**1. Introduction**

Worldwide, work-related musculoskeletal symptoms (WRMSs) are a major public health issue, with musculoskeletal conditions contributing the greatest proportion of lost work productivity [1]. Workers with WRMSs may experience pain, numbness, stiffness, and aching in various body parts [2]. Of all working adults, grassroots working-class workers are the most vulnerable to WRMSs because the nature of their work mostly exposes them to the identified risk factors [3].

Evidence has shown multidimensional programs, including exercise, can prevent/manage WRMSs in different contexts: office workers in Portugal [4]; bus drivers in South Korea [5]; workers requiring prolonged standing in Portugal [6]; and underprivileged migrant workers in a community setting in Korea [7]. A review of 61 studies of the effectiveness of workplace intervention for WRMSs showed that exercise programs, such as stretching and resistance-training, have demonstrated positive effects with moderate-to-strong levels of evidence [8]. Although nearly 60% of the WRMS intervention studies were conducted in office-based workplaces, the review indicated that it would be possible to implement the interventions in non-office settings.

Although the workplace has been recommended as a place for promoting physical activity, this concept is not yet accepted widely [9], particularly in small or medium-sized enterprises. Low-skilled workers might not be a beneficiary group because their workplaces most likely do not offer WRMSs prevention/management programs. In addition to their low education levels, this group might have limited access to WRMSs prevention/management knowledge. Community-based WRMSs prevention/management programs could be an innovative way to target low-skilled workers. Despite limited relevant studies having been conducted, one community-based study with Korean–Chinese migrant workers has shown promising results [7]. These Korean–Chinese migrant workers (of whom 90% were domestic or restaurant workers) rarely received any interventions, such as stretching exercises, which are commonly provided to native Koreans in their workplaces. The results of this randomized control trial study were encouraging, and suggested that community-based stretching exercise programs are feasible and effective in helping low-skilled working-class workers to prevent/manage WRMSs [7].

The aim of this study was to evaluate the effectiveness of a multidisciplinary exercise program in managing WRMSs for low-skilled workers in low-income community settings. The hypotheses were that the workers would reduce the number of their body parts with WRMSs (primary outcomes) and improve their workstyles, self-rated health status, self-rated job stress, self-efficacy, social support, mental health, handgrip strength, and spine flexibility (secondary outcomes) between baseline (T0) and immediate (T1) follow-up measures after the program.

#### **2. Methods**

#### *2.1. Study Design, Setting, and Participants*

This was a repeated-measures, single-group design study. The participants were recruited from low-income communities through seven Caritas Community Development Service (CCDS) [10] centers from all three regions of Hong Kong. The inclusion criteria were low-skilled workers aged ≥ 18 working full-time or part-time, with at least one body part that had been affected by WRMSs for at least one month. A "low-skilled" job was defined as work not requiring a high level of skill or without formal qualification requirements [11]. This type of job usually plays a supporting role in an organization [12]. Some of the typical low-skilled jobs are cleaning, laboring, general office work, equipment operation, and helping in restaurants [12–14]. The exclusion criteria were engagement in WRMS prevention programs in the workplace at the time of the study, undergoing medical treatment, or doing stretching exercises more than three times a day, five days per week [7].

#### *2.2. Multidisciplinary Exercise Program*

Ethical approval was obtained from The Hong Kong Polytechnic University (HSEARS20170815005). The program involved eight weekly 90-min sessions, with a group size of 13–22 participants. Each session consisted of a 45-min workshop and then a total of 45 min of stretching (10.5–15 min) and muscle strengthening (5–10 min) exercises, followed by a question-and-answer period (about 20 min).

The eight WRMS workshop topics, conducted by nurses, physiotherapists, traditional Chinese medicine (TCM) doctor, and social workers, included: (1) Exercise; (2) Health and work; (3) Manual handling; (4) Non-pharmaceutical pain management; (5) TCM; (6) Mental health; (7) Laws and compensation; and (8) Community resources. The teaching material was validated by a panel of seven experts representing the fields of nursing, occupational health, TCM, physiotherapy, and social workers. The content validity of the teaching material was 1.00.

The exercise regime was based on exercises suggested by occupational safety and health authorities [15,16] and validated by four physiotherapists and three nurses. The 21-movement stretching exercise involved neck (6 movements), shoulders (1), upper extremities (2), wrists (4), back (6), and lower extremities (2). The 10-movement muscle strengthening exercise involved neck (6 movements), upper extremities (2) and lower extremities (2). Each movement took 10 seconds and was repeated three times.

