Study setting, design, and ethics

The study was a cross-sectional analysis conducted at the Department of Physical Medicine and Rehabilitation at KKU Hospital, a tertiary care hospital specializing in musculoskeletal and rehabilitation services. The study was conducted from 13 July 2024–18 March 2025, with Institutional Review Board approval from KKU Hospital (REC# 2024–6743) to ensure adherence to ethical standards. The extended data collection period was necessary to ensure recruitment of a representative sample across all physical activity levels and to accommodate fluctuations in patient inflow across outpatient departments. This also allowed for balanced stratification across demographic subgroups relevant to the moderation analysis. Before enrollment, all participants provided written informed consent, and the study procedures complied with the ethical principles of the Declaration of Helsinki.

Participants

This study recruited participants from the outpatient physiotherapy department of KKU Hospital, where individuals with musculoskeletal disorders were receiving rehabilitation. The clinical environment included general and specialty musculoskeletal physiotherapy clinics within a tertiary hospital setting, offering services such as manual therapy, electrotherapy, and individualized exercise-based rehabilitation. Recruitment was conducted via direct referral from treating physiotherapists during routine clinical visits. Eligible patients were approached in private consultation rooms, and confidentiality was maintained throughout consent and data collection procedures. Personal identifiers were removed from all datasets prior to analysis. Diagnostic criteria included clinically confirmed musculoskeletal conditions affecting mobility and function, as documented by a licensed physician based on physical examinations, imaging studies, and patient-reported symptoms [12]. Common diagnoses among participants included osteoarthritis, lower back pain, and joint injuries, which significantly impaired functional ability and required structured physiotherapy interventions [13]. The distribution of diagnoses in the sample was as follows: chronic lower back pain (38.46%), osteoarthritis (29.02%), shoulder or knee joint injuries (22.73%), and other musculoskeletal conditions such as tendinopathies or post-fracture rehabilitation (9.79%). Diagnosis was based on clinical examination and imaging, where applicable, as documented in participants’ medical records. The inclusion criteria for this study were as follows: adults aged 18 years and older who had been diagnosed with a musculoskeletal disorder requiring physiotherapy, with an estimated recovery time of at least eight weeks. Participants were also required to have the cognitive capacity to understand and complete self-reported assessments on QoL, physical activity, and psychological well-being. Exclusion criteria included patients with acute infections, recent fractures, or neurological disorders that could interfere with physiotherapy outcomes. Individuals with severe mental health conditions, such as major depressive disorder or psychosis, which might limit adherence to the rehabilitation protocol or influence self-report accuracy, were also excluded. Patients who were non-ambulatory or had contraindications to physical activity, such as unstable cardiovascular disease, were not eligible.

Participants were selected through purposive sampling based on study eligibility criteria. No follow-up assessments were conducted; all data were collected at a single time point consistent with the cross-sectional design. A total of 412 patients were initially screened for eligibility during the recruitment period. Of these, 286 met the inclusion criteria and consented to participate, while 126 were excluded. Exclusions included 53 patients who did not meet diagnostic or cognitive inclusion criteria, 41 who declined participation, and 32 who were unable to commit to the assessment schedule. A participant flow diagram has been provided (Fig 1) to illustrate the recruitment process, including numbers assessed for eligibility, excluded participants (with reasons), and those included in the final analysis, as recommended by STROBE. The final enrolled cohort thus represented 69.4% of those screened. While purposive sampling enabled targeted recruitment of individuals receiving active physiotherapy for musculoskeletal disorders, this approach may limit generalizability and introduce selection bias due to its reliance on clinical judgment and participant availability. These limitations should be considered when interpreting the findings. Their physiotherapists approached eligible individuals and provided detailed information about the study’s purpose, procedures, and role. This recruitment process aimed to recruit a diverse sample that reflects typical patients undergoing physiotherapy for musculoskeletal disorders, ensuring enough variability in physical activity levels and other demographic characteristics relevant to the study’s objectives.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Participant flow diagram for recruitment and inclusion.

https://doi.org/10.1371/journal.pone.0341335.g001

Variables

The primary outcome variables in this study included functional recovery, QoL, and psychological well-being. Functional recovery was evaluated using the Patient-Specific Functional Scale [6], where participants rated their limitations in daily activities on a scale from 0 to 10, with higher scores representing greater functional capacity. Participants were instructed to identify three to five daily activities they found difficult due to their musculoskeletal condition. Each activity was rated from 0 (unable to perform) to 10 (able to perform at prior level). The final PSFS score represented the mean of all selected activities per participant, following standard PSFS scoring recommendations. The number of selected activities was recorded to ensure consistency in score interpretation. QoL was assessed using the Short Form-36 Health Survey (SF-36) [15], a validated instrument comprising eight subscales that cover physical functioning, physical role limitations, bodily pain, general health, vitality, social functioning, emotional role limitations, and mental health [7]. Higher scores on each subscale and overall indicate better perceived QoL [7]. Psychological well-being was measured through the Hospital Anxiety and Depression Scale (HADS) [8], a 14-item tool that separately measures anxiety and depression levels, with higher scores reflecting greater psychological distress [8].

