Physical therapists (PTs) have a high risk of developing musculoskeletal pain (MP) due to the physically demanding nature of their work tasks. Experience or the specialty area, have been associated with MP, however, previous studies are few and small. The aim of this study was to investigate the association between work-related factors and MP among PTs.
In this cross-sectional study, we collected information about MP and work-related factors of 1006 PTs using an online questionnaire. Associations between various work-related factors and MP were modelled using logistic regression controlled for various confounders.
Neck (57%) and low back pain (49%) were most common. Work-related factors associated with higher risk for having moderate-to-high MP (≥ 3 on a scale of 0–10) were “treating more patients at the same time” [OR 2.14 (95% CI 1.53–2.99)], “working ≥45 h per week” [OR 1.73 (95% CI 1.05–2.84)], and “work in a seated position” [OR 2.04 (95% CI 1.16–3.57)] for the low back. “More years of experience” showed a negative association for elbow pain [OR 0.41 (95% CI 0.21–0.78)] and low back pain [OR 0.48 (95% CI 0.29–0.79)] compared with their less experienced counterparts.
The lack of professional experience, working in private clinics, working in a seated position and high workload are associated with the higher risk for MP among PTs. These results add further insight about the relevance of such factors, which might be considered for developing effective interventions to prevent work-related MP and better working conditions among PTs.
Work-related musculoskeletal disorders (WRMDs) are defined as a variety of conditions which can affect the musculoskeletal system and occur in relation to work-related activities (Luttmann et al. 2003). These disabling yet in many cases preventable conditions are a common source of musculoskeletal pain (MP) and workplace absenteeism (Luger et al. 2017), and negatively impact the quality of life, which can lead to a decrease in productivity and associated healthcare costs for workers, employers and healthcare professionals (Bhattacharya 2014).
MP is common among the healthcare workforce, where physical therapists (PTs) are especially at high risk. Actually, according to a recent systematic review, lifetime prevalence of MP in PTs ranged between 53 and 91%, with the low back being the most commonly affected body area, followed by the neck, thumbs, upper back and shoulders (Vieira et al. 2016).
Despite the multifactorial and complex nature of pain (Brodal 2017; Ji et al. 2018), these rates seem to be associated to the physically demanding nature of their work tasks and the exposition to multiple factors (Mansfield et al. 2018). Indeed, some of these factors are related to the work environment, such as sustained and awkward postures, bending, carrying, repositioning or lifting patients (Devreux et al. 2012). However, the physical and mental demands of this profession can be variable depending on the setting, the specialty, the working position, or the quantity of work (e.g., number of patients and total working hours).
Accordingly, different authors have reported that those therapists who perform manual techniques and treat a large number of patients per day are more prone to be affected by MP in the thumbs, hands or wrists (Caragianis 2002; McMahon et al. 2006; Power and Fleming 2007), whereas other body areas seem to be more commonly affected in other settings. For example, a previous study showed that those working in hospital settings rather than in private clinics were more likely to develop MP, especially in body areas such as the lower back (Bork et al. 1996), as the level of physical dependence of the hospital patients is usually superior than those who attend to private clinics.
Albeit several epidemiological studies have described some work-related factors such as years of working experience, number of patients per week, or the total working hours as contributors of MP (Adegoke et al. 2008; Rozenfeld et al. 2010), few studies have been conducted with the aim of evaluating which of these factors are associated with higher risk for having MP among PTs. Therefore, a better understanding of these specific work-related risk factors is needed. Such knowledge could be used to highlight work-related risk factors that need further attention and to develop effective interventions adequately targeted, for improving working conditions and preventing musculoskeletal disorders among PTs. This could potentially contribute to a longer and healthier working life of this part of the workforce.
Thus, the aim of the present study was to investigate the association between work-related factors and MP in the back, neck and upper extremities among PTs. We hypothesized that work-related factors such as not having enough professional experience, working in public hospital settings, and treating a higher number of patients per week could increase the odds for MP among PTs.
This cross-sectional study collected data on MP and work-related factors from an online questionnaire sent to PTs in Spain. Potential participants for this study included practicing PTs who were registered in the professional association of PTs of different communities across Spain. Those participants who were retired or were not actively working at the time of the investigation were excluded. The present study was approved by the University of Valencia’s Ethical Committee (H1530736596718), in accordance with the principles of the Helsinki Declaration. To ensure comprehensive reporting of the data of this cross-sectional study, the STROBE guidelines were followed (von Elm et al. 2007). The data collection was conducted from January to June 2017.
