Musculoskeletal Disorders and Sick Leave among Peruvian Construction Workers: Implications for Safety Management

This study used a quantitative cross-sectional design to obtain a comprehensive understanding of WMSDs in the Peruvian construction sector. The methodological framework was adapted from the protocol described by Al-Khiami et al. [24] to fit the local occupational context. The process comprised two main stages: first, administration of a modified version of the Nordic Musculoskeletal Questionnaire (NMQ), previously validated for construction workers [24], to collect quantitative data on musculoskeletal symptoms in nine body regions; and second, statistical analysis using IBM SPSS Statistics to describe prevalence rates and explore associations between pain reports and selected demographic and work-related variables. The overall methodological workflow is summarised in Figure 1.

Figure 1. Overview of the study process

2.1 Instrument selection, translation and contextual review

For data collection, this study used the questionnaire reported by Al-Khiami et al. [24], which is based on a modified version of the Nordic Musculoskeletal Questionnaire (NMQ) for construction workers. Rather than developing a new instrument, the study adopted this previously published questionnaire because it was aligned with the research objectives and occupational context. The questionnaire covers nine body regions and records symptoms during the previous 12 months, work impairment, and symptom severity using a five-point Likert scale from 1 (very mild) to 5 (very severe). This structure follows the base study [24] and is consistent with recent evidence supporting the semantic, cultural, and functional applicability of the NMQ in Portuguese- and Spanish-speaking settings [25-27]. For application in Peru, the questionnaire was translated into Spanish and underwent a contextual review by the co-authors to improve clarity, wording, and appropriateness for local construction workers. Minor wording adjustments were introduced to improve comprehension while preserving the meaning and structure of the source instrument.

2.2 Sample size determination and recruitment strategy

The sample size was calculated using Cochran’s formula for a finite population, considering an accessible population of approximately 300 workers employed on a building foundation project in Lima, Peru. With a 95 % confidence level (Z = 1.96), a conservative proportion of 0.5, and a 7 % margin of error, the initial sample size for an infinite population $n_0$ was estimated as:

$n_0=\frac{Z^2 \times p(1-p)}{e^2}$   （1）

where, $n_0$ is the required sample size for an infinite population, $Z$ is the standard normal deviate for the chosen confidence level, $p$ is the assumed proportion of workers with the outcome of interest and $e$ is the margin of error. For a finite population of $N \approx 300$, the corrected sample size $(n)$ was then obtained using:

$n=\frac{n_0}{1+\frac{n_0-1}{N}}$   (2)

which yielded a minimum required sample of 119 participants. Recruitment was carried out through consecutive sampling between September and October 2025, during the foundation stage of the project. Eligible workers were approached during the routine pre-shift meeting at the study site, where the study was explained and workers present at that time were invited sequentially to participate until the target sample was reached. Participation was voluntary and limited to active workers aged 18 years or older, with at least one month of continuous employment at the site. Workers with non-work-related injuries in the previous three months were excluded. The sample included workers from trades typically present in early-stage construction, such as steel fixers, concrete workers, bricklayers, helpers and general labourers. Although the data were collected from a single site, the work processes and environmental conditions are representative of urban building projects in Peru.

The study was conducted in accordance with the ethical principles of the Declaration of Helsinki for research involving human participants. No identifying personal data was collected. Participation was strictly voluntary, and all workers were informed about the aims and procedures of the study and provided informed consent before completing the questionnaire. Seventeen invited workers declined to participate. Because recruitment was limited to workers present at the time of field administration, and the presence or absence of other workers was not systematically recorded, a healthy-worker effect or selection bias cannot be ruled out. In addition, the use of a 12-month recall period may have introduced recall bias. Symptom reporting may also have been affected by underreporting, since workers may normalise pain or hesitate to disclose discomfort in a work environment where complaining can be perceived negatively.

2.3 Questionnaire structure and administration

Quantitative data were collected using the questionnaire described in Section 2.1. The instrument began with basic demographic and occupational variables, specifically age, sex, and years of experience in construction. Participants then reported musculoskeletal symptoms across nine body regions: neck, shoulders, elbows, wrists/hands, upper back, lower back, hips/thighs, knees, and ankles/feet. For each region, respondents indicated whether they had experienced pain during the previous 12 months and whether the discomfort had limited their ability to perform work during the same period. When pain was reported, symptom intensity was rated using a five-point Likert scale ranging from 1 (very mild) to 5 (very severe). The questionnaire was administered directly to workers at the study site under researcher's supervision to ensure comprehension and completeness of responses. All completed questionnaires were subsequently reviewed and coded for statistical processing.

