Participants will complete 12 training sessions (twice weekly) spread over 7 weeks. The first week will be devoted to familiarization with the exercise protocol, while the subsequent 6 weeks will correspond to the actual intervention phase. Assessments will be conducted at baseline, immediately after the intervention is completed, and at a 4-week follow-up. The 4-week follow-up is included to evaluate the short-term maintenance of physical and psychological effects after the intervention; however, this timeframe does not allow conclusions regarding the long-term sustainability of behavior and psychological changes.

The groups will receive the designated intervention according to the following allocations: HBG, FxF, or CG. The 2 interventions will be compared to estimate the magnitude of differences in symptom relief and other outcomes, rather than to formally test equivalence. Participants allocated to the CG will be instructed to maintain their usual lifestyle and PA habits throughout the 7-week intervention period and the 4-week follow-up, and not to start any new structured exercise program (eg, supervised gym training, personal training, online exercise programs, or new exercise routines performed ≥2 times/wk). CG participants will receive no supervised training sessions or exercise prescription during the study period. To monitor potential contamination, all participants (including CG) will be asked at baseline, postintervention, and follow-up whether they started, stopped, or substantially changed their PA routines during the study. In addition, CG participants will be contacted weekly by message or phone call to confirm that no new structured exercise program was initiated and to record any relevant changes in PA behavior. Any initiation of a new structured exercise program during the trial will be documented and reported. These participants will remain in the intention-to-treat (ITT) analysis, and a per-protocol sensitivity analysis excluding participants with major protocol deviations will be conducted.

In addition to exercising, participants will receive educational instruction about their disease, surgery, eating habits, and PAs. A summary of the frequency, intensity, type, time, volume, and progression (FITT-VP) principles, as well as the progression factors proposed for each program modality can be found in Table 2. Researchers who participate in the clinical trial will undergo specific training to standardize the intervention. The professionals have more than 10 years of experience in the field of physical exercise and prehabilitation for individuals with obesity.

Intervention fidelity will be ensured through standardized training of the professionals involved, structured intervention protocols, and systematic documentation of session delivery. For the HBG, asynchronous video verification will also be used to assess exercise execution and adherence. Participants will be asked to submit video recordings of selected sessions, which will be reviewed by the research team to ensure correct performance and fidelity to the prescribed protocol.

The protocol will consist of 2 parts: aerobic training and resistance training, and it will last approximately 70 minutes and be performed twice a week on alternate days. The training program was developed specifically for individuals with obesity and can be applied in any environment.

At the beginning of the program, one of the researchers will meet with the participants online (via Google Meet) to explain how the training protocol will be conducted and how the exercises should be performed, and to answer any questions that may arise. Additionally, the meeting will be recorded and made available for participants to access, along with specific videos explaining each exercise. Participants will also have direct contact with the training supervisor for assistance with exercise execution. To monitor the training sessions, participants will be instructed to send a message informing the research team of the completion of the exercises and the time, and the researchers will collect this information weekly. The exercise program will consist of 30 minutes of aerobic training on a treadmill, stationary bike, or walking outdoors, followed by 30 minutes of strength exercises. The Borg scale (CR-10) will be used to monitor exercise intensity, and it is recommended that effort be maintained between 4 and 6, considered a moderate zone of effort, for both aerobic and strength exercises [39], or a 5% increase in speed or incline will be implemented every 2 weeks and in the load of resistance exercises after mastering 10 repetitions (3 sets of 8-10 repetitions, 40-60 seconds of rest). The strength exercises for the upper limbs will include arm flexion (inclined), cable row, and shoulder press, and those for the lower limbs will include free squats, lateral lunges, stepping hip thrusts, and plantar flexions. The exercises can be adapted if the participant has any mobility restrictions or is unable to perform any of the exercises.

Two researchers with experience in interventions for individuals with obesity will conduct the training sessions at the physiotherapy clinic at UDESC. The sessions will be held in the afternoon and last approximately 70 minutes. The first week will be dedicated to familiarizing participants with the exercise protocol.