The researchers contacted the main professional associations of PTs of different communities in Spain to ask for permission to invite their members to participate on a voluntary basis. The members received the invitation letter along with the project description via e-mail along with a link to the online questionnaire. By responding to the questionnaire each participant was giving consent to participate in the study and permission for the results to be published. The name and contact information of the researchers were included in the cover letter for solving any doubt or concern of the eligible participants before deciding to participate. The online questionnaire took about 20 min to complete. One month following the original e-mail, a reminder was sent to everyone inviting the PTs to participate if they had not done so previously. Due to the recruiting procedure, the exact number of invited participants was unknown.
The questionnaire was designed to collect information about self-reported MP and work-related factors among PTs. Preliminary questions were based on published instruments previously used (Bork et al. 1996; Salik and Özcan 2004; Nordin et al. 2011). To assess the content validity and question clarity of the questionnaire, ten PTs from academic, hospital and private-office settings, reviewed each question and pilot tested the survey. Their feedback was taken into consideration, and some items were reformulated by the researchers to ensure that each question was clear and easy to respond to. Once the questions were reviewed and amended, an online questionnaire was created using the online-tool Google Forms (Google Inc., Mountain View, CA, USA) to collect all responses and storage of the data. Due to data privacy reasons the setting of the survey system was set to “anonymous”, i.e. it was not possible to link the individual responses to neither individual emails nor IP-addresses of the participants.
According to an online tool (https://www.surveymonkey.com) and considering the estimated number of PTs in our country (i.e., 54,258) and in Europe (i.e., 554,000), a sample size of 783 was appropriate to have a confident level of 95% and a margin of error of 3.5%.
Demographic, lifestyle and work-related questions
The first section of the questionnaire consisted of closed-ended questions about participants demographics, lifestyle and work-related information. Participants provided data about their age, gender, height, weight, alcohol consumption, smoking habits, education and leisure physical activity. Work-related questions comprised years of professional experience, working hours per week in the main physical therapy job, number of patients treated per week, if they treated more than one patient at the same time, primary type of patients, primary type of treatment, if they adjusted the examination table when necessary, work position and practice setting of the main physical therapy job.
Musculoskeletal pain questions
The second section included modified questions from the Nordic Musculoskeletal Questionnaire (Kuorinka et al. 1987) to report the prevalence and severity of MP in the upper extremities and the trunk during the last month. Using a simple body diagram highlighted with specific body areas (neck, shoulders, upper back, low back, elbow/forearm and hand/wrist), subjects reported the presence of MP responding the question “Have you had pain or discomfort during the last month in your [body area]?” with options to answer ‘yes’ or ‘no’. When the answer was ‘yes’, they were asked to rate pain intensity using a 0–10 analogue scale, where 0 meant “no pain at all” and 10 was considered “pain is as bad as it could possibly be”. The Nordic Musculoskeletal Questionnaire has been reported to be a valid screening tool (Kuorinka et al. 1987), with sensitivity ranging between 66 and 92%, and specificity between 71 and 88% (Ohlsson et al. 1994).
All statistical analyses were performed using the SAS statistical software for Windows (Proc Logistic, SAS v9.4). Descriptive statistics were used to report the prevalence MP in the upper body, and demographic characteristics (age, height, weight, gender, education, smoking, alcohol units per week and levels of physical activity). Using binary logistic regression, odds ratios (ORs) and 95% confidence intervals (CI) were calculated for having moderate to high MP (≥ 3 on a scale of 0–10, reference category: MP 0–2) in different body areas (dependent variables). The independent variables were work-related factors, such as other works, years of experience, sector, type of employment, working hours per week, number of patients per week, treating more patients at the same time, primary type of patients and treatments, adjusting the examination table when needed and work position as mutually adjusted independent variables.
According to a previous study that compared ORs with effect sizes (Cohen’s d), ORs of 1.68, 3.47 and 6.71 correspond to small, medium and large effect sizes, respectively (Chen et al. 2010). As we analyzed effects rather than associations, we used the terms ‘weak’, ‘moderate’ and ‘strong’ positive associations for ORs of 1.68, 3.47 and 6.71, respectively. For ORs lower than 1, the reciprocal of the OR should be considered, that is, ORs of 0.60, 0.29 and 0.15 correspond to ‘weak’, ‘moderate’ and ‘strong’ negative associations, respectively.