The primary independent variable was physical activity level, measured in metabolic equivalent of task (MET) hours per week [17]. Participants completed the Arabic-translated IPAQ-SF, which records physical activity in MET-minutes per week. These scores were computed using standard IPAQ scoring protocols and subsequently converted to MET-hours per week by dividing total MET-minutes by 60, to enhance interpretability in regression models [17]. The physical activity data were collected using the short form of the International Physical Activity Questionnaire (IPAQ-SF), which has been translated and culturally validated for Arabic-speaking populations. The IPAQ-SF is a widely used tool for estimating MET-minutes per week across walking, moderate, and vigorous activities, and was scored according to standard IPAQ protocols. Based on this measure, participants were categorized into three groups: low, moderate, and high physical activity. The classification of physical activity levels was based on the official IPAQ-SF scoring protocol. Participants with less than 600 MET-min/week were categorized as having low physical activity, those with 600–2999 MET-min/week as moderate, and those with 3000 MET-min/week or more as high physical activity. These thresholds are consistent with WHO recommendations and have been validated for use in adult populations across different cultural contexts. In the present sample, 27.97% of participants were categorized as low activity, 42.31% as moderate, and 29.72% as high activity, ensuring a balanced distribution that supports group comparisons. Demographic and lifestyle variables, such as age, gender, socioeconomic status, and comorbidities, were also collected to examine their potential moderating effects on the primary outcomes [18]. Baseline severity was operationalized as the initial functional recovery score reported by each participant using the Patient-Specific Functional Scale (PSFS), where lower scores at intake indicated more severe functional impairment. Socioeconomic status was classified as low, middle, or high based on a composite of two criteria: (1) highest level of education attained and (2) self-reported monthly household income. High socioeconomic status was defined as having a college degree or higher and a monthly income above the national median (as per Saudi General Authority for Statistics, 2024 figures [9]). Medication use was recorded as a binary variable (yes/no) based on medical records and self-reports, encompassing any prescribed medications for chronic conditions relevant to musculoskeletal health (e.g., analgesics, nonsteroidal anti-inflammatory drugs, antihypertensives, or diabetes medications). Comorbidities were similarly coded as a binary variable (present/absent) and included physician-confirmed diagnoses of cardiovascular disease, diabetes, or obesity. These were verified through patient medical records and used both as covariates and moderators in the regression models to assess their interaction with physical activity on recovery outcomes. Age was recorded in years, while gender was categorized as male or female. Socioeconomic status was determined based on a combination of education level and self-reported income and categorized into low, middle, or high socioeconomic status. Comorbidities, including cardiovascular disease, diabetes, and obesity, were recorded based on participants’ medical records and self-reports, with a binary variable indicating the presence or absence of any comorbidity [19]. Smoking status and employment status were likewise collected as categorical variables, with smoking categorized as current smoker or non-smoker, and employment as employed or unemployed.

Data collection process

Data were collected through a combination of patient self-reports, medical records, and standardized assessments, ensuring the use of validated and reliable tools for each outcome and independent variable. This methodological approach facilitated a comprehensive examination of the relationships between physical activity levels and functional recovery, QoL, and psychological well-being, with adjustments for potential confounding variables. To ensure data accuracy and minimize potential biases, several rigorous data collection procedures were implemented. To further address potential sources of bias, blinding was implemented during data entry and analysis, whereby research assistants were unaware of participants’ physical activity groupings. Additionally, validated instruments were used for all primary outcomes to reduce measurement bias. Selection bias was minimized through purposive sampling across multiple clinics within the hospital, and the inclusion of diverse musculoskeletal diagnoses enhanced external validity. Data collectors, including research assistants and physiotherapists involved in the study, underwent specialized training before the commencement of data collection. This training covered standardized protocols for administering questionnaires, recording demographic and clinical information, and addressing common sources of error in self-reported measures. To minimize recall bias, participants were encouraged to reflect on recent activity patterns during questionnaire completion, in accordance with IPAQ-SF instructions. In addition, where feasible, self-reported clinical data were cross verified with participants’ medical records to confirm the presence of comorbidities and validate baseline health status. This cross-verification process ensured consistency between self-reports and documented clinical data, enhancing the accuracy of key variables used in the analysis. While the study relied primarily on self-reported scales, blinded data entry and scoring procedures were implemented. Baseline health status was operationalized as a composite clinical assessment score based on the physiotherapist’s initial evaluation, incorporating range of motion limitations, pain interference during daily activities, and overall physical function. Scores were recorded on a 0–10 scale, with higher scores indicating greater impairment. This score was extracted from standardized intake records and used as a covariate in the regression models.

Sample size calculation

An a priori sample size was determined using G*Power software (version 3.1) to ensure sufficient statistical power for detecting significant effects. Based on effect sizes from similar studies examining how physical activity influences functional recovery and QoL in musculoskeletal rehabilitation, a medium effect size (f² = 0.15) was chosen for the primary analyses. With a significance level of α = 0.05, a target power of 80%, and a model including up to 10 predictor variables (covering physical activity levels, demographic factors, and clinical covariates), the calculation indicated a minimum required sample size of 286 participants.

