Perceived Heat-Related PPE Discomfort and Self-Reported Routine Use Among Construction Workers in Riyadh, Saudi Arabia: An Exploratory Survey with Descriptive Site-Observation Summaries

1. Introduction

Construction remains a high-risk occupational sector because work environments change rapidly and multiple hazards can coexist. Work at height, excavation, lifting, moving plant, welding, dust, noise and manual handling are common examples. Personal protective equipment (PPE) is not a substitute for elimination, substitution, engineering or administrative controls, but it remains an important final barrier where residual risk cannot be adequately controlled [1,2,3].

Although PPE sits near the bottom of the hierarchy of controls, it can only protect workers when it is selected correctly and worn throughout the exposure. Non-compliance is not limited to complete non-use. Poor fit, incorrect adjustment, temporary removal and use only when a supervisor is present can all reduce protection. Construction research has linked these behaviours to PPE availability, comfort, fit, risk perception, supervision, training and safety management [4,5,6,7,8,9,10].

Hot-climate construction adds a practical difficulty: environmental heat, physical work and protective clothing can combine to increase sweating, fatigue and physiological strain. Equipment that is heavy, poorly fitted or restricts airflow may become harder to tolerate as a shift progresses. Occupational heat guidance and recent PPE studies support these concerns, although the reasons for adjusting or removing a particular item depend on the task and workplace [3,11,12,13,14,15,16,17,18,19,20,21,22].

Riyadh is a relevant setting because high-temperature working conditions are predictable rather than exceptional. Construction workers may move between outdoor, partially shaded and temporary recovery areas while using several PPE items. The sector also relies heavily on a multinational workforce, making understandable, two-way safety communication particularly important [14,23,24,25,26,27,28].

Previous studies have examined PPE use in relation to knowledge, training, supervision, safety climate, comfort, fit, communication and management practices [4,5,6,7,8,9,10,11,23,24,29,30,31,32]. Direct Saudi research has also documented substantial occupational heat exposure among construction workers [14,33]. However, evidence specifically connecting perceived heat-related discomfort with observed PPE behaviour in Riyadh remains limited.

This exploratory study examined perceived heat-related PPE discomfort and self-reported routine PPE use among construction workers in Riyadh. It assessed whether heat-related discomfort, supervisor or leadership influence, and the language accessibility of training and safety instructions were associated with workers describing PPE as part of their daily routine rather than something worn when reminded. It also described unlinked site-observation summaries from February to April 2026 and considered practical implications for health, safety and environment (HSE) teams.

2. Literature Review and Theoretical Framework

2.1. PPE Compliance in Construction Environments

PPE compliance has been widely examined because construction workers encounter changing hazards that may require continuous protective behaviour. PPE does not replace higher-order controls, but it remains essential where residual risk persists. Studies identify interacting personal, technological and organisational influences, including task demands, safety climate, supervision, training, equipment availability, perceived usefulness, fit and comfort [4,5,6,7,8,9,10,11,23,29,30,31,32].

There is seldom a single explanation for PPE non-compliance. Across construction and other blue-collar settings, recurring influences include weak risk perception, physical discomfort, poor fit, inconsistent enforcement, limited training, equipment availability and the wider safety climate [4,6,7,8,10,11,31]. PPE use is therefore better understood as a behavioural and organisational issue as well as an equipment issue.

2.2. Heat Stress and Construction Safety

Occupational heat stress arises from the combined effects of environmental conditions, metabolic workload, clothing or PPE, and insufficient recovery. Construction workers are vulnerable because much of their work is physically demanding and performed outdoors or in semi-exposed areas. Heat exposure is associated with physiological strain, dehydration, reduced work capacity and increased occupational-injury risk [1,3,12,13,14,15,16,17,18,19,20,21,22,33].