Initially, participants will perform 30 minutes of aerobic exercise (treadmill or stationary bike). Once this stage is complete, they will begin a protocol of strength exercises, which will include arm flexion (incline), cable row, and shoulder press. The exercises for the lower limbs will be free squats, lateral lunges, stepping hip thrusts, and plantar flexions, and participants will perform the training in the same manner as the HBG. The loads from all training sessions of the FxF group patients will be recorded, with patients being encouraged to reach concentric failure (ie, the inability to perform one more repetition) in the last series, as prescribed. We chose to prescribe training using maximum repetitions, where the number of repetitions to be performed is fixed and load variations occur based on this reference. A progressive increase in the load level will be initiated when the patient correctly performs the exercises with an ideal movement pattern for 2 consecutive days. Figure 2 presents an illustration of the resistance exercises.

The health education intervention will consist of 3 previously recorded videos, which will be made available after the week of familiarization, between the second and third weeks of the intervention, and between the fourth and fifth weeks of training.

The sessions will be recorded by physiotherapists and will address essential aspects of the relationship between PA and BMS. The videos will each be a maximum of 10 minutes long and will be structured to achieve the following objectives: (1) understand the importance of PA in the context of BMS, (2) identify and overcome barriers to the regular practice of PA, (3) learn behavioral and cognitive strategies to establish a more active lifestyle, and (4) create a personalized action plan, with realistic and sustainable goals.

Each video will address the 3 pillars of the self-determination theory (SDT): autonomy, competence, and relatedness, in addition to the steps called the “5As”: assess, advise, agree, arrange, and assist [40-42]. To ensure intervention fidelity and quality, participant engagement will be monitored via video analytics (defining adherence as ≥80% viewing completion). Furthermore, the content has been culturally adapted to the Brazilian context, featuring Portuguese audio with subtitles for accessibility. To specifically address and mitigate obesity-related stigma, self-compassion strategies have been embedded throughout the video content, reinforcing the “autonomy” pillar.

Table 3 provides a detailed view of the contents of each session, aligning with the pillars of the SDT and the steps of the “5As.”

Assessments will be conducted before the start of the intervention, after the 7-week program, and at the 4-week follow-up (Figure 3). Data collection will be conducted in person at the UDESC facilities for both the face-to-face and home-based groups. In the initial assessment, a characterization questionnaire will be administered to obtain socioeconomic and health information (age, marital status, educational level, occupation, PA level, and main symptoms). Patient data will be maintained solely by the principal researcher to protect confidentiality throughout the study, from inception to conclusion. The variables analyzed, along with their respective instruments, are listed below.

The primary outcome will be the total score of the Hospital Anxiety and Depression Scale (HADS). The HADS was developed in England [43] and was validated for Brazilian Portuguese by Botega et al [44]; the instrument showed internal consistency (Cronbach α) of 0.68 for anxiety and 0.77 for depression in its validation study. The instrument is a scale completed through an interview, preferably answered by the patient, and contains 14 questions, of which 7 assess anxiety (HADS-A) and 7 assess depressive symptoms (HADS-D). The scale emphasizes the psychological signs or consequences of anxiety and depression, excluding clinical symptoms such as dizziness and headache. The questions alternate, with half of them written in a positive tone and the other half in a negative tone. Each question is assigned a score from 0 to 3, with 3 indicating a state associated with more depressive symptoms or anxiety. Analyses of the HADS-A and HADS-D subscales will be conducted as secondary exploratory outcomes. In this study, HADS scores ≥8 will be used as a cutoff for each domain, as this is the score that indicates the presence of depressive and anxious symptoms and possible cases of depression and anxiety.

Secondary outcomes will include analyses of the HADS subscales (HADS-A and HADS-D), sleep quality assessed using the Pittsburgh Sleep Quality Index (PSQI), mood states assessed using the Brunel Mood Scale (BRUMS), quality of life assessed using the World Health Organization Quality of Life–BREF (WHOQOL-BREF), and self-efficacy assessed using the Physical Exercise Self-Efficacy Scale. Physical function will be assessed using the 6MWT, handgrip strength, and the 5-repetition sit-to-stand test. Adherence will be evaluated according to group allocation: attendance at supervised sessions for the FxF group, and, for the HBG, aerobic exercise recorded via the Strava smartphone app (Strava, Inc) and resistance training reported via text message.