Of the 1006 questionnaires which were returned by registered PTs, 25 questionnaires with missing information on self-reported pain or on work-related factors were excluded from the analysis. Thus, data from the remaining 981 questionnaires were analyzed.
Participant characteristics are described in Table 1. The study population of PTs had a mean age of 34.3 ± 8.0 years, 29.4% were male and 70.6% were female, whom on average had a BMI of 23.3 ± 3.4 kg/m2. The prevalence of MP (≥ 3 on a scale of 0–10) in the upper body areas is also shown in Table 1. Neck pain was the most commonly reported MP (57.0% of the participants reported to have experienced this symptom during the last month), followed by low-back pain (49.4%), upper-back pain (36.1%) and shoulder pain (33.8%).
Table 2 shows ORs for having moderate to high MP (≥ 3 on a scale of 0–10) in upper body areas (neck, shoulders, upper back, low-back, elbow/forearm and hand/wrist) in relation to different work-related factors. Of all the factors of the present study, those which presented higher risk for having low back pain were “treating more patients at the same time”, “working more than 45 h per week” and “work in a seated position”.
With the public sector as a reference, the private sector indicated a positive weak to moderate association with neck and shoulder pain. For the therapists that were working both in the public and private sector, there was a stronger positive association with pain in the shoulders, being still a weak to moderate association.
Working more than 45 h per week showed a positive association with upper back and low back pain, considering less than 35 h per week as reference. These associations were weak to moderate. For the number of patients per week, there was a weak to moderate positive association between “30 and 50 patients per week” and “more than 50 patients per week”, and shoulder pain, in comparison to “less than 30 patients per week”.
For the years of experience, with 0–5 years of work experience as reference, years that ranged between 6 and 15 years, were negatively associated (lower odds) with shoulders, low back and elbow/forearm pain. In PTs with more than 15 years of experience, there were negative weak to moderate associations with the aforementioned body areas, and neck pain, respectively.
PTs who use physical exercise as the primary type of treatment tend to have lower rates of neck pain compared with those who use manual therapy. In contrast, when the main type of treatment is the use of machines, PTs consistently report higher rates of upper back pain.
However, for other work-related factors like the type of employment, adjusting the stretcher when necessary or having more jobs, the association appeared to be less pronounced.
The main findings of the present study suggest that several work-related factors are associated with MP among PTs. Partially supporting our hypothesis, the lack of professional experience was associated with upper limb, and low back pain and working in private clinics showed associations with neck and shoulder pain. Treating more patients at the same time and working in a seated position was associated with low back pain, treating more than 30 patients per week with shoulder pain, and working more than 45 h per week showed associations with both upper and low back pain.
These results are consistent with a recent systematic review which suggested that the high prevalence rates of MP in PTs with fewer years of professional experience could be explained due to the lack of patient management skills and the dearth of practice about how to reduce the risk of developing MP (Vieira et al. 2016). As a matter of fact, one previous study (Nyland and Anne 2003), reported that even undergraduate physiotherapy students have a higher likelihood of developing low back pain during their training, suggesting that new PTs may be entering the workforce with existing low back pain. Other studies suggested that the low prevalence of MP in older therapists might be related with the development of injury-prevention strategies for coping with the physical demands of their jobs, such as modification of treatment techniques or increasing the use of support staff when required (Bork et al. 1996). The healthy workers effect may also be at play, i.e. therapists who do not adopt preventive strategies may leave the profession earlier or change their job, being a possible explanation of the low prevalence rates of MP in this age group (Bork et al. 1996). It could be also plausible that less experienced PTs are less familiarized with the physical demands from their workplace while more experienced PTs developed a higher pain threshold due to a higher work volume.