Heat may shape day-to-day behaviour through fatigue and discomfort. Workers may slow down, seek shade, take additional breaks, or adjust clothing and equipment. The present cross-sectional data cannot determine whether heat caused any such response, but international reviews and field studies provide a sound basis for including heat in task planning, recovery arrangements and PPE management [3,12,13,14,15,16,17,18,19,20,21,22].

2.3. PPE Discomfort, Usability and Behavioural Adaptation

PPE discomfort is a recognised barrier to sustained use. Comfort is affected by weight, breathability, fit, material, task compatibility and wear duration. Construction studies report that physical discomfort, poor fit, perceived ease of use and safety-management practices can influence PPE acceptance and use [4,6,7,8,9,10,11,31,34].

From a site HSE perspective, PPE adjustment should prompt investigation of the immediate reason as well as correction of the exposure. Fogging, poor fit, reduced dexterity or breathing resistance can make safe task performance more difficult. Enforcement should therefore be combined with suitable selection, fit, worker feedback and higher-order controls; the protective performance required for the hazard must not be compromised [6,8,9,11,34].

2.4. Supervision, Safety Climate and Heat-Stress Controls

Supervisors can influence PPE behaviour by modelling correct use, correcting unsafe practices, explaining requirements, reinforcing toolbox talks and checking availability and suitability. Construction research links management practices, supervision and safety climate with safety behaviour and PPE acceptance, although effects vary across settings and measures [4,5,8,9,10,24,32].

In extreme heat, supervision should extend beyond enforcement to monitoring fatigue, hydration, recovery and heat-illness symptoms. International and Saudi guidance supports acclimatization, accessible water, shaded or cooled recovery, work-rest planning and environmental assessment such as the Wet Bulb Globe Temperature (WBGT) method [3,12,13,14,18,21,26,35].

2.5. Communication and Multinational Workforce Challenges

Communication is important because workers must understand what PPE is required, why it is required, how it should be worn and when it should be replaced. Research in construction workgroups shows that safety climate, crew cohesion and supervisor-worker communication patterns shape the exchange and understanding of safety information [25,27,28].

Many respondents reported language understanding outside Arabic and English. In this setting, safety information needs plain wording, competent translation, visual explanation, practical demonstration and a check that the message was understood. Research on multinational construction projects also warns against depending uncritically on co-workers as informal interpreters. These considerations are especially important for heat symptoms and for reporting PPE that is damaged, uncomfortable or poorly fitted [11,25,26,27,28].

2.6. COM-B Model as the Theoretical Framework

The Capability, Opportunity, Motivation-Behaviour (COM-B) model explains behaviour through capability, opportunity and motivation [36]. Capability includes the knowledge, skill and physical ability needed to use PPE correctly. Opportunity includes environmental and social conditions such as suitable PPE, shade, hydration, supervision and accessible communication. Motivation includes the reflective and automatic processes that influence behaviour. A recent multinational blue-collar study provides direct support for using an accessible COM-B framework to examine PPE use [11].

In this study, COM-B was used as an interpretive framework rather than a tested latent measurement model. Heat-related discomfort may affect physical capability, while suitable PPE, shade, hydration, understandable training and clear instructions form part of physical and social opportunity. Supervisor or leadership influence may also shape social opportunity and motivation. These mechanisms remain interpretive because the cross-sectional design cannot establish mediation, temporality or causality [11,36].

2.7. Research Gap and Study Contribution

The literature establishes that PPE use is influenced by behavioural, organisational, ergonomic and communication-related factors, while occupational heat is an independent health and safety hazard. Less evidence is available on how workers describing PPE as routine experience heat-related discomfort, supervisor or leadership influence and language accessibility in Riyadh construction settings. This study addresses that context-specific gap but does not test a causal heat-exposure pathway.

The study adds anonymous survey evidence from seven active construction sites in Riyadh together with descriptive monthly site-observation summaries. The survey and observations were not linked to the same workers and are therefore interpreted separately. The study is intended to identify associations and questions for further research, rather than to confirm causal effects.

