A Pilot Internet-Based Behavioral Weight Loss Intervention with or without Commercially Available Portion-Controlled Foods


  • Funding agencies: This study was funded by a grant from the Obesity Society and Nutrisystem, Inc.

Correspondence: Kelly H. Webber (kelly.webber@uky.edu)


Objective: To evaluate the short-term impact of portion-controlled food provision in combination with an Internet behavioral weight loss program on weight, blood cholesterol, and blood glucose levels.

Design and Methods: Fifty participants, mean age 46 ± 10.7 years and mean body mass index 35.1 ± 3.8 kg/m2, were randomized to one of two study groups, an Internet behavioral weight loss program (Internet-alone; n = 25) or an Internet behavioral weight loss program plus a commercially available portion-controlled diet (Internet + PCD; n = 25) for 12 weeks.

Results: An intent-to-treat analysis found that the mean weight change in the Internet + PCD group was −5.7 ± 5.6 kg and in the Internet-alone group (n = 25) was −4.1 ± 4.0 kg (P = 0.26). Participants in the Internet + PCD group achieved significantly greater improvements in blood glucose (−2.6 ± 5.7 vs. 1.4 ± 11.0 mg/dl; P = 0.05) and LDL cholesterol (−8.2 ± 18.0 vs. −0.6 ± 21.0 mg/dl; P = 0.04), compared with Internet-alone group.

Conclusions: These data suggest that there may be short-term clinical benefit in using a PCD in conjunction with a behavioral Internet-based weight loss program to enhance weight loss and improve health indicators.


Currently 68% of U.S. adults are overweight or obese [1]. Excess weight increases the risk for several chronic diseases including cardiovascular disease, stroke, and type 2 diabetes [2] and as a result has high economic costs [5]. Weight loss of 5-10% of initial body weight has been shown to decrease the risk factors for chronic diseases such as type 2 diabetes and cardiovascular disease [2]. Modest weight loss has also been shown to reduce healthcare costs [6].

Currently, the most successful nonsurgical treatments for weight loss are behavioral face-to-face programs; however, there are drawbacks to the face-to-face format [7]. These face-to-face clinic-based treatments are effective but costly, require patients to travel to a clinic, reach a limited audience, and do not provide a choice of formats for patients. In an effort to overcome some of the limitations of face-to-face programs, behavioral Internet-based weight loss programs have been developed and evaluated [8]. These Internet programs provide many benefits including convenience, privacy while still allowing for interaction with peers or counselors, decreased cost, and a potential to reach large audiences [16]; however, they have not been as effective as face-to-face behavioral programs in producing weight loss [15].

There is a need to improve the effectiveness of these Internet behavioral weight loss programs. Research has shown that adding structure to a low-calorie diet increases adherence and weight loss [19]. In addition, a 2003 meta-analysis found that the provision of portion-controlled foods as part of a low-calorie weight loss program produced an average of 2.5 kg more weight loss than conventional reduced calorie diets over 3 months of intervention [20]. Therefore, one possibility for improving weight loss in Internet-based programs is with the addition of portion-controlled foods.

To our knowledge, the comparison of an Internet intervention plus a portion-controlled diet to an Internet-intervention alone has not been conducted. This study compared the efficacy of a 12-week Internet-based behavioral weight loss program, prescribed with a diet of self-selected conventional foods (i.e., Internet-alone group), to the same 12-week program supplemented with a commercially available portion-controlled diet (PCD) for producing weight loss (Internet + PCD). In addition, this study compared the impact of the two programs on chronic disease indicators including blood pressure, blood lipids, and blood glucose levels.

Methods and Procedures


Males and females between the ages of 25 and 65 years, with a body mass index (BMI) of ≥30 and ≤45 kg/m2, and a home computer with access to the Internet were recruited through advertisements in the local newspaper. Exclusion criteria included: a medical diagnosis of orthopedic or joint problems that may prohibit regular exercise; reported heart problems, chest pain, faintness, or dizzy spells; hospitalization for a psychiatric disorder within the last year; a history of anorexia or bulimia nervosa; blood pressure >160/100; a medical diagnosis of cancer (except skin cancer) or HIV; pregnant, nursing, or planning to become pregnant within the study period; less than 9 months postpartum; weight loss of ≥10 pounds in the last 6 months; on more than one medication for hypertension control; taking glipizide, glyburide, insulin, or any agent to control diabetes that could cause hypoglycemia; taking warfarin; taking a diuretic with or without potassium supplementation; and food allergies.