2.4 Qualitative data collection

During the survey process, several participants spontaneously shared brief reflections on their experiences of work-related discomfort. In total, 17 reflective comments were collected and analysed using thematic analysis, following the six-phase framework proposed by Braun and Clarke [28]. The researcher independently reviewed and coded all responses, identified recurring ideas and grouped them into broader categories to capture shared perceptions. This process emphasised conceptual clarity and internal consistency rather than inter-rater comparison. The thematic analysis was conducted inductively, allowing patterns to emerge from the data without applying predefined categories. The resulting themes were then contrasted with the quantitative findings to contextualise symptom reporting within broader sociocultural attitudes towards pain in construction work.

2.5 Data analysis

All quantitative data were manually entered into Microsoft Excel, checked for completeness and exported as a CSV file for statistical analysis. The analytical process was conducted using IBM SPSS Statistics version 27, following procedures consistent with those described by Al-Khiami et al. [24]. The analysis involved three main stages. First, normality testing was performed. Data distributions were visually assessed using quantile–quantile (Q–Q) plots to determine whether the use of parametric or non-parametric tests was appropriate.

Second, a descriptive analysis was conducted. Measures of central tendency (mean and median) and dispersion (standard deviation, minimum and maximum) were computed for quantitative variables such as age and years of experience, while frequency distributions were generated for categorical variables representing the presence or absence of WMSD symptoms across the nine body regions. Overall 12-month prevalence (“any body region”) was calculated as the proportion of participants with at least one positive response across the nine NMQ regions. In addition, 95% confidence intervals for overall and regional prevalence estimates were calculated using the Wald (normal approximation) method as implemented in SPSS. This approach was retained as a simple and consistent method for reporting descriptive prevalence estimates across body regions.

Third, an inferential analysis was carried out. Given the binary nature of the pain variables, Kendall’s tau-b (τ) coefficients were computed to explore pairwise associations between body regions. This method is robust to ties and is suitable for non-parametric, ordinal and dichotomous data. To check robustness, phi (φ) coefficients were also calculated. Correlations between pain occurrence and demographic factors (age and years of experience) were likewise tested using non-parametric methods. Given the exploratory nature of this study and the limited sample size, no multivariable models were fitted, and results are interpreted as bivariate associations rather than causal relationships. No formal adjustment for multiple comparisons was applied; therefore, the correlation findings were interpreted as exploratory and with caution.

2.6 Data quality and reproducibility

To ensure data accuracy, a random 10 % of the survey entries were double-checked against the original paper questionnaires to verify correct transcription into the electronic database. No inconsistencies were detected. Data validation procedures included internal range checks and verification of missing values prior to statistical analysis. All analyses were conducted using standardised syntax files in IBM SPSS Statistics version 27 to ensure replicability, and analytical decisions and coding steps were documented to maintain traceability throughout the process. An anonymised dataset and analysis scripts are available from the corresponding author upon reasonable request.

This study examined the prevalence and reporting patterns of WMSDs among construction workers in Peru. The findings showed that 70.0% of participants experienced musculoskeletal discomfort in at least one body region during the previous 12 months, with the lower back (41.7%), shoulders (20.8%), and knees (15.0%) identified as the most affected areas. These results are consistent with the predominance of lumbar and upper-limb strain as key ergonomic concerns in construction work. Pain intensity was predominantly mild to moderate, with severity levels 2 and 3 representing the majority of reports and only isolated cases of severe discomfort. Absenteeism analysis indicated that lower-back pain was the main factor associated with work leave, followed by shoulder-related problems, underscoring the occupational relevance of trunk and upper-limb disorders. Kendall’s tau-b correlation analysis demonstrated weak and localised associations (|τ| < 0.20) between body regions, suggesting a predominantly localised pattern of self-reported discomfort rather than a generalised multi-region pattern in this sample. Although these patterns may be compatible with task-specific loading, objective task or exposure metrics were not collected, and causal inferences cannot be drawn. Age and work experience showed negligible correlations with WMSD occurrence, indicating that demographic characteristics exert limited influence on symptom distribution.

Overall, these findings support the importance of strengthening ergonomic prevention and early symptom recognition in the Peruvian construction sector. However, given the cross-sectional, single-site design and the reliance on self-reported symptoms, the results should be interpreted as descriptive and exploratory rather than causal. Within these limits, practical next steps for occupational safety management include posture-oriented ergonomic training, early reporting and symptom-recognition mechanisms, and awareness programmes that help reduce the normalisation and stigma of pain in construction work.