3. Materials and Methods

3.1. Research Design and Setting

Two quantitative data sources were used: a cross-sectional anonymous worker survey and monthly descriptive site-observation summaries. The survey examined associations among selected self-reported and rule-coded variables, whereas the observation summaries described work-front patterns from February to April 2026. The two sources were not linked at worker level. The report follows relevant STROBE recommendations for observational studies [37].

3.2. Population, Sampling and Response Rate

Workers at seven participating construction sites formed the accessible study population. About 120 workers were approached and 85 valid questionnaires were returned, giving an approximate response yield of 71%. General labourers made up most of the sample. Site selection and participant recruitment were non-probability based; the findings therefore describe the participating sites and are not intended to represent all construction workers in Riyadh or Saudi Arabia.

3.3. Data Collection Instruments

Data were collected using a structured questionnaire and a PPE observation checklist. The questionnaire included five-point Likert items, a multiple-response question on workplace barriers and short open-text questions addressing PPE discomfort, supervisor or co-worker influence and whether PPE use was routine or reminder-dependent. Monthly observation summaries reported correct use, incorrect use, non-compliance, supervisor presence, heat discomfort and heat-attributed PPE removal or adjustment. The observation summaries contained percentages but not underlying event counts, observer identifiers, worker and task distribution, repeat-observation structure or inter-rater agreement; they were therefore analysed descriptively.

3.4. Operational Definitions and Data Analysis

For this exploratory secondary analysis of the original Microsoft Forms survey export, the primary outcome was self-reported routine PPE use. Responses to the question asking whether PPE was part of the daily routine or worn only when reminded were coded as routine use when they referred to routine, habit or an equivalent affirmative description, and as reminder-dependent use when they referred to being reminded or forgetting. All 85 responses were classifiable.

Four predictors were defined before contingency tables were generated. Heat-related discomfort was coded from the open-text discomfort question: responses containing heat, hot, sweat, sweaty or sweating were coded as heat-related, while other valid discomfort responses formed the comparison category; one missing response was excluded. Supervisor or leadership influence was coded from the relevant open-text question when a response explicitly referred to a supervisor, leader, leadership modelling or leading by example; co-worker or other valid responses formed the comparison category, and three missing responses were excluded. Training-language accessibility and safety-instruction language accessibility were based on their exact Likert items, with scores of 4-5 coded as accessible and scores of 1-3 coded as lower or neutral. Full item wording and coding rules are provided in File S1.

Survey data were checked against the original Microsoft Forms export and analysed in Microsoft Excel and jamovi 2.6 [38]. Pearson chi-square tests were used for the four 2 × 2 tables; all expected cell counts exceeded five. Results are reported with two-sided unadjusted p-values, Holm-adjusted p-values, Cramer’s V, odds ratios (ORs) and 95% confidence intervals (CIs). A significance threshold of p < 0.05 was used. Cronbach’s alpha was not calculated for these variables because the analysis used separate single items and rule-coded responses rather than an interchangeable multi-item scale.

3.5. Ethical Considerations

Participation was voluntary and anonymous, and no directly identifying information was collected. Written gatekeeper permission to access the participating project sites and conduct the proposed surveys and observational activities was obtained from the relevant project and site management representatives in November and December 2025. Gatekeepers facilitated site access but did not access completed responses. Data were stored securely and used only for academic research. The Atlantic Technological University (ATU) Faculty of Science and Health Research Ethics Committee confirmed ethical approval on 5 February 2026 (REC Reference No. MSC_EHS_2025-26_33), valid until 31 August 2026. Participant recruitment, survey administration and research-specific observational data collection commenced only after ethical approval and the necessary organisational permissions were in place.

3.6. Use of Generative AI-Assisted Tools

OpenAI ChatGPT (OpenAI, San Francisco, CA, USA; accessed 5 August 2026) was used as an assistive tool for language refinement, manuscript organisation and checking the presentation of statistical calculations derived from the original survey export. The authors independently reviewed and verified the coding rules, calculations, references, interpretations and final wording and retain full responsibility for the manuscript. The tool was not used to generate, replace or alter the original research data.