Study design

All participants were screened initially via telephone and eligible subjects were invited to a study information session. All participants also provided a signed physician's consent form at the information session. At the study information session, interested participants were asked to sign a consent form and to schedule a baseline assessment at the laboratory. The laboratory assessment included measurement of height, weight, waist circumference, blood pressure, and a fasting blood draw. Blood was analyzed for total cholesterol, LDL-C, HDL-C, and blood glucose.

After baseline assessments were completed, participants were randomized to one of two study groups, Internet-based behavior therapy alone (Internet-alone; n = 25) or in combination with a PCD (Internet + PCD; n = 25). Participants were asked to return for a follow-up assessment visit at the laboratory at 12 weeks. All study procedures were approved by the University of Kentucky Institutional Review Board.


Each group received a separate face-to-face 2-h weight loss session at the beginning of the study led by a registered dietitian. The sessions took place in a classroom at the university and included a presentation of basic research-based information on weight loss and an orientation to the study website. During the sessions, an emphasis was placed on self-monitoring of diet, exercise, and weight and research to support these recommendations. All participants received a calorie book, references to online websites with diet and physical activity calorie information, and self-monitoring diaries to support these recommended behaviors. Participants were asked to keep paper diaries of their daily caloric intake and expenditure and then to enter that information, along with a weekly weight, into the online self-monitoring form once each week. Information from the National Weight Control Registry was also presented.

A gradual increase in exercise, in keeping with guidelines from the American College of Sports Medicine [21], was recommended for all participants. Participants who were not currently exercising were told begin with exercise bouts of 10-20 min on 2-3 days per week, participants who were exercising sporadically were encouraged to increase their physical activity to 30 min on 5 days per week, and participants who were already exercising regularly were encouraged to continue their current habits. Locally available resources for physical activity were discussed among the groups.

Calorie intake targets were intended to induce a weight loss of 1-2 lb per week and were the same in both groups (i.e., ∼1,200 and 1,500 kcal/day for women and men, respectively). Participants in the Internet-alone group were provided instruction on reducing calorie intake with self-selected foods, whereas those in the Internet + PCD group were encouraged to follow the Nutrisystem weight loss program, which was provided to participants at no cost. This program includes portion-controlled entrees and snacks and a meal plan for supplementing those items with grocery foods (e.g., vegetables, fruit, and dairy). Finally, all participants were introduced to the idea of beginning to work on reasons for overeating that were not due to a lack of knowledge, but instead due to reasons commonly referred to as “emotional eating.”

At the conclusion of the baseline face-to-face session, participants were given access to the password-protected study website. The website contained weekly weight loss tips, weekly lesson postings, a message board feature, and links to self-help diet, exercise, and behavioral modification resources available on the web. The site also had a link to a personal online self-monitoring report form which participants were asked to use to report, at least weekly, daily caloric intake and daily exercise. Participants could also record weight in the online diary, which was then graphed on the website. The website was identical for the two groups. Topics for the 12-week behavioral weight loss lessons were similar to the core of the Diabetes Prevention Program (DPP) [22]. The DPP lessons were developed for use in a face-to-face individual setting where counselors interacted with participants and provided problem solving and support verbally. These lessons were modified for use in this study to be more self-directed and suitable for posting to the website and have been used in previous studies [11, 13]. All participants received the same basic weekly email from the study coordinator to remind them of the new weekly lessons on the website.

Participants in the Internet + PCD group received 2 weeks of shelf-stable foods (three entrees and one to two snacks per day) at the baseline session. For the remainder of the 12-week study, foods were delivered directly to participants' homes.

At the conclusion of the study, participants were contacted for scheduling a final laboratory assessment. Each participant received $40 for returning to the final study assessment to ensure adequate follow-up rates. Participants in the Internet-alone group also received 12 weeks of Nutrisystem program, at no charge, upon conclusion of the study.

