Abstract

Objective: To summarize the determinants of adherence and dropout and the characteristics of intervention programs with physical exercise in adults with obesity.
Methods: The Scopus, PsycINFO, PubMed/Medline, SPORT Discus, Lilacs, Cochrane, and Web of Science databases were consulted in August 2021. The articles should come from clinical trials conducted with adults with obesity and identify the determinants of adherence and/or dropout from physical exercise interventions in adults with obesity.
Results: Seven studies were included in the synthesis. No study reported determinants associated with dropout. The intrapersonal adherence determinants found were depressive symptoms (n=2), mood conditions (n=2), stress (n=1), change in body composition (n=2), pleasure (n=2), aggravated health conditions (n=1), commitment to an established goal (n=1), body image (n=1), self-efficacy (n=1), and self-concept (n=1). The interpersonal dimension had two determinants of adherence, social support (n=1) and health-related leadership (n=1). The intervention characteristics regarding allocating the training group, using resistance exercises, and determining the intervention duration were related to increased adherence. No determinants related to the environment were found for adherence.
Conclusion: The determinants of adherence to physical exercise programs for adults with obesity were essentially intrapersonal, interpersonal, and intervention characteristics.

Keywords: Adherence; Obesity; Exercise training; Adults

Abbreviations: BMI: Body Mass Index;

Introduction

Obesity is a chronic non-communicable disease affecting more than 650 million adults, representing about 13% of the world’s population [1]. Its occurrence has practically tripled since 19751 and results from complex interactions between genetic, psychological, socioeconomic, cultural, and environmental factors [2]. Given this scenario, numerous interventions using the prescription of physical exercises to reduce obesity in adult populations have been conducted and summarized in the literature [3,4]. Nevertheless, regardless of the strategy used in these interventions, there are reports that approximately 80% of participants drop out before the final scheduled conclusion.

Besides, the weekly frequency of sessions, understood here as adherence, has shown low proportions to the point of reverberating in the finding of less expressive results in the various outcomes studied [5]. Socioecological models emphasize that lifestyle behaviors, including the practice of physical exercise, are complex and bi-directionally influenced by variables of different levels, covering intrapersonal (biological, psychological, cognitive, behavioral), interpersonal (social, cultural), and environmental (physical, community, organizational, and political environment) characteristics [6,7]. A systematic review [8] summarized studies aimed at identifying barriers to behavior change and predictors of adherence in intervention programs focused on lifestyle in adults with obesity.

The main barriers encountered were essentially intrapersonal, such as low motivation, lack of time, poor health, etc. Similarly, the main predictors of adherence reported in the review were all intrapersonal, some of which were an early success in weight loss, lower baseline body mass index (BMI), and better baseline mood. As for the factors associated with dropout in weight loss interventions involving overweight adults, limited conclusions were found in the literature5. The reasons presented were the diversity of variables explored, the small number of studies and the variety of settings and methodologies used, the inconsistent reporting of the results, and the conflicting findings between the studies, indicating an inappropriate set of predictors.

On the other hand, implementing financial incentives, multicomponent interventions, and self-monitoring technology helped reduce the dropout of adults in weight loss programs [9]. Multilevel and multi-component approaches to treating obesity present promising results, especially when integrating components at policy, community, and interpersonal levels [10]. However, there are proposals in the literature which only take account of exercise practice [11]. These interventions have greater feasibility for their execution and more practical applicability in real-life scenarios. Also, they seem to promote benefits for physical fitness, body composition, biochemical markers, and even mental health indicators in this population, even without weight loss [11].

In this sense, we aim to advance in this area of investigation and to identify the variables related to adherence and dropout in interventions in which the focus was on the practice of physical exercise without interfering with other behaviors. The findings may contribute to the advancement and construction of proposals for physical exercise treatment, considering specific aspects related to training prescription, increased adherence, and reduced dropout from programs, especially for those who are more vulnerable to dropout. Furthermore, the determinants of adherence to physical exercise may differ among adults with a condition of installed obesity compared to those with overweight. The worsening of health conditions, as well as the increase in body volume, added to the psychosocial impacts can give individuals with obesity motivations different from those with overweight [12].