4. Results

4.1. Respondent Profile

The survey yielded 85 valid responses from seven construction sites. Of these, 75 respondents (88.2%) were general labourers, eight were skilled labourers and two did not report a role. Forty respondents selected a language category other than Arabic or English. The broad other-language category does not permit language-specific conclusions.

Table 1. Respondent profile and key demographic characteristics.

Variable	Category	Frequency
Age	Under 25	17
Age	25-34	36
Age	35-44	24
Age	45+	8
Construction experience	Less than 1 year	10
Construction experience	1-3 years	23
Construction experience	3-5 years	25
Construction experience	More than 5 years	26
Construction experience	Not specified / not reported	1
Role	General labourer	75
Role	Skilled labourer	8
Role	Not specified / not reported	2
Language understanding	Arabic	21
Language understanding	English	22
Language understanding	Other languages	40
Language understanding	Not specified / not reported	2

Table 1. Respondent profile and key demographic characteristics.

Variable	Category	Frequency
Age	Under 25	17
Age	25-34	36
Age	35-44	24
Age	45+	8
Construction experience	Less than 1 year	10
Construction experience	1-3 years	23
Construction experience	3-5 years	25
Construction experience	More than 5 years	26
Construction experience	Not specified / not reported	1
Role	General labourer	75
Role	Skilled labourer	8
Role	Not specified / not reported	2
Language understanding	Arabic	21
Language understanding	English	22
Language understanding	Other languages	40
Language understanding	Not specified / not reported	2

4.2. Descriptive Findings on Heat, PPE Comfort and Training

All respondents agreed that PPE helps protect against serious injury, and 84 of 85 (98.8%) reported that PPE becomes uncomfortable in hot weather. Table 2 reports the number of agreeing or strongly agreeing responses and the valid denominator for each item. These figures describe perceptions and are not direct measurements of heat exposure or observed PPE removal.

Table 2. Worker knowledge, training and behavioural perceptions. Percentages use the valid response denominator shown for each item. PPE, personal protective equipment.

Statement	Agree/Strongly Agree, n/N (%)
Understand when PPE is required	65/85 (76.5%)
Know how to wear and adjust PPE correctly	52/85 (61.2%)
PPE training helps correct PPE use	84/85 (98.8%)
Hands-on demonstrations improve PPE understanding	84/84 (100%)
Practical demonstrations are more effective than classroom-based training	84/85 (98.8%)
PPE training delivered in an understandable language	34/85 (40.0%)
Understand the process for reporting PPE defects	21/85 (24.7%)
PPE helps protect against serious injury	85/85 (100%)
Motivated to wear PPE even without supervisor presence	53/85 (62.4%)
PPE becomes uncomfortable during hot weather	84/85 (98.8%)
Confident speaking up about uncomfortable PPE	20/85 (23.5%)

Table 2. Worker knowledge, training and behavioural perceptions. Percentages use the valid response denominator shown for each item. PPE, personal protective equipment.

Statement	Agree/Strongly Agree, n/N (%)
Understand when PPE is required	65/85 (76.5%)
Know how to wear and adjust PPE correctly	52/85 (61.2%)
PPE training helps correct PPE use	84/85 (98.8%)
Hands-on demonstrations improve PPE understanding	84/84 (100%)
Practical demonstrations are more effective than classroom-based training	84/85 (98.8%)
PPE training delivered in an understandable language	34/85 (40.0%)
Understand the process for reporting PPE defects	21/85 (24.7%)
PPE helps protect against serious injury	85/85 (100%)
Motivated to wear PPE even without supervisor presence	53/85 (62.4%)
PPE becomes uncomfortable during hot weather	84/85 (98.8%)
Confident speaking up about uncomfortable PPE	20/85 (23.5%)

Table 3. Reported environmental and operational factors making PPE use difficult. Multiple selections were permitted; percentages use N = 85 and therefore do not sum to 100%.