Outcome measures

Body weight, height, waist circumference, and blood pressure were measured in the clinic at baseline and 12 weeks by a nurse blinded to group assignment. Weight measurements were taken with a Tanita digital scale, model # BWB-800 (Arlington Heights, IL). Participants wore a hospital gown with no shoes for weight and height measurements. Waist circumference was measured at the umbilicus. Blood pressure was measured once in the seated position. A blood sample was also taken to be analyzed for blood glucose, total cholesterol, LDL-C, and HDL-C.

Motivation for weight loss was assessed at baseline using the autonomous and controlled regulation subscales of the Treatment Self-Regulation Questionnaire (TSRQ) for weight loss treatment [23].

Self-monitoring was measured by online food and exercise diary forms. Completion of at least 5 days of monitoring was required for a participant to receive credit for monitoring in a given week. The number of logins to the website was also monitored and recorded. Participants in the Internet + PCD group were asked, at the final visit, approximately how many times they had consumed the PCD foods; the possible responses were as follows: every meal, almost every meal, most of the time, about half of the time, less than half of the time, and not at all. The responses were scored on a scale of 1-6, with 1 indicating greatest adherence to the PCD.

Statistical analysis

This study was designed to have 80% power to detect a between-groups difference of 3.0 kg with a pooled standard deviation of 4.0 kg and α = 0.05 on the primary outcome of weight loss [19, 20, 24]. All analyses were performed using Windows version 16.0 of the Statistical Package for the Social Sciences (SPSS, Chicago, IL). Descriptive statistics were used to characterize both study groups at baseline. Categorical variables were analyzed using Pearson χ2 analysis.

The primary analysis was an intent-to-treat repeated measures analysis of variance to test for a between-groups difference in weight change. Baseline values were carried forward in the case of outcome data. Similar intent-to-treat analyses were conducted to test for differential changes in secondary outcomes. Difference in attrition between groups was examined using a Fischer's exact test. The relationship of baseline and process variables to weight loss were first examined using the Pearson R coefficient. Cohen's d was calculated to estimate the effect size. Alpha was set at 0.05 for all tests.



There were 50 participants enrolled in the study. Follow-up data were available for 47 participants, for a completion rate of 94% (Figure 1). At baseline, the mean age of all participants was 46 ± 10.7 years. The sample was 86% female and 76% Caucasian. Most participants (62%) had completed a college degree or higher. Participants had an average BMI of 35.1 ± 3.8 kg/m2 and an average waist circumference of 109.9 ±12.4 cm. The two groups did not differ on any baseline characteristic (Table 1).

Figure 1.

Participant flow chart.

Table 1. Baseline characteristics of study participants
 Internet-alone group n = 25 Mean (SD)Internet + PCD group n = 25 Mean (SD)P-value
Age47.6 (9.5)44.4 (11.7)0.28
Gender92% Female80% Female0.22
Race76% Caucasian76% Caucasian0.57
Education60% College graduate or greater64% College graduate or greater0.37
Autonomous motivation5.9 (0.8)6.0 (0.8)0.91
Controlled motivation3.4 (0.9)3.5 (0.8)0.74
BMI35.2 (3.9)35.0 (3.8)0.90
Waist circumference (cm)109.2 (11.2)110.7 (13.7)0.66
Fasting glucose (mg/dl)98.8 (19.0)98.2 (12.5)0.90
Total cholesterol (mg/dl)202.2 (30.4)197.0 (29.4)0.54
HDL-C (mg/dl)55.0 (18.5)52.6 (16.3)0.62
LDL-C (mg/dl)121.5 (34.8)120.4 (20.3)0.88
Systolic blood pressure (mmHg)129.5 (11.8)126.7 (11.7)0.40
Diastolic blood pressure (mmHg)82.8 (7.2)82.2 (9.6)0.80

Final measurements were completed by 100% of participants in the Internet-alone group, and 88% of those in the Internet + PCD group. Attrition was not significantly different between groups (P = 0.23).