For example, adults with higher BMI report less willingness to adhere to general health care recommendations than their lower BMI peers [13]. Besides, physical and health limitations, embarrassment, anxiety, and fear of pain are barriers that can affect adults with obesity more than healthy weight adults. Environmental and social pressures, norms, and expectations can also harm adults with this health condition, especially those surrounded by families and cultures with unhealthy lifestyles [14]. Thus, this systematic review aims to summarize the determinants of adherence and dropout and the characteristics of physical exercise intervention programs in adults with obesity.

Methods

This systematic review examined original scientific articles from intervention studies that identified the determinants of adherence and dropout from physical exercise programs in adults with obesity. In this study, the term dropout refers to the participants who, for some reason, abandoned the treatment at a given moment of the intervention, not maintaining contact with the creators of the proposal. Adherence was understood as the maintenance/permanence of the practice [15], expressed by the attendance control of the participants in the intervention, indicated by the number of physical exercise sessions completed until the scheduled end. The method of this review followed the guidelines of the PRISMA checklist [16] and PERSiST [17] (Supplementary Material 1), and the protocol of this review was registered in PROSPERO (ID=238898).

The searches were conducted in the Scopus, PsycINFO, PubMed/Medline, SPORTDiscus, Lilacs, Cochrane, and Web of Science databases, during August 2021. The search was organized in blocks based on the use of MeSH terms, following the PICO strategy [18]: obes* AND adult* AND (exercise OR training) AND (adhe* OR attrition* OR “patient compliance” OR “patient non-compliance” OR “treatment adherence and compliance” OR dropout* OR attendance OR retention OR compliance). Filters were used in the Scopus, Web of Science, Embase, and PubMed databases for studies with humans, age (from 18 to 65 years), and languages (English, Spanish, or Portuguese). No restrictions were applied as to the year of publication. The specific strategies of each database and the results obtained are available in Supplementary Material [2]. In addition to the searches in the virtual libraries, we consulted the reference lists of potentially eligible studies retrieved in the primary research. Articles from the personal collection and reference lists of systematic reviews on the subject were also considered [5,8,9,19].

Eligibility Criteria

The studies that met the following criteria were included: (i) to have the sample represented by adults aged 18 to 65 years; (ii) to present specific results for participants with obesity, given by BMI ≥30kg/m². It is noteworthy that when the eligible documents did not specify whether the population was composed of adults with obesity, the mean value and measures of BMI dispersion and/ or fat percentage of the study sample were considered; (iii) having the full original text available in English, Spanish, or Portuguese; (iv) identifying the determinants of adherence and/or dropout from physical exercise programs; (v) reporting the prescription of physical exercises, with control of duration, attendance, and intensity; and (vi) being an original article and with data from a clinical trial, even if in a cross-sectional view (presenting the baseline data).

The following exclusion criteria were applied:
(i) studies conducted in pregnant or breastfeeding women;
(ii) that included people with other comorbidities (e.g., hypertension, diabetes, dyslipidemia) diagnosed in the sample;
(iii) that had in the intervention procedures the inclusion of strategies such as bariatric surgery, pharmacotherapy, psychological care, or prescription of specific diets and/or very low caloric value (studies that used only nutritional counseling, collectively and without prescriptive character were considered);
(iv) conducted in stays, residences, or places for rehabilitation;
(v) that used monetary incentives or provision of exercise equipment to participants; and
(vi) that included participants with physical or intellectual disabilities.

Study Selection

The articles were retrieved through the initial electronic search conducted simultaneously by two independent researchers (ARS and CB) at all stages of the study. In cases of disagreement, a third assessor was consulted (GFDD). The references reached in the databases were exported to the Zotero®5.0 reference manager. Subsequently, duplicates were automatically eliminated by the software and manually when necessary. Reading the title and abstract was the first phase of screening relevant studies. The full texts of the remaining references were read in full order to confirm compliance with all eligibility criteria.

Data Synthesis

The data extracted from the studies included in the synthesis were organized in blocks: (i) Study identification; (ii) Sample characteristics; (iii) Intervention characteristics; (iv) Percentage of adherence and dropout; and (v) Predictors of adherence and/ or dropout reported by the studies and the main results related to them. The extraction of all information was systematized in an Excel spreadsheet by two researchers simultaneously (ARS and CB). In the end, the review was carried out by one of them (CB). A descriptive narrative approach was used to synthesize the key findings. The characteristics of the training prescribed in each study were described considering the FITT principle (frequency, intensity, time, and type) [20].