The PSQI, developed by Buysse et al [45], demonstrates high internal consistency (global Cronbach α=0.83). It consists of 19 questions assessing the quality and pattern of sleep, grouped into 7 components: subjective quality, latency, duration, habitual efficiency, disorders, medication use, and daytime dysfunction. Each component is evaluated on a scale of 0 to 3, yielding a global score (range 0‐21); a score ≥5 indicates poor sleep quality.

Mood states will be evaluated using the BRUMS [46]. Validated for the Brazilian population, the instrument shows robust psychometric properties, with subscale Cronbach α coefficients ranging from 0.70 to 0.85 (eg, tension α=0.78) [47]. It consists of 24 items across 6 domains: tension, depression, anger, vigor, fatigue, and mental confusion. Participants rate each item on a 5-point scale (0=“not at all” to 4=“extremely”), resulting in a total score ranging from 0 to 16 for each mood state.

Preoperative quality of life will be assessed using the WHOQOL-BREF instrument, developed by the World Health Organization [48]. Validated for the Brazilian population, the instrument demonstrates satisfactory internal consistency (Cronbach α=0.91) [49]. This self-assessment questionnaire consists of 26 items: 2 general questions regarding overall quality of life and health, and 24 items distributed across 4 domains: physical health, psychological, social relationships, and environment. The instrument enables a comprehensive, multidimensional assessment, in which higher scores indicate a better perception of quality of life.

To assess self-efficacy, the Physical Exercise Self-Efficacy Scale will be used [50]. This instrument consists of 5 items that assess an individual’s confidence in performing physical exercises under different emotional states. The items are as follows: feeling worried and in trouble, feeling depressed, feeling nervous, feeling tired, and feeling busy. The instrument was translated, culturally adapted, and validated for the Portuguese population by Martins et al [51]. The questionnaire can be self-administered or administered via interview. Each item is graded on a 5-point Likert scale, defined as follows: 1=“not true at all,” 2=“hardly true,” 3=“probably true,” and 4=“exactly true.”

The interviewer should take care not to influence the patient’s response. When a respondent does not understand the meaning of a question, the interviewer should reread the question slowly, without using synonyms, and neither the question nor its meaning should be discussed, nor should the answer scale be explained. In cases where this is impossible (illiteracy, severe visual impairment, or lack of clinical condition), the instrument is administered by the interviewer, and the effort to avoid influencing the individual’s answers should be redoubled. The score for each facet is the corresponding value marked, and the average of the items is used to obtain the score for each domain answered. The final score is calculated by averaging the scores of all domains (physical, psychological, social relationships, and environment) and ranges from 0 to 100 points. The higher the score, the better the quality of life in the domain or overall.

The 6MWT will be used to assess functional capacity, following the recommendations of the American Thoracic Society [52]. The procedure will be conducted by a trained evaluator in a standardized corridor at the university. Participants will be instructed to walk the longest distance possible for 6 minutes, receiving standardized encouragement. Heart rate, SpO₂, blood pressure, and subjective perception of exertion (perceived exertion using the Borg scale for dyspnea and lower limb fatigue) will be monitored at baseline, every 2 minutes, and at the end of the test. Continuous monitoring of heart rate and SpO₂ will be performed using a pulse oximeter. For greater reliability, each participant will perform 2 tests on the same day, with a minimum interval of 20 minutes between them. The longest distance covered will be considered for analysis.

Handgrip strength will be assessed using a Saehan DHD-1 digital dynamometer, following the recommendations of the American Society of Hand Therapists [53]. Participants will be seated comfortably in an armless chair with their feet flat on the floor, hips and knees flexed at approximately 90°, shoulders adducted and neutrally rotated, elbows flexed at 90°, forearms in a neutral position, and wrists in a range of 0° to 30° extension and 0° to 15° adduction. Three consecutive 3-second maximal contractions will be performed with each hand, with intervals of 30 seconds between attempts and 2 minutes between hands. The arithmetic mean of the 3 tests for each limb will be used for analysis.