Clinical outcomes

In the intent-to-treat analysis, the mean weight reductions observed in the Internet + PCD and Internet-alone groups were 5.7 ± 5.6 and 4.1 ± 4.0 kg, respectively. The between-groups difference was not statistically significant (P = 0.26) despite an estimated effect size (d) of 0.33. In the completers analysis, the mean weight loss in the Internet + PCD group increased to 6.5 ± 5.5 kg and the estimated effect size (d) increased to 0.50, but the between-groups difference remained nonsignificant (P = 0.10). A similar percentage of participants who completed both treatments (52% of the Internet-alone group and 56% of the Internet + PCD group) achieved at least a 5% reduction in initial weight (P = 0.31).

As shown in Table 2, participants in the Internet + PCD group achieved significantly more favorable changes in blood glucose (−2.6 ± 5.7 vs. ±1.4 ± 11.0 mg/dl, P = 0.05) and LDL cholesterol (−8.2 ± 18.0 vs. −0.6 ± 21.0 mg/dl, P = 0.04) than those in the Internet-alone group. Participants in the Internet + PCD group (−13.2 ± 29.9 mg/dl, P = 0.01), but not in the Internet-alone group (−6.4 ± 42.8 mg/dl, P = 0.18), achieved a significant reduction in total cholesterol; however, the change was not significantly different (P = 0.32). No other between-group changes were statistically significant.

Table 2. Mean changes in outcome measures from baseline to 12-week follow-up
Outcome measureInternet-alone group Mean (SD) n = 25 (P-value for within-group change)Internet + PCD group Mean (SD) n = 25 (P-value for within-group change)P-value for between-group differences
Weight (kg)−4.1 (4.4) (P < 0.001)−5.7 (5.6) (P < 0.001)0.26
Waist circumference (cm)−5.8 (5.3) (P < 0.001)−6.9 (5.3) (P < 0.001)0.48
Systolic blood pressure (mmHg)−0.7 (10.8) (P = 0.76)+1.6 (10.7) (P = 0.45)0.45
Diastolic blood pressure (mmHg)+1.6 (6.5) (P = 0.22)0.0 (5.8) (P = 0.97)0.36
Blood glucose (mg/dl)+1.4 (11.0) (P = 0.54)−2.6 (5.7) (P = 0.04)0.05
Total cholesterol (mg/dl)−6.4 (42.8) (P = 0.18)−13.2 (29.9) (P = 0.01)0.32
LDL cholesterol (mg/dl)−0.6 (21.0) (P = 0.89)−8.2 (18.0) (P = 0.03)0.04
HDL cholesterol (mg/dl)−2.0 (6.0) (P = 0.12)−2.6 (7.0) (P = 0.08)0.75

Process measures

Program usage was assessed through website visits and the number of self-monitoring diaries completed. The Internet-alone group visited the study website an average of 20.1 times, whereas the Internet + PCD group visited the website 15.2 times (P = 0.14) over the 12-week study. The number of visits to the website was not associated with weight loss in either group (Internet-alone: r = 0.37, P = 0.07; Internet + PCD: r = 0.35, P =0.11). The average number of weekly diaries completed by the Internet-alone group was 8.9 and the Internet + PCD group completed an average of 7.8 diaries (P = 0.35). The number of diaries completed was associated with weight loss in both groups (Internet-alone: r = 0.42, P = 0.04; Internet + PCD: r = 0.52, P = 0.01). In the Internet + PCD group, 77% of participants reported eating the PCD at every meal, almost every meal, or most of the time. More consistent consumption of prepackaged foods consumed during the study was found to correlate with weight loss (r = 0.45, P = 0.04).


This pilot study sought to determine if supplementing a 12-week Internet behavioral weight loss program with a portion-controlled diet would produce greater weight loss than a 12-week Internet behavioral weight loss program with a self-selected diet. Results indicate that both groups lost a significant amount of weight. Although not significant in this small sample, the addition of PCD was associated with a 1.6 kg increase in mean weight loss. There were significant advantages for Internet + PCD vs. Internet-alone with respect to blood glucose and LDL cholesterol.

Weight loss in both groups was greater than some previously reported short-term weight loss studies. Face-to-face 16-week behavioral weight loss programs have produced weight losses ranging from 2.9 to 5.5 kg [25], whereas previous 16-week Internet behavioral weight loss programs have produced weight losses ranging from 3.6 to 4.5 [11, 13]. It appears that in this small short-term study that the average weight loss achieved for both groups was higher than previous Internet behavioral weight loss programs and within the range of what has been found for 16-week face-to-face programs, whereas the weight loss achieved by the Internet + PCD group was higher than previous 16-week face-to-face programs. The differences in weight loss seen in this study could have been due to the differences in populations between studies.

