|Home | About | Journals | Submit | Contact Us | Français|
The purposes of this study were to investigate weight loss expectations and goals in a population sample of US adults who planned to make a weight loss attempt, and to examine predictors of those expectations and goals. Participants were 658 overweight and obese adults (55% female, mean age = 47.9 yr, body mass index [BMI] = 31.8 kg/m2) who responded to a telephone survey about weight loss. Respondents reported weight loss expectations (i.e., reductions they realistically expected) and goals (i.e., reductions they ideally desired) for an upcoming “serious and deliberate” weight loss attempt. They also reported the expectations they had, and the reductions they actually achieved, in a previous attempt. Respondents' weight loss expectations for their upcoming attempt (8.0% reduction in initial weight) were significantly more modest than their goals for that attempt (16.8%), and smaller than the losses that they expected (12.0%) and achieved (8.9%) in their most recent past attempt (ps ≤ .003). Women and participants with higher BMI had greater weight loss expectations and goals. After controlling for BMI, age, and gender, previous weight loss was unrelated to expectations, and inversely related to goals, for the upcoming weight loss attempt. Results suggest that overweight and obese individuals can frame realistic weight loss expectations that are more modest than their ideal goals. BMI and gender appear to be more important than previous weight loss experiences in determining expectations among persons planning a weight loss attempt.
Overweight and obese participants in clinical weight loss studies have been shown to have unrealistic expectations and goals for weight loss (1–5). Weight loss expectations refer to the amount of weight participants realistically expect to lose with the treatment provided, whereas goals reflect the amount they would ideally like to lose (6). Several studies have found that patients enrolled in weight loss treatment ideally wanted to lose about 33% of their initial weight with 1 year of behavioral and/or pharmacological treatment, and thought that losing over 20% was realistic (1,3,6). Women, younger patients, and patients with a higher initial body mass index (BMI) had greater expectations and goals for weight loss (2,6). The weight loss expectations and goals of persons in the general population are less well known.
A recent investigation examined desired weights (from which weight loss goals were calculated) among several cohorts of the Behavioral Risk Factor Surveillance Survey (BRFSS) (7). (The BRFSS is a state-based system of health surveys, sponsored by the Centers for Disease Control and Prevention, that collects information on health-related behaviors and associated constructs.) Overweight (BMI = 25.0–29.9 kg/m2) men and women desired reductions of approximately 5% and 12% of initial weight, respectively. These goals were fairly consistent with consensus guidelines that recommend the loss of 10% of initial weight for overweight persons with weight-related comorbidities (8). These participants' goals also may be considered realistic, given that behavioral and pharmacological programs induce mean reductions of 5%–10% of initial weight (9). Obese persons (BMI ≥ 30 kg/m2) in this study, however, desired far greater reductions of approximately 15% and 24% of initial weight for men and women, respectively (7). Such goals are lower than those frequently reported in clinical weight loss trials but are still beyond what is typically achieved with nonsurgical methods. The BRFSS did not ask respondents how much weight they realistically expected they could lose.
The present study sought to describe weight loss expectations and goals in a large, non-clinical population sample of overweight and obese persons in the US, and to explore determinants of those cognitions. Participants completed a survey that assessed their expectations and goals for a weight loss attempt that they planned to make in the next 6 months. Respondents also indicated, for their most recent weight loss attempt, how much weight they had expected to lose and how much they actually lost. Drawing from clinical studies, we predicted that respondents' ideal weight loss goals for the upcoming attempt would be significantly higher than their realistic expectations. We further hypothesized that expectations and goals for weight loss in the upcoming attempt would be related to reductions that were actually achieved in the previous attempt, such that persons who were more successful in the past would have greater weight loss expectations and goals for their future attempt. Based on previous research, we also anticipated that women, younger participants, and those with higher BMIs would have greater weight loss expectations and goals. Finally, we tested the independent contributions of age, gender, BMI, and previous weight loss experiences to weight loss expectations and goals for the upcoming attempt.