Factor	Frequency	Percentage
Heat / hot weather	85	100%
Time pressure	25	29.4%
Long working hours	24	28.2%
Physically demanding tasks	24	28.2%
Heavy workload	16	18.8%
Other	10	11.8%

Table 3. Reported environmental and operational factors making PPE use difficult. Multiple selections were permitted; percentages use N = 85 and therefore do not sum to 100%.

Factor	Frequency	Percentage
Heat / hot weather	85	100%
Time pressure	25	29.4%
Long working hours	24	28.2%
Physically demanding tasks	24	28.2%
Heavy workload	16	18.8%
Other	10	11.8%

Table 4. Coded themes from 84 valid open-text responses describing PPE discomfort. A response could contribute to more than one theme, so percentages do not sum to 100%.

Coded discomfort theme	Frequency	Percentage of 84 valid responses
Weight or rigidity	29	34.5%
Heat and sweating	26	31.0%
Fit or pressure	12	14.3%
General discomfort or fatigue	8	9.5%
Pain or physical discomfort	6	7.1%
Breathing restriction	5	6.0%
Fogging or visibility	2	2.4%
Slipperiness	1	1.2%

Table 4. Coded themes from 84 valid open-text responses describing PPE discomfort. A response could contribute to more than one theme, so percentages do not sum to 100%.

Coded discomfort theme	Frequency	Percentage of 84 valid responses
Weight or rigidity	29	34.5%
Heat and sweating	26	31.0%
Fit or pressure	12	14.3%
General discomfort or fatigue	8	9.5%
Pain or physical discomfort	6	7.1%
Breathing restriction	5	6.0%
Fogging or visibility	2	2.4%
Slipperiness	1	1.2%

4.3. Exploratory Association Analyses

Four exploratory Pearson chi-square tests examined associations with self-reported routine PPE use. Open-text variables were coded using the rules described in Section 3.4, and the two language-accessibility items were grouped as 4-5 versus 1-3. Both unadjusted and Holm-adjusted p-values are reported because four comparisons were made.

Table 5. Observed frequencies: heat-related discomfort vs self-reported routine PPE use.

Heat-related discomfort	Routine PPE use	Reminder-dependent use	Row total
Heat-related discomfort mentioned	21	5	26
Other discomfort	32	26	58
Column total	53	31	84

Table 5. Observed frequencies: heat-related discomfort vs self-reported routine PPE use.

Heat-related discomfort	Routine PPE use	Reminder-dependent use	Row total
Heat-related discomfort mentioned	21	5	26
Other discomfort	32	26	58
Column total	53	31	84

Heat-related discomfort was associated with self-reported routine PPE use, χ²(1, N = 84) = 5.05, unadjusted p = 0.025, Holm-adjusted p = 0.025, Cramer’s V = 0.245, OR = 3.41, 95% CI [1.13, 10.29]. Routine use was reported by 21 of 26 workers (80.8%) whose discomfort response mentioned heat or sweating, compared with 32 of 58 (55.2%) whose response described other forms of discomfort. This positive direction does not indicate that heat improved compliance; routine users may have worn PPE for longer and therefore experienced or reported heat-related discomfort more often.

Table 6. Observed frequencies: supervisor or leadership influence vs self-reported routine PPE use.

Reported influence	Routine PPE use	Reminder-dependent use	Row total
Supervisor/leadership influence	45	5	50
Co-worker/other influence	7	25	32
Column total	52	30	82

Table 6. Observed frequencies: supervisor or leadership influence vs self-reported routine PPE use.

