Participants in weight loss programs typically set unrealistically high weight loss goals that some believe are detrimental to success. This study examined outcomes associated with goal and ideal body mass index (BMI). Participants (N=1801) were enrolled in a weight loss trial comprised of low-intensity mail or telephone interventions vs usual care. Goal and ideal weight losses were assessed by asking participants how many pounds they expect to lose in the program (goal) and how much they would like to weigh (ideal). Goal and ideal weight losses were unrealistically high (men: −16 and −19%, women: −21 and −27%). For women, less realistic goals were associated with greater weight loss at 24 months. Goals were not associated with participation or weight loss for men. Results are more supportive of the idea that higher goals motivate women to lose weight than of the hypothesis that high goals undermine effort.
Many obesity treatment programs recommend weight loss goals of 5–10% of body weight, in the belief that such goals are achievable and clinically valuable.1, 2 Program participants typically select much higher goals, however.3, 4, 5, 6, 7 Although some argue that these unrealistic goals undermine treatment success, empirical evidence for this found in a search of online resources (eg, PsycInfo, Medline), supplemented by citations from our earlier work in this area7 was mixed.3, 4, 6, 7, 8, 9 To further clarify this issue, the present study examined the relationship between weight loss goals, program participation, and weight loss over 2 y in a large sample of men and women.
Data are from Weigh-to-Be, a collaboration between the University of Minnesota and HealthPartners, a large Minnesota managed care organization (MCO). The study enrolled 1801 health plan members (508 men, 1293 women) recruited by flyers, Internet advertisement, or physician referral. Participants were randomized to mail intervention, telephone intervention, or usual care. Mail and phone interventions were offered over 24 months, although participation was largely limited to the first 12 months. Participants in usual care were offered weight loss services available to all health plan members. Procedures and weight loss outcomes are described elsewhere.10, 11
Measures included in this report are demographics and measured height and weight assessed at baseline, self-reported weight at 12 months, measured weight at 24 months, and number of intervention lessons completed in treatment (range=0–20). Based on average discrepancies between self-reported and measured baseline weights, 1.5 and 1.7 kg were added to men's and women's 12–month weights, respectively.12 Weight changes were expressed in body mass index units (BMI; kg/m2). Weight goals were calculated in two ways: goal weight loss from ‘How many pounds do you expect to lose in the Weigh-to-Be program?’ and ideal weight loss from ‘How much would you like to weigh?’ Both were expressed in BMI units, with higher values indicating smaller, more realistic weight losses, and lower values indicating greater, less realistic weight losses. Program participation was assessed by number of intervention lessons completed by mail or phone during the 24–month study period.
Statistical analyses were conducted using the Statistical Analysis System (SAS), Version 8.2.13 General Linear Models were used to analyze associations of outcomes with baseline, goal, and ideal BMI for men and women. Baseline BMI was entered as a covariate in goal and ideal BMI analyses, and treatment group was entered as a covariate in outcome analyses. Effect sizes were calculated using Cohen's d statistic (small=0.20, medium=0.50, large=0.80).14
Table 1 presents descriptive statistics on demographics, baseline, goal, and ideal BMI, weight loss in treatment, and program participation. As expected, weight loss goals were unrealistically high (16% in men, 21% in women). At 12 months, 4.9% of men (n=19) and 2.8% of women (n=26) had reached or surpassed their goal weight, and 5.6% of men (n=22) and 2.1% of women (n=20) had reached or surpassed their desired weight. By 24 months, 6.5% of men (n=19) and 3.1% of women (n=22) had reached or surpassed their goal weight, and 3.4% of men (n=10) and 1.9% of women (n=13) had reached or surpassed their desired weight.
Table 2 presents associations between baseline, goal, and ideal BMI and: (a) lesson completion through 12 months, and (b) weight loss at 12 and 24 months. The only predictor of lesson completion was lower baseline BMI in women (d=0.22). Lesson completion predicted weight loss at 12 months in both men (d=0.32) and women (d=0.35), and 24–month weight loss in men (d=0.32). For men and women, greater 12- and 24-month weight losses were significantly associated with higher baseline BMI (12 months: d=0.24 and 0.32; 24 months: d=0.51 and 0.49). Three of the eight associations between weight loss goals and actual weight loss were statistically significant. At 12 months, less realistic ideal weight losses predicted greater weight loss in both men (d=0.21) and women (d=0.16). In women only, greater 24-month weight losses were associated with less realistic goal weight loss (d=0.28). Addition of psychosocial or psychiatric variables known to be associated with weight outcomes in this sample (self-reported depression status, self-reported binge eating status, and weight control self-efficacy)15 as covariates in these models did not change the magnitude or significance of the findings for baseline, goal, or desired BMI; the analyses without these covariates are presented here.
This study examined prospective associations between weight goals and treatment outcomes. Strengths of this study are the inclusion of a greater percentage of men (28%) and the large sample size. Weaknesses are a low intensity weight loss program that may limit generalization to higher-intensity weight loss programs, and lack of measured weights at 12 months.
The most important finding here was that the predictive value of weight loss goals for actual weight loss was modest across the 24–month study, and that all significant effects were in the direction of larger, less realistic goals predicting greater actual weight loss. Although effect sizes were small to medium,14 as has been the case in other published empirical examinations,6, 7, 8, 9 the evidence suggests that unrealistically high goals do not discourage and may encourage weight loss, independent of any effect of psychosocial or psychiatric variables that may negatively influence weight change, such as binge eating, depression, or self-efficacy.15 The evidence does not convincingly support the prevalent clinical practice of counseling people to adopt modest weight loss goals.
The pattern of goal weight losses being more realistic than ideal weight losses seen in this study is consistent with previous responses to similar items.3 The observation that men's goal and ideal weight losses were significantly more realistic than those of women in this study is a new finding. Women's goals and ideals were more consistent with data from previous studies;3, 5, 16 however, previous study samples were comprised of 80–100% women.
It is noted that if attained, the average goal BMI for men and women and average ideal BMI for men in this study (range=26.80–27.83 kg/m2), would still classify participants as overweight. Similarly, the lowest goal or ideal BMI reported here, that of average ideal BMI for women (mean=24.75 kg/m2), would classify participants in the normal range. Thus, obese participants in this sample (mean BMI=33.10 kg/m2 for men, 33.86 kg/m2 for women) are not so unrealistic as to desire excessively thin, unhealthy body weights as a result of weight loss. In a general sense, participants' expectations were realistic, as they involved medically feasible weights.
Regardless of gender, treatment participation was positively associated with weight loss. Greater participation previously has been linked with greater weight loss.17 Interestingly, for women, greater participation showed a modest association with baseline BMI: those with lower weights at baseline were likely to have completed more lessons through 12 months. As declining participation is a concern in weight management,18 this finding warrants attention.
Overall, our results augment the literature suggesting that unrealistic goals are not detrimental to weight loss.6, 7, 9 As long as treatment outcomes remain imperfect, more benefit may be derived from focusing on factors known to improve outcomes, such as self-monitoring or physical activity18, 19 than by focusing on perceived harmful effects of weight loss goals.
This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant 1R01-DK53826.