Statistical analysis

Data were analyzed using SPSS software (Version 24.0). Prior to model interpretation, assumptions for linear regression were evaluated. The normality of residuals was confirmed via visual inspection of Q–Q plots and Shapiro–Wilk tests (p > 0.05), while homoscedasticity and linearity were assessed using residuals versus fitted value plots. No major violations of these assumptions were identified. Multicollinearity diagnostics showed variance inflation factor (VIF) values ranging from 1.04 to 1.28, indicating low multicollinearity and stable model estimation. Associations between physical activity levels and the primary outcomes—functional recovery, quality of life (QoL), and psychological well-being—were examined using Pearson correlation analyses and multiple linear regression models. Physical activity served as the main independent variable, with each outcome assessed in a separate regression model. Covariates were selected a priori based on theoretical relevance and prior evidence. For all models, age, gender, socioeconomic status, and presence of comorbidities were included. Additional covariates—BMI, medication use, smoking status, and employment status—were incorporated into models examining QoL and psychological outcomes due to their known influence on health-related variables. These covariates, described in Table 1, were entered hierarchically to evaluate their individual and combined effects, enhancing the accuracy of physical activity estimates.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Participant Demographics and Clinical Profiles.

https://doi.org/10.1371/journal.pone.0341335.t001

Missing data were assessed for each variable and found to be < 5% overall. As the missingness appeared completely at random (MCAR), listwise deletion was employed to maintain analytical consistency while avoiding the assumptions associated with data imputation. Sensitivity analyses excluding these cases yielded results consistent with the main models. Group differences in functional recovery scores across physical activity levels (low, moderate, high) were evaluated using one-way ANOVA. Significant effects were further examined using Tukey’s post hoc tests, and effect sizes (η²) were calculated to assess magnitude. For QoL and psychological outcomes, ANCOVA was used with physical activity group as the independent factor and relevant covariates included. The ANCOVA for QoL showed a significant main effect of physical activity, F(2, 281) = 5.23, p = 0.006, partial η² = 0.27. Adjusted mean SF-36 scores were 58.76 (SD = 14.21) for the low activity group, 66.04 (SD = 13.56) for moderate, and 70.91 (SD = 12.67) for high activity participants. Bonferroni-adjusted post hoc comparisons indicated significant differences between the low and high groups (p = 0.004, 95% CI [2.45, 11.96]) and between low and moderate groups (p = 0.031, 95% CI [0.45, 10.01]).

To examine potential moderating effects, hierarchical regression models were extended to include interaction terms (e.g., physical activity × age, × gender, × socioeconomic status). Changes in explained variance (ΔR²) were used to assess the contribution of these interaction terms, providing insight into whether the strength of association varied across subgroups.

Clinical significance of the study

The clinical significance of this study lies in demonstrating that higher levels of physical activity are strongly associated with better functional recovery, improved QoL, and enhanced psychological well-being in individuals undergoing physiotherapy for musculoskeletal disorders. Notably, the study highlights that demographic and lifestyle factors, such as age, gender, socioeconomic status, and comorbidities, significantly influence these outcomes. These findings suggest that tailoring physiotherapy treatments to account for a person’s demographic and health background—especially in older adults and those with multiple comorbidities—can lead to better recovery results. Furthermore, the positive impact of physical activity underscores the importance of integrating structured physical activity programs as a key component of rehabilitation. Clinicians should consider these moderating factors when designing personalized treatment plans, ensuring to address obstacles to physical activity, such as socioeconomic challenges or health issues, in order to optimize patient recovery and overall well-being.

Limitations of the study

Reliance on self-reported physical activity data introduces several limitations. First, recall bias may affect the accuracy of responses, particularly in estimating frequency and duration of physical activity over a typical week. Second, social desirability bias may lead participants to overreport their physical activity levels to align with perceived normative expectations, potentially inflating MET-min/week values recorded through the IPAQ-SF. Third, psychological status itself may influence self-report accuracy; individuals experiencing anxiety or depression may underreport activity due to negative self-perception or motivational deficits, while others may overreport as a form of compensation or perceived compliance. These factors introduce measurement error that may attenuate or distort observed associations between physical activity and health outcomes. Although weekly activity logs and trained data collectors were employed to mitigate these risks, the inherent limitations of self-report data should be considered when interpreting the strength and direction of the study’s findings. Although validated questionnaires were used, objective tools like accelerometers or pedometers would yield more precise data and lessen dependence on memory. Additionally, the study did not incorporate objective severity scores such as the numeric rating scale for pain, which may have provided further granularity in assessing baseline condition severity. Future studies should include such metrics to enhance clinical characterization. Future studies could incorporate these devices or weekly activity logs to improve data accuracy. Additionally, the cross-sectional design restricts causal interpretations of how physical activity affects recovery, indicating that longitudinal research is necessary to track changes over time and determine causation.