Participants in the present study were respondents to a random digit-dial telephone survey conducted by the Center for Survey Research and Analysis (CSRA) at the University of Connecticut. Respondents were inhabitants of the 48 contiguous states, with stratified sampling (by US Census Bureau geographic regions) to ensure proportional regional representation. A total of 64,025 telephone numbers was dialed from November 2005 to January 2006. Calls were made during the evening, with a maximum of 10 attempts per number. To randomize sampling within a household, the interviewer asked to speak with the adult with the most recent birthday. Interviews were conducted in English or Spanish, according to the respondent's choice, and took an average of 24 minutes to complete. A text telephone system was used, when necessary, to accommodate hearing-impaired respondents. A respondent was contacted in 12,599 households. Of these, 3,500 individuals completed the survey in its entirety, yielding a response rate (i.e., number of completed interviews over an estimate of the number of eligible respondents in the sample [n=18,442]) of 19%. The cooperation rate (i.e., the proportion of completed interviews over the proportion of eligible persons who were contacted) was 30% (10). Respondents were not compensated for their participation.
In addition to demographic information, the survey assessed several aspects of respondents' weight history and weight control practices. A draft survey was piloted in a sample of 232 respondents. Only the items pertinent to the present study are described below. (Data from this survey on the use of dietary supplements for weight loss have been reported by Pillitteri et al. .)
Respondents were asked to report their height and weight, in English units, without shoes. These data were converted to metric units and used to calculate BMI. Participants also categorized their weight status as very underweight, slightly underweight, about the right weight, slightly overweight, or very overweight. Respondents were asked whether they planned to “make a serious and deliberate attempt to lose weight” in the next 6 months and, if yes, what methods they planned to use. They also were asked the following questions in reference to their planned attempt: “Realistically, how many pounds do you think you will actually lose in your next attempt?” and “Ideally, how many pounds do you want to lose?” (emphases in original interview scripts). These responses were defined as weight loss “expectations” and “goals,” respectively.
Additionally, participants were asked if they had ever made a “serious and deliberate” weight loss attempt that lasted more than 3 days. For their most recent attempt, they were asked to report whether the attempt was completed or ongoing, their weight at the start of the attempt, the amount of weight they had expected to lose, and the amount of weight they actually lost (converted to percentage lost). A discrepancy score was computed to represent respondents' achieved reductions in relation to their expectations for their most recent weight loss attempt. This discrepancy is expressed as a percentage (achieved/expected reductions) with values less than 100% indicating smaller-than-expected reductions and values greater than 100% indicating that respondents exceeded their weight loss expectations.
Of the 3,500 survey respondents, 1,844 (55.3%) reported heights and weights that reflected a BMI ≥ 25 kg/m2. Overweight and obese respondents who did not perceive themselves as ever being at least slightly overweight (n = 279), who reported no previous weight loss attempts (n = 435), or who did not plan to lose weight in the next 6 months (n = 208) were excluded. Persons who reported that their most recent weight loss attempt was ongoing (n = 237) also were excluded because reported reductions would not reflect the total weight losses that would be achieved upon completion of the attempt. Additionally, those who reported having had bariatric surgery (n = 10) or who were planning to undergo bariatric surgery (n = 5) were excluded from the present analyses, as the large reductions that can be achieved with bariatric surgery had the potential to bias results regarding expectations and goals. One person who said he ideally wanted to lose zero pounds was excluded. Finally, 11 participants who indicated they did not want to weigh less were not asked about future weight loss expectations and goals and, therefore, were excluded. Thus, the sample for the present study included 658 overweight (n=338) and obese (n=320) individuals who gave information about past and planned nonsurgical weight loss attempts (i.e., 18.0% of all survey respondents). The demographic characteristics of these participants are shown in Table 1.
Data were weighted to mirror the US population in gender, age category, educational attainment, ethnicity, and region. Paired t-tests were computed to test for differences among the following: weight loss expectations and goals for the upcoming attempt, and expected and achieved reductions for the past attempt. Independent t-tests were used to test for gender differences in BMI, expectations, goals, and achieved weight losses. To determine predictors of weight loss expectations and goals, we examined their correlations with demographic variables and previous weight loss attempts. Multiple regression analyses were conducted in which demographic variables were entered simultaneously into the first step of each equation, and previous weight loss experiences were added simultaneously to the second step of each equation. Data were analyzed using SAS software, version 9.1.