To report on the percentage of adherence of each study, the ratio between the number of sessions attended and the total number of possible sessions throughout the program multiplied by one hundred was considered. The dropout percentage was obtained by the ratio between the number of participants who abandoned the intervention before its end versus the number of participants multiplied by one hundred. The synthesized determinants were grouped in the light of the intrapersonal, interpersonal, and environmental aspects as proposed by the ecological model [7]. Besides, the determinants related to the intervention itself were presented.

Assessment of Risk of Bias

All included studies were assessed for risk of bias using the Cochrane collaboration tool [21]. This tool uses an assessment based on seven domains and classifies each domain as low, high, or uncertain. The seven domains refer to: random sequence generation, allocation concealment, blinding of participants and researchers, blinding of outcome assessment, incomplete outcomes data, selective reporting outcome, and other sources of bias. The assessment consisted of two stages. The first stage referred to individual consultation of each article included in the synthesis. For each domain, the methodological description of the information reported in the study was collected in detail and put on a spreadsheet.

The information was obtained from the original article, in addition to consulting the protocols and other studies from the same database when cited. This fact was exposed when there was no information related to any domain. The second stage consisted of judging the risk of bias for each analyzed domain, classified into three categories: low, high, or uncertain. These classifications were assigned according to the guidelines provided in the tool used (Supplementary Material 4, Table S4.1). Two reviewers (ARS and CB) performed the quality assessment independently. Disagreements were resolved by a third reviewer (GM). The percentage of inter-rater disagreement and the Kappa coefficient of agreement were calculated using the IBM SPSS Statistics software, version 20.0.

Results

The initial search in the databases retrieved a total of 11,842 articles. After excluding duplicates [6,811] titles and abstracts were read, and 72 articles were read in full. In the end, 7 articles met the eligibility criteria and were included in the review (Figure 1). In Supplementary Material 3, there is a list of all studies read in full and the reasons for their exclusion. The assessment of the risk of bias in the articles included in this synthesis can be seen in (Figure 2). The disagreement between the assessors during the risk of bias assessment was 34.7% (kappa = -0.16; 0.52), ranging from 14.3% to 71.4%. Among the seven domains assessed by the tool, the lowest disagreement among the assessors was the random sequence generation and incomplete outcome data, with only one disagreement each.

The domain “other sources of bias” showed greater disagreement among the assessors and is divergent in five of the seven articles assessed (Supplementary Material 4, Table S4.2). Among the synthesized studies, the domains blinding of participants and researchers and other sources of bias had the most uncertain and high results for the risk of bias. While the blinding of the outcome assessment and the selective reporting were classified more frequently as at low risk of bias. Notably, in non-randomized studies [22], the domain random sequence generation was automatically rated with a high risk of bias. A study [23] classified almost all items as uncertain risk, as it presented little information that could allow the judgment proposed by the tool. The assessment of each study can be found in (Table S4). 3 of Supplementary Material [4].

Table 1 presents the main characteristics of the reviewed studies. Most studies were conducted in the United States [23-27] and had a duration of six months [22,24-26]. All studies included in the synthesis involve adults aged between 40 and 60 years of both sexes. Two studies [25,26] included a true control group consisting of randomized participants who were instructed to maintain their usual activities. The exercise interventions varied in frequency, ranging from one [22] to three [23-26] sessions per week. Two studies [22,24] allowed participants to self-select the intensity of the exercises. The duration of each session ranged from 30 minutes22 to 90 minutes [28].

Aerobic exercise was prescribed in all studies, making it the predominant modality. Resistance training was included in five of the seven studies [23,24,26-28] (Table 1). One study25 involved treatment overseen by specialists; however, the physical exercise sessions were not directly supervised. Participants had access to exercise prescriptions and physical facilities, but training was unsupervised. In contrast, two other studies24,26 offered group exercise sessions led by instructors, aiming to promote social cohesion. Additionally, these studies provided support protocols through extra meetings between participants and professionals, which included guidance on self-management and self-regulatory strategies.

Notes: NR: not reported. FITT= frequency, intensity, time, and type, respectively. HRmax= maximum heart rate. ±sd= standard deviation. OB-GYM = group with participants with obesity performing aerobic and resistance physical exercise. OB-NW = group with participants with obesity walking. DM2-GYM = group with participants with type 2 diabetes performing aerobic and resistance physical exercise. DM2-NW= group with participants with type 2 diabetes walking.