The 5-repetition sit-to-stand test will be used to assess lower limb strength, balance control, fall risk, and functional capacity [54,55]. The participant will remain seated in a standardized chair, with feet flat on the floor and arms crossed across the chest, and will be instructed to stand up and sit down 5 consecutive times as quickly as possible. Before and after the test, heart rate, SpO₂, and perception of dyspnea will be recorded using the modified Borg scale. The total time to complete the test will be considered for analysis.

Adherence will be assessed based on the characteristics of each group. In the in-person group, attendance at supervised training sessions will be considered. In the HBG, adherence to aerobic exercise will be monitored by recording activities on the Strava app, while resistance training will be tracked using the Physitrack app (Physitrack PLC), which provides video and text instructions, automatic reminders, and self-monitoring functions. This tool has already demonstrated good acceptability in home-based programs and contributed to greater participant engagement [56]. Based on the records obtained, it will be possible to quantify the proportion of completed sessions relative to the total number planned, as well as verify training feedback and access to the health education videos. For analysis purposes, good adherence will be considered to be participation in at least 75% of the proposed sessions, a value close to that adopted in international multimodal prehabilitation programs that used more rigorous cutoffs, such as 80% [57].

Different approaches to measuring adherence across groups are justified by the distinct modes of intervention delivery. Adherence will be reported descriptively and explored in sensitivity analyses; no formal moderation analyses based on adherence are planned.

The G*Power 3.1 (Heinrich-Heine-Universität Düsseldorf) program was used to determine the sample size. The calculation was based on the HADS total score, which is defined as the primary outcome of the study, assuming a repeated-measures design with 3 groups (home-based, face-to-face, and control) and 3 assessment time points (baseline, postintervention, and 4-week follow-up). Considering an α level of .05 and power of .80, an effect size of f of 0.25 (moderate) was assumed for the group × time interaction, with an assumed correlation among repeated measures of 0.50 and a nonsphericity correction (ε) of 1.0. Under these assumptions, a sample of at least 42 participants, randomized into the 3 groups, is required. This sample size was calculated to detect differences between groups and was not designed to formally demonstrate equivalence between the interventions. Anticipating a dropout rate of approximately 30% (18/60), a total of 60 participants will be recruited to ensure the necessary final sample size of 42 participants.

The analysis will be conducted using the ITT principle, including all randomized participants. Per-protocol sensitivity analyses will also be performed. For continuous variables, a linear mixed model will be used, with fixed effects for group (face-to-face, home-based, and control), time (baseline, postintervention, and follow-up), and group × time interactions. Participants will be modeled as a random effect. Baseline values will be included as part of the repeated-measures structure rather than as covariates, allowing the estimation of within- and between-group changes over time. Age and sex will be considered as covariates in adjusted models if baseline imbalances or clinically relevant associations with the outcomes are identified. The home-based vs face-to-face comparison will be interpreted based on effect estimates and confidence intervals and will not be treated as a formal equivalence analysis. This model, which treats repeated within-participant measures as a random effect, will be used. An appropriate covariance structure for the repeated measures will be selected based on model fit criteria. This method allows for the management of missing data under the assumption that they are missing at random. In cases of systematic losses, multiple imputations may be used as a complementary analysis.

For categorical or dichotomous variables, the chi-square or Fisher exact tests will be applied, where appropriate. For all hypothesis tests, the α level for statistical significance will be set at P<.0, corresponding to a 95% confidence level. To assess the magnitude of significant differences between the preintervention and postintervention moments and between groups, the Hedge g (x ̅1–x ̅2⁄ SD grouped) will be calculated. The effect size magnitudes will be interpreted as follows: <0.2, no effect; 0.2 to 0.4, small; 0.5 to 0.7, moderate; and ≥0.8, large [58]. Secondary outcomes will be analyzed using the same modeling strategy; however, these analyses will be considered exploratory, and no formal adjustment for multiple comparisons is planned. The data will be analyzed using IBM SPSS software, version 20.0, licensed to UDESC, and Microsoft Office Excel for Windows.