As shown in Table 2, respondents had a mean (± standard deviation) expectation of losing 8.0% ± 6.4% of their initial weight in their upcoming attempt. Their goals, of losing 16.8% ± 9.5% of initial weight, were significantly higher than their expectations (p < .0001). Despite having similar mean BMIs (p = .87), women reported significantly (ps < .0001) greater weight loss expectations than men (9.1 ± 6.6% vs. 6.7 ± 5.8%), as well as greater goals (19.7 ± 8.5% vs. 13.4 ± 9.7%). Weight status also was related to expectations and goals, with obese respondents expecting and desiring significantly (ps < .0001) greater reductions than overweight respondents in their upcoming weight loss attempt (values of 9.2 ± 7.8% vs. 6.8 ± 4.5% and 21.2 ± 10.5% vs. 12.1 ± 5.8%, for expectations and goals, respectively). Pearson's correlations showed that age was inversely related to weight loss expectations (r = −.12, p = .001) and goals (r = −.13, p = .001) such that younger respondents has higher targets.
Details of respondents' most recent weight loss attempt are shown in Table 3. In that attempt, respondents reported that they had expected to lose 12.0% ± 8.0% of initial weight but actually lost 8.9% ± 7.2% of their weight (p < .0001). The majority of respondents (62.2%) achieved smaller-than-expected weight losses, whereas 18.5% lost as much as they had expected to lose, and 19.3% lost more than expected.
Women had greater weight loss expectations than men (14.0 ± 8.2% vs. 9.7 ± 6.8%), and achieved greater reductions than men in that past attempt (10.1 ± 7.5% vs. 7.4 ± 6.5%)(ps < .0001). Men and women did not differ, however, in the proportion of their expected weight loss they actually achieved or in the distribution of respondents who lost more than, less than, or exactly as much as, they had expected to lose (ps ≥ .36).
Obese respondents reported having had significantly (p < .0001) greater weight loss expectations than overweight respondents in their previous attempt (13.6 ± 9.6% vs. 10.5 ± 5.8%) and achieving marginally (p = .06) greater reductions (9.4 ± 7.7% vs. 8.3 ± 6.7%). The percentage of overweight and obese respondents who lost more than, less than, or exactly as much as, expected did not differ significantly (p = .52). Furthermore, overweight and obese participants achieved a similar percentage of their expected weight loss (p = .34).
Respondents expected to lose significantly (ps ≤ .003) less weight in their upcoming attempt (8.0 ± 6.4%) than they had expected to lose in their most recent past attempt (12.0 ± 8.0%), or had actually lost in that attempt (8.9 ± 7.2%)
Table 4 shows that BMI (higher), age (younger), gender (female), and expected reductions in the previous weight loss attempt (greater) were related to having more ambitious weight loss expectations for an upcoming attempt, as determined by both zero-order and multivariate analyses. Thus, each of these variables was an independent predictor of the amount of weight respondents expected to lose in their upcoming attempt. The reduction that was achieved in the previous attempt was significantly and positively related to weight loss expectations for the upcoming attempt in the zero-order analysis, but this relationship was not significant in the multivariate model. Thus, the amount of weight actually lost in the previous attempt was not a significant independent predictor of respondents' weight loss expectations for the upcoming attempt, once other factors were accounted for. The discrepancy variable (previously-achieved weight loss divided by previously-expected weight loss) was not related to upcoming expectations in either the zero-order or multivariate analyses.
Women and respondents with a higher BMI had greater weight loss goals for their upcoming weight loss attempt, as determined by both zero-order and multivariate analyses (see Table 4). Age was significantly and inversely related to weight loss goals in the zero-order, but not the multivariate, analysis. The reductions that respondents expected in their previous attempt positively predicted goals for the upcoming attempt in the zero-order and multivariate analyses. Achieved reductions in the previous attempt were positively related to goals for the upcoming attempt in the univariate analysis, but negatively related to goals in the multivariate model. The discrepancy variable was again a nonsignificant predictor in the zero-order and multivariate analyses.
This study is the first to compare weight loss expectations and goals in a non-clinical population sample of overweight and obese persons who plan to lose weight. This is also the first examination of how current weight loss expectations and goals are related to expectations and actual achievement of weight loss in a prior attempt.
We found that respondents' expectations for an upcoming weight loss attempt were significantly more modest than their ideal goals. This finding replicates a pattern that has been found in clinical studies (i.e., expectations < goals)(1,3,6). However, the reductions that respondents expected and desired (8% and 16.8% of initial weight, respectively) are much smaller than the weight loss expectations and goals (~20% and ~33%, respectively) reported by clinical trial participants. This raises the question of what accounts for the apparent difference between clinical and non-clinical samples with respect to their weight loss expectations. Clinical samples tend to have higher mean BMIs than that of the current sample, and obesity severity predicts weight loss expectations and goals (1–6). Additionally, clinical samples typically are predominantly (or exclusively) female, and women have consistently been shown to expect and desire greater losses than men (1–6). The context may also be relevant: the expectations and goals of clinical samples have typically been assessed at the onset of an effort, when motivation and confidence are likely to be at peak levels. Furthermore, the structure of a formal program may increase participants' outcome expectations, relative to their expectations for self-directed weight loss attempts.