Note: NA = no association; – = not tested.

Notes: NR: not reported. FITT= frequency, intensity, time, and type, respectively. HRmax= maximum heart rate. ±sd= standard deviation. OB-GYM = group with participants with obesity performing aerobic and resistance physical exercise. OB-NW = group with participants with obesity walking. DM2-GYM = group with participants with type 2 diabetes performing aerobic and resistance physical exercise. DM2-NW= group with participants with type 2 diabetes walking. The adherence and dropout rates for each study are presented in (Figure 3). Adherence rates ranged from 19% 27 to 83% 28. The highest reported dropout rate was 72% 27. One study24 did not report adherence or dropout data, while another28 reported no participant losses.

In the investigation of determinants, none of the synthesized studies reported predictors of dropout. (Table 2) summarizes all adherence-related determinants identified in the included articles, which were: depressive symptoms [24,27], health conditions [22], stress [24], mood [24,26], changes in body composition [24,26], allocation group [23,26], social support [23], body image [23,26], characteristics of the intervention [23,28], program duration [23], pleasure [23], commitment to achieving a set goal [23], health-related leadership [23], self-efficacy [26], and physical self-concept [26]. One study [25] found no association between mood and session frequency or continued participation in the exercise program. Additionally, sociodemographic variables such as age, sex, ethnicity, education, employment status, and place of residence were not associated with adherence (Table 2 and Supplementary Material 5).

Supplementary Material 5 provides a detailed overview of the independent variables included in each study, the measurement tools used, and the respective conclusions. One study [22] included the greatest number of independent variables in the association models examining adherence to exercise sessions. In contrast, another study [23] identified the highest number of determinants associated with adherence. At the end of approximately four months of intervention, participants in this study shared their perceptions of the intervention and reasons for adherence through interviews.

Figure 4 presents the adherence determinants identified in this synthesis, categorized into intrapersonal, interpersonal, environmental domains, and intervention characteristics. No environmental determinants were identified in relation to adherence to physical exercise programs for adults with obesity. Most adherence determinants were classified within the intrapersonal domain (n = 9). The interpersonal domain included only two factors; both associated with positive perceptions of social support and the presence of health-related leaders. Intervention characteristics-specifically group allocation, inclusion of resistance training, and program duration-were also associated with increased adherence among adults with obesity.

Discussion

This systematic review aimed to identify the determinants of adherence and dropout in physical exercise interventions for adults with obesity. The determinants of adherence identified were related to intrapersonal and interpersonal factors, as well as characteristics of the intervention. These findings are consistent with previous reviews [8,19], which addressed both barriers to behavior change and predictors of adherence and/or dropout among overweight adults. Pratt (1989) [19], in a synthesis of studies published between 1977 and 1988, highlighted key variables associated with adherence to weight-loss interventions. These included the intention to remain in the program, social support, motivation, outcome expectations, and selfreinforcement, in addition to factors related to the intervention design and demographic variables. Similarly, Burgess, Hassmén, and Pumpa (2017) [8] found that early success in weight loss, lower baseline BMI, and better baseline mood were predictors of adherence to interventions promoting a healthier lifestyle.

Intrapersonal determinants were the most prominent in the present review. This may be directly related to the frequency with which variables within this domain are investigated and reported in literature. A greater number of studies tend to focus on factors related to self-knowledge and the individual’s internal processes. Additionally, the influence of the domains of the socioecological model on individual behavior should also be considered [7]. In this context, intrapersonal variables appear to exert a greater influence on adherence to physical exercise interventions among adults with obesity when compared to interpersonal and environmental factors.

In a cross-sectional study, variables related to mental healthsuch as depressive symptoms-and health conditions, including comorbidities, were previously identified as predictors of dropout among overweight adults [29]. However, the study by Annesi (2010) [25] found no association between mood and adherence to physical exercise in adults with obesity. In contrast, two other studies did report such an association [24,26]. Differences in the exercise protocols applied may help explain these divergent findings. In Annesi (2010)25, participants performed unsupervised exercise sessions, without professional guidance. Meanwhile, in the studies by Annesi (2008) [24] and Annesi and Whitaker (2008) [26], participants engaged in group-based exercise sessions led by instructors, which aimed to foster social cohesion.

