Predictors of attrition from a weight loss program. A study of adult patients with obesity in a community setting

Definition of attrition

Patients who could not participate in a scheduled visit were provided with a new appointment. If the patient did not attend the visit, he/she was contacted by telephone and a new appointment was offered. Patient who did not attend a scheduled appointment and either did not answer the follow-up call or declared that he/she no longer wants to participate in the program was considered a dropout.

We defined patients who dropped out early as those who did not complete the initial 12 weeks of treatment (performed < 3 visits) and did not re-engage; we defined patients who dropped out late as those who disengaged from our weight loss program after the first 3 months and did not re-engage. Re-engaging patients were those who dropped out (at any time) and then requested to be included again in the program. Finally, we defined completers as those who missed no more than one meeting and attended the final meeting, while attenders were defined as those who missed > 1 meeting and attended the final meeting.

Measurements

Weight (measured with the participants wearing light clothes and no shoes to the nearest 0.1 kg by a mechanical column scale, SECA model 711, Hamburg, Germany), height (measured to the nearest 0.1 cm with a Stadiometer, SECA 220 measuring rod, Hamburg, Germany), and waist and neck circumferences (measured by a plastic tape meter at the umbilicus level or under the cricoid cartilage, respectively) were assessed during each visit.

At baseline, all patients completed the following questionnaires:

• State-Trait Anxiety Inventory (STAI) for both trait (STAI-1) and state (STAI-2) anxiety [19];

• Binge Eating Scale (BES): a binge eating disorder was considered unlikely with scores 0–16, possible with scores 17–27 and likely with scores > 27 [20];

• -Short-Form (SF)-36 health survey: used to measure general health-related quality of life; the SF-36 has different domains with scores ranging from 0 to 100 (higher values indicate a better quality of life, a value < 50 corresponds to low health status for each domain [21].

• The above psychometric questionnaires were checked and assessed by a trained psychologist.

• Arterial blood pressure values were measured on the left arm, in a sitting position, after at least 10 min of rest, with a mercury sphygmomanometer with appropriate cuff sizes (ERKA Perfect-Aneroid, Germany). Two measurements were taken with the arm supported at heart level and the values reported were the means of the two. Blood samples were collected after an overnight fast. Laboratory methods have been previously described [22].

Statistical analyses

The t-test for paired data was applied to investigate within-group variations; the t-test for independent samples was performed to assess between-group differences. Because the distributions of the BDI score and anthropometric changes were skewed, the Mann–Whitney test was employed to evaluate between-group changes. A multiple logistic regression model, using attrition as the dependent variable, was used to evaluate the association between attrition and the variables of interest.

Attrition

Our finding of ~ 50% attrition is in line with the literature, since a high rate of attrition to weight-loss programs is common to most obesity units, with rates up to 80% after starting treatment [1–3, 13]. The dropout rates increased with the duration of follow-up, ranging from 20 to 30% in the first 4 months [5, 23] to 85% at 36 months [2, 24]. Accordingly, our attrition rates were 18.8%, 35.2%, 49.2% and 53.6% at 3, 6, 9, and 12 months, respectively.

Many different factors have been previously reported as potentially responsible for attrition with inconsistent results, including socioeconomic, demographic, weight-related and/or psychological conditions, such as depression, anxiety, alexithymia, presence of binge eating, body dissatisfaction, lack of motivation, unrealistic weight-loss expectations, low goal ownership, lower self-esteem, high levels of stress, and difficulty in coping with emotions [2, 3, 5–14, 25–30]. However, the data were highly contrasting [1, 14, 15], and the definitions of dropout greatly varied among the authors [4, 8, 16]. Furthermore, many of these studies were conducted within the context of randomized controlled trials or lifestyle intervention studies [1, 7, 10–12, 17, 30–32]; thus their results are not generalizable to a real-world setting.

We found a strong protective effect of BMI on attrition. Highly conflicting data are available on this topic [1]; no association [5, 7, 13, 15, 33, 34], a direct relationship [11, 25] and an inverse [23, 28, 35, 36] relationship have been reported. Individuals with higher BMI may be more motivated to lose weight; furthermore, healthcare professionals may pay more attention to the active involvement of patients with more severe obesity. Accordingly, the presence of comorbidity has been reported to cause greater adherence to weight-loss treatments, and completers were characterized by increased concern for one’s future or present health [13, 26, 35].

The weight loss achieved by our patients was in line with studies evaluating the effects of weight-loss programs in real clinical practice [6, 13, 15, 28, 37]. Weight loss was lower but not significantly different in late dropouts when compared to completers. Indeed, the differences between the two groups increased with time, and early weight loss has been suggested to be one of the factors motivating participants to continue the weight-loss program by many studies [6, 10, 13, 23, 28, 31, 38]. On the other hand, weight loss has been reported to be directly associated with the length of time that the individual remained in the weight-loss program [7]. We cannot exclude that the between-group difference could reach statistical significance in a larger sample; however, weight loss probably represents a less important factor in influencing attrition in our patients.