Reported influence	Routine PPE use	Reminder-dependent use	Row total
Supervisor/leadership influence	45	5	50
Co-worker/other influence	7	25	32
Column total	52	30	82

Supervisor or leadership influence was associated with self-reported routine PPE use, χ²(1, N = 82) = 39.03, unadjusted p < 0.001, Holm-adjusted p < 0.001, Cramer’s V = 0.690, OR = 32.14, 95% CI [9.23, 111.90]. Routine use was reported by 45 of 50 workers (90.0%) whose response referred to supervisor or leadership influence and by 7 of 32 (21.9%) whose valid response referred to co-workers or another influence. The large effect should be treated as exploratory because both variables were derived from short self-reported responses and the confidence interval was wide.

Table 7. Observed frequencies: training-language accessibility vs self-reported routine PPE use.

Training-language accessibility	Routine PPE use	Reminder-dependent use	Row total
Accessible (Likert 4-5)	29	5	34
Lower/neutral (Likert 1-3)	25	26	51
Column total	54	31	85

Table 7. Observed frequencies: training-language accessibility vs self-reported routine PPE use.

Training-language accessibility	Routine PPE use	Reminder-dependent use	Row total
Accessible (Likert 4-5)	29	5	34
Lower/neutral (Likert 1-3)	25	26	51
Column total	54	31	85

Training-language accessibility was associated with self-reported routine PPE use, χ²(1, N = 85) = 11.59, unadjusted p = 0.001, Holm-adjusted p = 0.002, Cramer’s V = 0.369, OR = 6.03, 95% CI [2.02, 18.06]. Routine use was reported by 29 of 34 workers (85.3%) who agreed that PPE training was delivered in a language they understood, compared with 25 of 51 (49.0%) in the lower or neutral group.

Table 8. Observed frequencies: safety-instruction language accessibility vs self-reported routine PPE use.

Safety-instruction language accessibility	Routine PPE use	Reminder-dependent use	Row total
Accessible (Likert 4-5)	29	6	35
Lower/neutral (Likert 1-3)	25	25	50
Column total	54	31	85

Table 8. Observed frequencies: safety-instruction language accessibility vs self-reported routine PPE use.

Safety-instruction language accessibility	Routine PPE use	Reminder-dependent use	Row total
Accessible (Likert 4-5)	29	6	35
Lower/neutral (Likert 1-3)	25	25	50
Column total	54	31	85

Safety-instruction language accessibility was associated with self-reported routine PPE use, χ²(1, N = 85) = 9.59, unadjusted p = 0.002, Holm-adjusted p = 0.004, Cramer’s V = 0.336, OR = 4.83, 95% CI [1.71, 13.66]. Routine use was reported by 29 of 35 workers (82.9%) who agreed that safety instructions were communicated in a language they understood, compared with 25 of 50 (50.0%) in the lower or neutral group.

Table 9. Summary of Pearson chi-square analyses. The p column is unadjusted; Holm p accounts for four comparisons. All results remain exploratory because the variables were self-reported and two were rule-coded from open-text responses.

Relationship	N	χ2(1)	p	Holm p	Cramer’s V	Interpretation
Heat discomfort vs routine use	84	5.05	0.025	0.025	0.245	Significant
Supervisor/leadership influence vs routine use	82	39.03	<0.001	<0.001	0.690	Significant
Training language vs routine use	85	11.59	0.001	0.002	0.369	Significant
Safety-instruction language vs routine use	85	9.59	0.002	0.004	0.336	Significant

Table 9. Summary of Pearson chi-square analyses. The p column is unadjusted; Holm p accounts for four comparisons. All results remain exploratory because the variables were self-reported and two were rule-coded from open-text responses.

Relationship	N	χ2(1)	p	Holm p	Cramer’s V	Interpretation
Heat discomfort vs routine use	84	5.05	0.025	0.025	0.245	Significant
Supervisor/leadership influence vs routine use	82	39.03	<0.001	<0.001	0.690	Significant
Training language vs routine use	85	11.59	0.001	0.002	0.369	Significant
Safety-instruction language vs routine use	85	9.59	0.002	0.004	0.336	Significant

4.4. PPE Observation Findings

Across the February-April observation summaries, correct PPE use fell from 78% to 65%, while heat-attributed PPE removal or adjustment rose from 25% to 48% (Figure 1 and Figure 2). Table 10 gives monthly percentages for supervisor presence and observed correct PPE use during periods recorded as active supervision. Underlying monthly counts, worker and task composition, site distribution and measured heat exposure were unavailable. Confidence intervals and formal trend tests were therefore not calculated, and the changes cannot be attributed to heat or supervision.