In addition, these studies implemented a structured support protocol designed to enhance feelings of mastery, competence, and self-efficacy, thereby promoting adherence. Such strategies may have positively influenced depressive symptoms by modifying participants’ perceptions over the course of the intervention and encouraging greater commitment to the treatment. Obesity is often associated with other chronic conditions, including depression. A systematic review with meta-analysis30 indicated a longitudinal and bidirectional relationship between depression and obesity, suggesting that weight gain may be a late consequence of depressive symptoms. This highlights the importance of monitoring mood in patients with obesity. Recognizing the coexistence of these conditions enables prevention, early detection, and timely treatment for at-risk individuals, thereby reducing the negative impacts of both disorders.

Furthermore, obesity is sometimes perceived solely as a physical condition, rather than as a complex disease with serious cardiometabolic consequences. For this reason, it is essential that, in the early stages-before clinical complications develop, individuals are made aware of their condition. Early awareness and appropriate guidance regarding treatment strategies can help prevent the need for pharmacological interventions and reduce the burden on healthcare systems. The study by Marcon, Gus, and Neumann (2011) (22) included the largest number of variables in an adherence prediction model among the studies synthesized in this review.

However, the adherent and non-adherent groups were similar across all measured variables, except for higher systolic and diastolic blood pressure observed in the adherent group. It is important to interpret this finding with caution due to a potential risk of bias identified in the study’s methodological quality assessment. Specifically, the use of a convenience sample composed of individuals on a waiting list for bariatric surgery may limit the generalizability of the results. These participants are likely to have distinct experiences, motivations, and barriers compared to individuals with obesity who are not actively pursuing weight management or surgical treatment. This context should be considered when analyzing the study’s conclusions

The interpersonal factors identified in this review were social support and the presence of health-related leaders within the intervention, both reported by a single study [23]. This study employed interviews to explore determinants of adherence but did not conduct statistical analyses to confirm these associations. Furthermore, it was assessed as having an uncertain risk of bias across most domains, warranting cautious interpretation of its findings. Social support can be categorized as either structuralreferring to the quantitative aspects of social relationships-or functional, which relates to the perceived quality of available support, such as instrumental or emotional assistance [31].

In the context of physical exercise, social support has been associated not only with weight loss but also with adherence to training programs [32]. It is plausible that structured interventions particularly influence structural support by facilitating peer interaction and group formation, which may foster a sense of belonging and, consequently, enhance engagement [33]. This review highlighted that characteristics of the prescribed training—such as exercise type, group allocation, and program duration-were decisive for participant retention in physical exercise interventions. The study by Pippi et al. (2020)28 reported the highest adherence rate among the included studies, with no dropouts.

However, its risk of bias assessment revealed high risk in four out of seven evaluated domains. The authors observed that the walking group showed 14.4% lower adherence compared to the combined training group (aerobic and strength circuit), despite both interventions being supervised. Conversely, the study by Annesi (2010) [24], which employed a self-selected intensity protocol, reported the lowest adherence rates. This study also presented an unclear risk of bias in most assessed domains. Notably, it found that dropout rates were higher in the control group compared to the intervention groups. These findings may reflect real-world scenarios of unsupervised or “autonomous” physical activity in various settings.

Taken together, the evidence suggests that, for adults with obesity, the early stages of an intervention may benefit from structured training programs that include strength training with immediate implementation. Furthermore, professional supervision and frequent variation in exercise stimuli appear to enhance engagement and are likely more suitable for supporting long-term adherence. No environmental determinants were identified that could explain engagement in physical exercise interventions among adults with obesity. This absence may be specifically attributed to the design of the included studies. Clinical, controlled environments used in scientific research often represent scenarios that are far removed from real-world conditions, offering structures and resources not typically found in commercial or community settings accessible to the general population.

In everyday contexts, factors related to the built environment, perceived safety, and even political and policy dimensions may significantly influence the maintenance of an active lifestyle [34]. Nevertheless, scoping review [35] synthesized existing evidence on the relationships between the built environment, physical activity, and obesity. While the authors identified a small subset of environmental factors consistently associated with weight and related behaviors, the overall results were marked by heterogeneity in both the strength and statistical significance of the associations. In lifestyle interventions, disparities based on age and gender have been identified, with being male and older emerging as prominent predictors of adherence among overweight adults [8].