Our results suggest the importance of psychological distress (in particular, depression) in predicting attrition in weight-loss programs. Other authors have found that the presence of depression symptoms increased the risk of attrition [1, 9, 25, 31, 36]; an Italian study obtained the opposite result [35], and others reported no associations [14, 32].

Our participants had not been formally diagnosed with clinical depression; indeed, even mild scores of depressive symptoms could have an impact on attrition. Depression and attrition could be linked for several reasons, including patient lack of energy, impacting weight-loss compliance, the tendency to adopt maladaptive coping strategies (e.g., rumination, expressive suppression [39]), and the lower propensity to be engaged in the therapeutic alliance [40]. Further studies should investigate these clinically relevant topics.

Early vs late attrition

Many studies had short follow-up (approximately 6 months) and/or did not differentiate between early and late attrition [7, 14, 15, 28, 33–35]. This makes their results difficult to compare. Older studies [16] suggested that individuals in the early attrition group had fewer prior attempts to lose weight, a higher external locus of control, and lower functional scores for depression, breadth of interest, organization, and responsibility. Accordingly, patients in our early dropout group showed specific characteristics that are suggestive of greater frailty, such as living alone and decreased mental well-being.

Living alone was a predictor of early dropout, in line with some authors [37], but differently from others [5, 13]. Living alone could be a proxy for multiple conditions including depression, the possible presence of other unreported psychiatric conditions, and being less motivated to cook healthy foods. However, more data would be necessary to draw more firm conclusions on this finding.

Decreased mental well-being was a significant risk factor for early attrition among our patients. Intriguingly, a questionnaire assessing many psychological aspects was more predictive of dropout risk than those assessing a single mental state, thus suggesting that the mental discomfort of those individuals was a complex mix of emotional, personality and social factors, rather than being characterized by specific symptomatology. This was in line with studies showing that scores combining depression, anxiety, social dysfunction, and somatic symptoms were better predictors of early attrition [8] and could justify the highly divergent results of previous studies that focused on only a few specific traits [1]. Greater mental stress can make it difficult to conduct many activities or handle different situations, such as traveling to the medical clinic, attending visits, interacting with specialists, comparisons with other participants, contact with new people, the change of one's lifestyle habits and tolerating the distress of potential failure. Our findings suggest the need for a broad assessment of mental health in these patients.

Early attrition leads to less weight loss [16]; psychological and psychiatric screening performed at baseline in all patients may be able to preemptively identify early dropouts. Patients experiencing greater mental stress might be better served by first addressing this burden before starting the weight loss program.

Even if late attrition is also regarded as a failure, its consequences seem to be less serious, since those individuals at least lost some weight, even though the weight loss was less than that achieved by the attender group. In a large sample of individuals mainly composed of participants who dropped out late, after 36 months of follow-up, the completer and attender groups showed a greater percentage of weight loss, but those in the dropout group who were satisfied or confident with their ability to reduce their weight reported further weight loss even without professional help, thus suggesting that not all attrition must be considered treatment failure [24]. It is of note that before starting treatment, patients tended to gain a considerable amount of weight [41], and even modest weight-loss could confer health benefits in the long term [42].

Strategies to enhance engagements may include increasing knowledge about the health problems linked to obesity, heightening awareness about the specific individual weight-loss barriers and obstacles and offering potential solutions, raising confidence about one’s control skills, acting on motivation in an individualized and realistic way, avoiding exclusively focusing on weight loss as an outcome but emphasizing the importance of healthy behaviors, better quality of life, and social functioning.

Limitations

We did not assess other psychological factors, such as motivation to change, personality traits and coping abilities, which have been proven to be associated with attrition. Furthermore, we did not assess the reasons for attrition. Nevertheless, our evaluation was performed in a community setting, where the administration of many questionnaires to the patients is not feasible. The data were collected at baseline but may have changed over the course of treatment. Our sample size was small, and our study was limited to a single center with a specific weight-loss program, thus limiting the generalizability of our results. However, the post-hoc power analysis indicated that the study had a > 90% power to evaluate the between-group differences in BMI.

The strengths of our study were: the comprehensive assessment of multiple variables in all participants and the fact that all participants underwent an individual psychological assessment during which answers to the psychological questionnaires were verified and discussed instead of being simply self-reported, thus reducing the risk of either overestimating or underestimating psychopathology.

Conclusions

Specific features seem to be related to the early weight-loss program attrition, suggesting the need for early evaluation and identification of those individuals at greatest risk to target tailored programs focused on improved long-term attendance.

Below is the link to the electronic supplementary material.

Conflict of interest

The authors declare that they have no conflict of interest.

Ethics approval

All procedures performed were in accordance with the ethical standards of the institutional research committee and with the Helsinki declaration and its subsequent amendments or comparable ethical standards. The study was approved by the Ethics Committee of “Città della Salute e della Scienza” Hospital of Torino (Prot. no. 0036671).

Consent to participate

Informed consent was obtained from all participants included in the study.