23 Marston AR, Marston MR, Ross J. A correspondence course
behavioral program for weight reduction. Obes/Bariatric Med
1977; 6: 140–147.
24 Linstrom LL, Balch P, Reese S. In person versus telephone
treatment for obesity. J Behav Ther Exp Psychiat 1976; 7:
367–369.
25 US Preventive Services Task Force. Screening for Obesity in Adults.
US Preventive Services Task Force, Agency for Healthcare
Research and Quality: Rockville, MD, December 2003.
26 Jeffery RW, McGuire MT, Brelje KL, Pronk NP, Boyle RG, Hase KA
et al. Recruitment to mail and telephone interventions for obesity
in a managed care environment: the Weigh-To-Be project. Am J
Manag Care 2004; 10: 378–382.
27 Pronk NP, Boucher JL, Gehling E, Boyle RG, Jeffery RW. A
platform for population-based weight management: description
of a health plan-based integrated systems approach. Am J Manag
Care 2002; 8: 847–857.
28 Jeffery RW, Sherwood NE, Brelje K, Pronk NP, Boyle R, Boucher JL
et al. Mail and phone interventions for weight loss in a managedcare setting: Weigh-To-Be one-year outcomes. Int J Obes Relat
Metab Disord 2003; 27: 1584–1592.
29 Moher D, Schulz KF, Altman DG. The CONSORT statement:
revised recommendations for improving the quality of reports of
parallel-group randomised trials. Lancet 2001; 357: 1191–1194.
Mail and phone interventions for weight loss in a managed-care setting
NE Sherwood et al
1573
30 Finkelstein EA, Fiebelkorn IC, Wang G. National medical
spending attributable to overweight and obesity: how much,
and who’s paying? Health Aff (Millwood) 2003; (Suppl): W3219-226.
31 The Diabetes Prevention Program (DPP). Description of lifestyle
intervention. Diabetes Care 2002; 25: 2165–2171.
32 Jeffery RW, French SA. Preventing weight gain in adults: design,
methods and one year results from the Pound of Prevention
study. Int J Obes Relat Metab Disord 1997; 21: 457–464.
33 Baker RC, Kirschenbaum DS. Self-monitoring may be
necessary for successful weight control. Behav Ther 1993; 24:
377–394.
International Journal of Obesity