These factors also appear to influence the distribution of physical activity within the population [36]. However, in the studies synthesized in this review, no significant differences were reported between men and women or across different age groups in relation to adherence to physical exercise programs among individuals with obesity [22,27]. It is possible that, for this population, engagement in physical exercise is more strongly influenced by psychosocial factors-such as perceived changes in body composition and body image, increased pleasure, self-efficacy, and physical self-concept [24,26,27] than by sociodemographic characteristics alone. None of the synthesized studies proposed associations between variables and dropout rates.

Most focused on pre- and post-intervention data, without addressing follow-up outcomes. These data could provide valuable insight into the reasons for dropout and whether individuals who adhered to the intervention remained active on their own after the training program ended. It is important to recognize that participants who were technically adherent but had low attendance should not be equated with those who dropped out entirely. The challenges associated with behavior change and the integration of physical exercise into daily life differ significantly between individuals actively attempting to adopt a healthier lifestyle and those who disengage from the intervention altogether. In 2015, Miller and Brennan called for greater attention to the reporting and monitoring of dropout in obesity treatment programs. The authors highlighted key aspects that should be addressed in future studies, including clear definitions of program completers and dropouts, as well as factors distinguishing individuals who discontinue the intervention from those who persist at any stage [37].

Future Research

For the research to advance on this theme, new publications should be built according to the recommendations for reporting interventions, such as the use of the Consolidated Standards of Reporting Trials (CONSORT) [38] or specific guidelines for the measurement and the report of dropout from obesity treatment, suggested by Miller and Brennan [37]. This is due to the frequent classification as uncertain in assessing the risk of bias, a characteristic stemming from the lack of information that would allow a judgment of quality at high or low risk.

- Studies with a single focus on the process of adherence to physical exercise by adults with obesity are still necessary. The details of information related to adherence, such as program entry, dropout, and adherence rates, need to be clearly reported.

Moreover, studies should analyze and describe which variables, both from the baseline and those resulting from the intervention, are related to program permanence or dropout.

• More original studies that test the relationship of interpersonal and environmental variables with adherence to physical exercise programs for adults with obesity are also needed, especially those concerning social support. This is because few determinants have been identified within these constructions so far.

• New original studies should also propose strategies within their protocols to increase adherence and reduce dropout rates.

• Future reviews could advance this field by synthesizing interventions involving physical exercise in this population that reported high adherence rates. Detailed descriptions of the applied protocols may offer valuable insights into best practices, which can inform the design of future interventions aimed at enhancing adherence and minimizing dropouts.

• It is recommended that the determinants identified in this review be considered when developing physical exercise interventions for adults with obesity, aiming for outcomes that go beyond weight loss, such as promoting long-term engagement in an active lifestyle.

This systematic review presents several strengths, including the investigation of multiple outcomes related to adherence to physical exercise interventions in adults with obesity. The findings enable the identification of key information that reflects the effectiveness and success of exercise programs in the treatment of obesity. Additionally, the reporting of percentages associated with the synthesized variables represents a notable strength, as the literature still lacks comprehensive compilations of adherence and dropout rates in physical exercise programs targeting this population. The inclusion of these rates allows future studies to make comparisons and use them as benchmarks for setting goals and evaluating outcomes.

To the best of our knowledge, this is the first review to exclusively explore the determinants of adherence to physical exercise interventions in adults with obesity. Addressing this gap contributes not only to academic literature but also provides practical evidence to support professionals implementing exercise as a treatment strategy. This review identified a set of predictive variables for adherence. From this evidence, tools can be developed to assist in the screening process for interventions, identifying individuals at greater risk of dropping out. Early identification of low-adherence profiles enables the development of targeted strategies tailored to specific determinants, potentially increasing the success of the intervention.

Furthermore, the classification of adherence determinants according to the socioecological model represents a methodological strength of this study. Analyzing determinants within their respective domains, rather than in isolation, allows for a clearer understanding of which level exerts the greatest influence-in this case, the intrapersonal domain. This approach can substantially aid in identifying mechanisms behind subgroup differences and in designing focused, domain-specific strategies for enhancing adherence [6,7]. As for limitations, the review highlights the small number of eligible studies, which restricts the depth of understanding and discussion on this topic. Moreover, the findings must be interpreted with caution due to the risk of bias in the included studies and the heterogeneity of the interventions, even when limited to physical exercise.