We also found that respondents expected to lose less weight in their upcoming attempt than they had expected to lose or actually lost in their most recent past attempt. These mean differences, interpreted in isolation, might suggest that respondents moderated their expectations based on past experience. However, in multivariate analyses controlling for BMI, gender, and age, previously expected and achieved weight losses no longer predicted expectations for the upcoming attempt. Thus, those demographic characteristics were more important than previous weight loss experiences in predicting weight loss expectations.
In contrast, weight loss goals were lower among participants who lost more weight in the past, after controlling for BMI, gender, and age. It may be that the previously achieved weight losses brought respondents closer to their ultimate goal weights, such that smaller reductions would be required to reach target weights in subsequent attempts. This interpretation implies that the previously achieved reductions were well-maintained. Given that significant regain is the norm following weight loss, a more likely explanation for the negative relationship is that those who lost less in the past desired to lose more in the future because they had further to go to reach their target weights.
Our finding that women and respondents with higher BMIs had greater weight loss expectations and goals was consistent with results from clinical studies (2,6) and from the only known population study to examine weight loss goals (7). In our study, BMI and gender appeared to contribute more strongly to goals (βs = .51 and −.27, respectively) than to realistic expectations (βs = .16 and −.11, respectively). The contributions of BMI and gender to weight loss expectations and goals may reflect greater body image concerns and poorer health-related quality of life among women and more obese individuals (12–14). These factors may increase the perceived urgency of weight loss and the desire to obtain greater reductions.
This investigation had the advantage of a large population sample that was weighted to represent US adults. Other strengths are its assessment of expectations as well as goals, and its examination of those constructs in past as well as upcoming attempts. Weaknesses include the use of self-reported heights and weights, and the 6-month timeframe to define an upcoming weight loss attempt. Retrospectively-reported expectations for the past attempt may have been confused with goals for that attempt or biased by actual reductions achieved. Additionally, analyses did not include several modifiable (e.g., body image) and non-modifiable (e.g., family history of obesity) factors that may be related to weight loss expectations and goals.
Should providers encourage overweight and obese individuals to adopt more realistic weight loss expectations and goals? The answer may depend on the setting. When treatment is provided for free, as in most clinical studies, the answer appears to be “no.” Having, and failing to meet, unrealistic targets did not increase the risk of dropout or adverse psychosocial outcomes under those conditions (1,3,5,6,15–17; see 18 for a notable exception). When people must pay for treatment, however, it is reasonable to assume that they may be more likely to discontinue treatment if they do not achieve a result that they consider to be realistic. Although data from commercial weight loss programs are not available, a large study from 23 medical centers – where patients paid for treatment – found greater attrition among persons who desired and expected the greatest reductions (19). Thus, it may be advisable for providers in those settings to encourage modest expectations and goals. We note, however, that it has proved quite difficult to alter patients' perceptions about what is a reasonable weight loss (3,16, 20).
In summary, the weight loss expectations of overweight and obese individuals in the general population appear to be more modest than those observed in clinical settings. Furthermore, respondents expected weight losses that can realistically be achieved with nonsurgical methods. Our findings replicate previous research that found that women and those with a higher BMI realistically expect, and ideally want, to lose more weight. The relationship of age to expectations and goals was less consistent. Our results also suggest that weight loss expectations become more modest in subsequent weight loss attempts. Those expectations, however, are largely unrelated to previous success, suggesting that BMI and gender are stronger determinants of what is a reasonable weight loss to expect.
This study was additionally supported by NIH grants K23 DK070777 and K24 DK065018.
Disclosure The survey that yielded study data was conducted by the Center for Survey Research and Analysis at the University of Connecticut and supported by GlaxoSmithKline Consumer Healthcare. Authors Harkins and Burton are employed by GlaxoSmithKline. Rohay, Pillitteri, and Shiffman perform consulting services on behalf of GlaxoSmithKline.