These weight loss results were also comparable to what has been found in previous studies using PCD. A 2003 meta-analysis of the use of meal replacements for weight management found a pooled difference of 2.5 kg between the groups that used PCD and those that did not [20], which is similar to the 2.4 kg difference in weight loss between the two groups found in the completers analysis of this study. Previous use of a PCD in a face-to-face behavioral weight loss program found weight loss at 3 months to be 8.2 ± 5.2 kg [30]. This is higher than the weight loss found at 3 months in this study; however, face-to-face programs typically produce greater weight losses than Internet-based studies [7, 15]. Finally, data from the provider of the PCD used in this study indicate that the average weight loss among customers who self-reported their weight online at the point of purchase and again after 3 months was 8.7 ± 6.6 kg [31]. However, these are completers data only and there are no published estimates of completion rates. Therefore, this value may not be representative of all persons who started the program.

Results also indicated that increased consumption of the PCD was correlated with increased weight loss. Raynor et al. found that subjects who received single serving size packages of food for breakfast consumed fewer calories than those who did not receive portion-controlled servings of the same foods [32]. This finding suggests that PCDs may help some subjects to eat fewer calories and to subsequently lose weight. The additional weight loss seen in the Internet + PCD group was likely due to the provision of a PCD; however, further research is needed to confirm these results.

The Internet + PCD group saw a significant improvement in the chronic disease risk factors of fasting blood glucose and cholesterol levels. The decrease in glucose, total cholesterol, and LDL-C is similar to what has been found in other weight loss studies [33]. HDL-C did not show an improvement in this study, which is similar to what has been found in a previous review of weight loss trials [36].

Both groups were engaged with the program at a similarly high level, as indicated by the number of logins to the study website and the number of online diaries completed. As in previous studies, the number of self-monitoring diaries completed was correlated with weight loss in both groups. This is commonly expected and possibly related to greater mindfulness in the amount of calories consumed and expended each day.


The primary limitation was the lack of statistical power to detect significant differences in this small pilot study. The observed difference in weight loss was slightly smaller, and the pooled standard deviation was larger, than the assumptions of our power analysis (2.4 vs. 3.0 kg and 4.7 vs. 4.0 kg, respectively). Furthermore, the small sample was generally well educated and healthy, and predominantly white; therefore, these study results may not be generalizable to other populations. Another limitation to the study was the short time frame of 12 weeks. Previous studies have found that continued use of PCD has a continuing positive effect on weight loss in face-to-face settings [20, 37, 38]. Whether a longer-term effect of additional dietary structure would also be found in an Internet-based intervention has yet to be examined. Finally, potentially important process variables (e.g., energy expenditure) were not assessed.


In the short term it appears that a PCD in conjunction with an Internet behavioral weight loss program may aid in weight loss. Although the between-groups difference was not statistically significant, the effect size was moderate. This combination also has a favorable impact on the health indicators of fasting blood glucose and cholesterol levels.

Future directions

Replication of this research using an adequately powered study is needed. Considering the effect size of the intent-to-treat analysis, a much larger sample is needed, 230 total participants vs. the 50 in this study. Long-term studies of the impact of portion-controlled foods on weight loss and health indicators are also needed. Future research could investigate the long-term effectiveness and cost-effectiveness of both Internet and face-to-face behavioral weight loss programs with and without adjunctive PCD. Future research could also investigate the efficacy of an Internet-based behavioral weight loss program supplemented with portion-controlled foods on a population with greater health risks including those with impaired fasting glucose, type 2 diabetes, or metabolic syndrome. Because both the Internet and commercial weight control programs are widely available, the successful combination of these two interventions may have significant public health benefits.


This study was supported by a grant from The Obesity Society, using a mechanism funded by Nutrisystem, Inc. Nutrisystem, Inc. also donated the portion-controlled diet used in this study. Nutrisystem, Inc. had no role in designing the study or selecting the grant application for funding.