Conclusion

The synthesis of physical exercise interventions for adults with obesity led to the conclusion that the primary determinants of adherence to these programs are predominantly intrapersonal. These include depressive symptoms, overall health and mood conditions, stress levels, changes in body composition, perceived pleasure, commitment to achieving a defined goal, body image, self-efficacy, and physical self-concept. In contrast, interpersonal determinants-such as social support and the involvement of health-related leaders-and aspects related to the intervention itself (e.g., type of exercise, group allocation, and program duration) have not yet been consistently consolidated in the literature. Additionally, no environmental variables were found to be associated with adherence to physical exercise interventions in this population. Notably, none of the reviewed studies investigated potential predictors of dropout, highlighting a significant gap in the literature.

Acknowledgments: The authors thank the Fundação de Amparo à Pesquisa e Inovação do Estado de Santa Catarina - FAPESC, for receiving scholarships.

Conflict of interest: All authors declare no potential conflicts of interest.

References

1. WHO (2021) World Health Organization. Obesity.
2. Francischi RPP de, Pereira LO, Freitas CS, et al. (2000) Obesidade: atualização sobre sua etiologia, morbidade e tratamento. Rev Nutr 13(1): 17-28.
3. Katzmarzyk PT, Lear SA (2012) Physical activity for obese individuals: a systematic review of effects on chronic disease risk factors. Obes Rev 13(2): 95-105.
4. Adrian Thorogood, Salvatore Mottillo, Avi Shimony, Kristian B Filion, Lawrence Joseph, et al. (2011) Isolated Aerobic Exercise and Weight Loss: A Systematic Review and Meta-Analysis of Randomized Controlled Trials. Am J Med 124(8): 747-755.
5. Moroshko I, Brennan L, O’Brien P (2011) Predictors of dropout in weight loss interventions: a systematic review of the literature. Obes Rev Off J Int Assoc Study Obes 12(11): 912-934.
6. Sallis JF, Cervero RB, Ascher W, Henderson KA, Kraft MK, et al. (2006) An ecological approach to creating active living communities. Annu Rev Public Health 27: 297-322.
7. Sallis JF, Owen N, Fisher E (2015) Ecological models of health behavior. Health Behav Theory Res Pract 5(43-64).
8. Burgess E, Hassmén P, Pumpa KL (2017) Determinants of adherence to lifestyle intervention in adults with obesity: a systematic review. Clin Obes 7(3):123-135.
9. Stephanie Pirotta, Anju Joham, Lisa Hochberg, Lisa Moran, Siew Lim, et al. (2019) Strategies to reduce attrition in weight loss interventions: A systematic review and meta-analysis. Obes Rev 20(10): 1400-1412.
10. Ewart-Pierce E, Mejía Ruiz MJ, Gittelsohn J (2016) “Whole-of-Community” Obesity Prevention: A Review of Challenges and Opportunities in Multilevel, Multicomponent Interventions. Curr Obes Rep 5(3): 361-374.
11. Petridou A, Siopi A, Mougios V (2019) Exercise in the management of obesity. Metabolism 92: 163-169.
12. Van Hout GCM, van Oudheusden I, van Heck GL (2004) Psychological Profile of the Morbidly Obese. Obes Surg 14(5): 579-588.
13. Sansone RA, Bohinc RJ, Wiederman MW (2015) Body mass index and self-reported compliance with general health care. South Med J 108(2): 79-81.
14. Šárka Slabá, Iva Málková, Martin Wagenknecht, Karel D Riegel, Lukáš Junek, et al. (2020) Psychological aspects of obesity. Cas Lek Cesk 159(3-4): 118-124.
15. World Health Organization (2003) Adherence to long-term therapies: evidence for action / [edited by Eduardo Sabaté].
16. Page MJ, McKenzie JE, Bossuyt PM, et al. (2021) The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 372: n71.
17. Clare L Ardern, Fionn Büttner, Renato Andrade, Adam Weir, Maureen C Ashe, et al. (2022) Implementing the 27 PRISMA 2020 Statement items for systematic reviews in the sport and exercise medicine, musculoskeletal rehabilitation and sports science fields: the PERSiST (implementing Prisma in Exercise, Rehabilitation, Sport medicine and SporTs science) guidance. Br J Sports Med 56(4): 175-195.