Figure 1 and Figure 2 show the three monthly summaries. The connecting lines are included only to make the month-to-month pattern easier to read; they are not fitted or statistically tested trends.

5. Discussion

5.1. Heat-Related Discomfort and Routine PPE Use

Workers whose open-text response mentioned heat or sweating were more likely to describe PPE as part of their routine. This should not be interpreted as evidence that heat improves compliance. A plausible alternative is that workers who wear PPE more consistently experience longer exposure to its thermal burden and are therefore more likely to report heat-related discomfort. The aggregate observation summaries moved differently between February and April, with lower correct use and more heat-attributed adjustment. As the survey and observations were unlinked and neither WBGT nor observation counts were available, the difference cannot be resolved from this dataset. Established evidence still supports preventive control of occupational heat [3,12,13,14,15,16,17,18,19,20,21,22,33].

5.2. PPE Discomfort and Worker Behaviour

The open-text coding identified weight or rigidity most frequently, followed by heat and sweating, fit or pressure, general discomfort or fatigue, pain, breathing restriction, fogging and slipperiness. Several responses contained more than one concern. These themes are consistent with construction PPE research on fit, comfort, ease of use and safety-management practices [4,6,7,8,9,10,11,31,34]. The findings identify usability concerns but do not establish which item was adjusted, when this occurred or whether it produced unsafe exposure.

5.3. Supervisor Influence and Compliance Reinforcement

Supervisor or leadership influence showed the strongest association with routine PPE use. This pattern is compatible with research linking supervision, management practice, safety climate and leadership with safety behaviour and PPE acceptance [4,8,9,10,24,32]. However, the very large odds ratio should not be interpreted as a causal effect. The predictor and outcome were brief self-reported responses, the coding was simplified and unmeasured site or crew characteristics may have contributed to the association. Replication with a validated scale and prospective observations is needed.

5.4. Training and Communication

Workers who agreed that training or safety instructions were provided in a language they understood were more likely to describe PPE use as routine. These associations remained significant after Holm correction, but they do not prove that language accessibility caused routine use. Both measures were single Likert items, neutral responses were grouped with lower responses and all data came from the same survey. The findings nevertheless support previous construction research emphasising two-way communication, practical training and understandable language [11,25,27,28,32].

5.5. COM-B Interpretation

COM-B provides a useful interpretive structure, although the model was not tested statistically. Heat-related discomfort may reflect the physical demands of sustained PPE use; suitable PPE, shade and hydration form part of physical opportunity; and supervisor or leadership influence and understandable communication form part of social opportunity. Routine use may also reflect motivation reinforced by workplace expectations. Future studies should use validated construct-specific measures before testing these pathways [11,36].

5.6. Contribution to Knowledge

The study contributes an auditable exploratory secondary analysis of the original Microsoft Forms survey export and shows that routine PPE use co-occurred with reported heat discomfort, supervisor or leadership influence and language accessibility. The direction of the heat association is important: it may reflect discomfort arising from consistent use rather than heat encouraging compliance. The separate observation summaries recorded lower correct use and more heat-attributed adjustment, but they cannot confirm or contradict the survey at worker level. The study therefore defines a clearer question for longitudinal research using measured exposure, validated measures and linked observations.

6. Practical Implications

The practical recommendations below draw on the study context and the cited literature; they were not evaluated as interventions. PPE checks should form part of heat-risk management, alongside accessible drinking water, shaded or cooled recovery, acclimatization and work-rest planning based on measured conditions [3,12,13,14,18,21,26,35].