The data were collected from a questionnaire (30 min) and health assessments (30 min) conducted at the baseline (T0) and immediately after the program completion (T1). To ensure the consistency of the data-collection procedure, the personnel responsible for collecting the data in each center were trained for two hours by a member of the research team, and their return demonstrations achieved satisfactory results. The same models of equipment were used for the health assessments in each center, to ensure consistency.

### *2.3. Measurements*

The questionnaire was developed based on previous studies and a literature review [17–25]; it was validated by the same panel as the teaching material. The content validity index was 0.99. Some parts of the questionnaire had been tested previously by the investigators or in other local studies. Below is a description of each section:


Spine flexibility was assessed using the sit-and-reach method. The subjects were instructed to sit on the floor with both legs straight and their soles touching the sit-and-reach box. They were then asked to bend their backs, with both palms facing down and fingertips straight ahead, to push the scale on the top of the box. The distance pushed was measured in centimeters. Other health assessments, such as body mass index (BMI), hip-waist ratio (HWR), blood pressure (BP), blood glucose (BG), cholesterol (using a glucometer), and handgrip strength, were also measured.

#### *2.4. Data Analyses*

All the analyses were computed by using the Statistical Package for Social Sciences (SPSS, version 24.0). To examine what factors influenced the number of body parts with WRMSs, which was treated as a continuous variable, at the baseline, non-parametric Spearman correlation analyses were used. A series of linear regression analyses with all variables entered at once into the regression model were employed. Factors found to be correlated significantly in the non-parametric test were entered into the univariate regression analyses, and a multivariate regression model was used to identify the contributing factors. The Wilcoxon Signed Ranks tests were adopted to observe any significant changes in the number of body parts with WRMSs and associated factors over the eight-week period. Non-parametric Spearman correlation analyses were then conducted to test whether there was any association between the measured factors and the changes in the number of body parts with WRMSs. For all significant correlated factors, mutually adjusted multiple linear regression analyses were employed, with age and gender included in the model as covariates, to confirm what factors truly contributed to the reduction in the number of body parts with WRMSs. Statistical significance was determined by two tailed tests, with a *p*-value of < 0.05.

#### **3. Results**

#### *3.1. Characteristics of Participants at Baseline (T0)*

The average age of the 105 participants was 50.5 ± 8.7 years (ranging from 31 to 67). Over 80% (n = 87) of them were female, 68.6% (n = 72) were married, and 68.6% (n = 72) had completed secondary school. Regarding their employment status, 30.5% (n = 32) were part-time employees. They reported having experienced WRMSs in an average of three body parts at T0. In addition, female workers with depression or anxiety had more body parts with WRMSs (*p* < 0.05). No significant differences were observed among other demographic characteristics (Table 1).

#### *3.2. Factors Associated with Body Parts with WRMSs at Baseline (T0)*

Table 2 shows that the number of body parts with WRMSs was significantly correlated with the workstyle of working through pain, social reactivity, the total score for workstyle, self-rated health status, and depressive symptoms.

Under univariate regression analyses, gender, the workstyle of working through pain, social reactivity, depressive symptoms, anxiety and self-rated health status were all found to be significantly associated with the number of body parts with WRMSs. Multivariate regression analysis indicated that gender and the workstyle of working through pain remained as significant factors. Workstyle seems to have been the most influential factor in affecting the number of body parts with WRMSs (Table 3).

#### *3.3. Factors Associated with the Change before and after Intervention (T1-T0)*

Immediately after the intervention (T1), 86 (81.9%) participants took part in the follow-up data collection. However, there were no significant differences found in the reported number of body parts with WRMSs between the participants who joined T1 and those who dropped out. Table 4 shows that the number of body parts with WRMSs, sit-and-reach distance, handgrip strength, workstyle of working through pain, and self-rated job stress had changed significantly over time. The number of body parts with WRMSs was reduced by a mean value of 1.05. The workstyle of working through pain and self-rated job stress was also reduced significantly from T0 to T1 (−1.70 and −0.29, respectively). Also, both the sit-and-reach distance and the right-hand grip strength were enhanced by mean values of 1.95 cm and 0.40 kg, respectively, at T1.

Under the univariate model, it was demonstrated that gender and change in the workstyle of working through pain and right-hand grip strength were significantly associated with the change in the number of body parts with WRMSs. After adjusting for potential confounders, changes in the workstyle of working through pain and self-rated health status were the significant factors influencing the reduction in the number of body parts with WRMSs (Table 5).

#### *3.4. Evaluation of the Program*

With a response rate of 93.02% (n = 80), the participants rated all aspects of the program positively, with means ranging from 3.33 to 3.44 out of 4.00.


**Table 1.** Characteristics of participants, and the association of factors and number of body parts with WRMSs at baseline (T0).


**Table 1.** *Cont*.