18. Santos CM da C, Pimenta CA de M, Nobre MRC (2007) The PICO strategy for the research question construction and evidence search. Rev Lat Am Enfermagem 15(3): 508-511.
19. Pratt CA (1989) Development of a screening questionnaire to study attrition in weight-control programs. Psychol Rep 64(3 pt 1): 1007-1016.
20. Morton RW, Colenso-Semple L, Phillips SM (2019) Training for strength and hypertrophy: an evidence-based approach. Curr Opin Physiol 10: 90-95.
21. De Carvalho A, Silva V, Grande A (2013) Avaliação do risco de viés de ensaios clínicos randomizados pela ferramenta da colaboração Cochrane. Diagn Trat 18(1): 38-44.
22. Marcon ER, Gus I, Neumann CR (2011) Impacto de um programa mínimo de exercícios físicos supervisionados no risco cardiometabólico de pacientes com obesidade mórbida. Arq Bras Endocrinol Metabol 55(5): 331-338.
23. Gillett PA (1988) Self-reported factors influencing exercise adherence in overweight women. Nurs Res 37(1): 25-29.
24. Annesi JJ (2008) Relations of mood with body mass index changes in severely obese women enrolled in a supported physical activity treatment. Obes Facts 1(2): 88-92.
25. Annesi JJ (2010) Dose-response and self-efficacy effects of an exercise program on vigor change in obese women. Am J Med Sci 339(2): 127-132.
26. Annesi JJ, Whitaker AC (2008) Weight loss and psychologic gain in obese women-participants in a supported exercise intervention. Perm J 12(3): 36-45.
27. Aubrey L Shell, Loretta Hsueh, Elizabeth A Vrany, Daniel O Clark, NiCole R Keith, et al. (2020) Depressive symptom severity as a predictor of attendance in the HOME behavioral weight loss trial. J Psychosom Res 131: 109970.
28. Roberto Pippi, Andrea Di Blasio, Cristina Aiello, Carmine Fanelli, Valentina Bullo, et al. (2020) Effects of a Supervised Nordic Walking Program on Obese Adults with and without Type 2 Diabetes: The C.U.R.I.A.Mo. Centre Experience. J Funct Morphol Kinesiol 5(3): E62.
29. E M Inelmen 1, E D Toffanello, G Enzi, G Gasparini, F Miotto, et al. (2005) Predictors of drop-out in overweight and obese outpatients. Int J Obes 29(1): 122-128.
30. Floriana S Luppino, Leonore M de Wit, Paul F Bouvy, Theo Stijnen, Pim Cuijpers, et al. (2010) Overweight, Obesity, and Depression: A Systematic Review and Meta-analysis of Longitudinal Studies. Arch Gen Psychiatry 67(3): 220-229.
31. Cohen S, Wills TA (1985) Stress, social support, and the buffering hypothesis. Psychol Bull 98(2): 310.
32. M Becky Marquez, Andrea Anderson, Rena R Wing, Delia S West, Robert L Newton, et al. (2016) The Relationship of Social Support with Treatment Adherence and Weight Loss in Latinos with Type 2 Diabetes. OBESITY 24(3): 568-575.
33. Verheijden M, Bakx J, Van Weel C, Koelen M, Van Staveren W (2005) Role of social support in lifestyle-focused weight management interventions. Eur J Clin Nutr 59(1): S179-S186.
34. Giles-Corti B, Donovan RJ (2002) The relative influence of individual, social and physical environment determinants of physical activity. Soc Sci Med 54(12): 1793-1812.
35. Dixon BN, Ugwoaba UA, Brockmann AN, Ross KM (2021) Associations between the built environment and dietary intake, physical activity, and obesity: A scoping review of reviews. Obes Rev 22(4): e13171.
36. Althoff T, Sosič R, Hicks JL, King AC, Delp SL, et al. (2017) Large-scale physical activity data reveal worldwide activity inequality. Nature 547(7663): 336-339.
37. Miller BML, Brennan L (2015) Measuring and reporting attrition from obesity treatment programs: A call to action! Obes Res Clin Pract 9(3): 187-202.
38. Schulz KF, Altman DG, Moher D (2010) CONSORT 2010 Statement: Updated Guidelines for Reporting Parallel Group Randomized Trials. Ann Intern Med 152(11): 726-732.