Supervisors should monitor PPE use and signs of heat strain together, model correct behaviour and ensure that workers can raise fit or comfort concerns without blame. When equipment has been adjusted, supervisors should establish why, correct any immediate exposure and escalate repeated usability problems. This recommendation is consistent with earlier research and the exploratory association in this sample; it is not based on a confirmed causal effect [4,9,10,24,32].

PPE procurement should consider fit, task compatibility, fogging, dexterity, breathability and worker feedback while maintaining the protection level required by the hazard and applicable standards. Climate suitability must not be achieved by reducing protective performance [6,8,9,11,34].

Toolbox talks should use plain language, appropriate professional translation, visual demonstrations and confirmation of understanding. Heat symptoms, hydration, recovery and PPE reporting arrangements should be explained in a form workers can apply at the work front [11,25,26,27,28].

Worker feedback should be incorporated into PPE selection, fit checks and replacement processes. Workers should be able to report uncomfortable, damaged or poorly fitting PPE without blame, while unsafe removal remains subject to immediate correction.

7. Limitations and Future Research

The study included 85 valid responses from seven accessible sites, and 88.2% of respondents were general labourers. Neither site selection nor participant recruitment was probability-based, and the survey export did not contain a site identifier that could support adjustment for clustering. Selection bias and unmeasured site or crew effects are therefore possible, and the results should not be generalised to all construction workers in Riyadh or Saudi Arabia. The cross-sectional, self-reported design is also open to recall, social-desirability and common-method bias and cannot establish temporal or causal relationships.

The observation period covered February–April 2026 rather than the peak summer months. WBGT, workload, clothing-adjustment factors and individual physiological strain were not measured. Monthly observation counts, site and task distribution, repeat-observation structure and inter-rater reliability were also unavailable, and observations were not linked to survey respondents. The percentages can therefore describe the recorded pattern but cannot establish an exposure-response relationship or a supervisory effect [3,14,18,21,22,35].

The primary outcome and two predictors were coded from brief open-text responses for this exploratory secondary analysis of the original Microsoft Forms survey export. Although a transparent rule-based codebook was used, the categories were not independently validated and may simplify workers’ intended meanings. The two language-accessibility predictors were single items, and grouping neutral with lower responses reduced the information in the five-point scale. Four tests were conducted and adjusted using Holm’s method; the associations remained statistically significant, but effect estimates—particularly for supervisor or leadership influence—had wide confidence intervals. Future studies should use validated multi-item measures, pre-specified hypotheses and designs that account for worker, crew and site clustering.

Future research should link task-level PPE observations with WBGT, workload, PPE ensemble, supervisor presence and heat-control availability across peak-summer and cooler periods. Qualitative interviews can then examine why workers adjust PPE and how fit, production pressure, reporting confidence and language influence behaviour. A prospective multilevel design would be better suited to testing whether supportive supervision or climate-suitable PPE moderates heat-related non-compliance.

8. Conclusions

Among 85 workers from seven Riyadh construction sites, heat-related PPE discomfort was widely reported. Workers whose discomfort response mentioned heat or sweating were more likely to describe PPE as part of their routine, but this positive association may reflect the discomfort experienced by consistent users rather than a beneficial effect of heat. The monthly site-observation summaries also changed between February and April, but the available data do not show that heat caused those changes.

Supervisor or leadership influence showed the strongest association with routine PPE use. Training delivered in an understood language and safety instructions communicated in an understood language were also associated with routine use after correction for multiple testing. These findings are exploratory and must be interpreted in light of non-probability sampling, short self-report measures, post hoc coding, missing site identifiers and unlinked aggregate observations.

Heat controls remain necessary regardless of these associations. International guidance and Saudi field evidence support heat-risk assessment, hydration, acclimatization, recovery and suitable PPE [12,13,14,22,26,35]. In practice, PPE comfort and worker feedback should be reviewed alongside visible supervision and multilingual communication. The next research step is a prospective study based on measured heat exposure, validated compliance measures and worker-linked observations over time.