Contrariwise, the study of Grooten et al. revealed that more than half of the PTs with more than 15 years of experience (53.5%) were affected by MP (Grooten et al. 2011), and similarly, the study of Darragh et al. reported that age equal or greater than 55 years was a risk factor for MP among occupational therapists, finding that the odds of having MP were 3.46 times as high among those being 55 years or older (95% CI = 1.14, 10.49) (Darragh et al. 2009). However, these results should be interpreted carefully because, according to the authors of the study, sample size was small (131 subjects), and only trade union members were invited to participate. Thus, the previous study did not perform compare other subgroups of participants with less experience (Grooten et al. 2011). In the study of Darragh et al. the odds of having MP were 1.49 times higher among occupational therapists than PTs. Thus, it seems that the factors involved in MP among occupational therapists are not equivalent to those affecting to PTs (Darragh et al. 2009). Nevertheless, both professions appeared to be at particularly high risk of developing MP in more experienced individuals. Considering these results, future studies should investigate in more detail the age-related aspects of MP in PTs. It could be speculated that the risk decreases after the first years due to better working routines and practice and that the risk then increases again after many years of exposing the body to physically strenuous working conditions.
The type of treatment seems to play an important role too in the prevalence of MP. Our results showed a positive association for having MP when manual therapy is the primary type of treatment. We found that there was a weak to moderate positive association for having moderate to high pain (≥ 3 on a scale of 0–10) in the hand/wrist and in the neck when comparing with other types of primary treatments such as physical exercise, which showed a negative association with neck pain. As reported previously, procedures such as joint mobilization, manual traction and/or orthopedic manual therapy techniques, were associated with MP in the hand/wrist (Bork et al. 1996; Cromie et al. 2000; Grooten et al. 2011). Indeed, Bork et al. reported that those PTs who habitually performed manual therapy were 3.5 times more prone to have wrist or hand symptoms than those who did not perform such techniques, suggesting that manual therapy techniques could increase mechanical stress on specific anatomical areas, being a major source of upper limb MP (Bork et al. 1996).
The significant association between those who were treating a higher number of patients per week and shoulder pain was not surprising. This could be explained by their primary role in the movement of the upper limbs and, therefore, be more prone to exhaustion after higher workloads. In fact, repetition and monotony have been reported as contributor factors for developing shoulder pain (Buckle and Devereux 2002). Interestingly, negative associations were found for upper back pain and treating more than 50 patients per week, in comparison to those PTs who treated less patients. Upper back muscles have a stabilizing function, so probably this musculature may be better adapted to higher work demands, and consequently, might play some role as a protective mechanism for MP. However, associations were weak, so further studies are needed to corroborate this assumption.
PTs who work more hours per week are also at greater risk for low back pain than those who work less. In line with our results, previous investigations have reported a strong relationship between working more hours per week and risk of injury among health professionals (Trinkoff et al. 2003), and more specifically among PTs (Cromie et al. 2000). Accordingly, a previous study found weak to moderate associations between the number of weekly hours performing rehabilitation treatments and an increased risk of MP in shoulder/elbow, as well as an increased risk in the wrist/thumb for those PTs who work more hours and perform manual treatments (Rozenfeld et al. 2010). However, these results should be interpreted with caution, as different risk factors can coexist in combination with others, and when two or more are present together, it may increase the odds for developing MP, especially when these professionals have excessive workload, prolonged duration of work, insufficient rest periods or monotonous work without task variations (Yassi 1997).
Although previous investigations have reported the association between working in public hospitals and a higher prevalence of MP compared with their non-hospital-based counterparts (Bork et al. 1996; Alrowayeh et al. 2010), the present study found opposite results. PTs who worked in the private sector (i.e., private clinic), compared with those who worked in public hospitals were more likely to report MP, especially in the neck and shoulders. These associations were even more pronounced in those PTs who were working in both public and private sectors. A possible explanation of these findings could be the nature of the physical therapy profession in Spain, as PTs that work in the private sector tend to have longer journeys compared with those who work in public settings, who have a fixed working day length of 7 h. During this time, they have several breaks, which allows them to move and walk. However, in private settings, working time may be variable, including more hours and less breaks, especially when the salary depends on the volume of treatments. In addition, PTs in private clinics usually have a more limited space than in hospitals, having a lower possibility to move (which can also determine the type of treatment used). According to a previous study (Liao et al. 2016), private physiotherapy clinics may have not adequate equipment and less undergraduate students to undertake primary care. In this sense, alternating work “which allows breaks in otherwise repetitive or maintained activities” is essential in the prevention of such musculoskeletal complaints (Cromie et al. 2000), being a possible explanation for the lower rates of MP among PTs working in public hospitals.
Our findings suggest that working mainly in a seated position increases the odds for developing MP, especially in the lower back. These results are in concordance with a previous study among a general working population, which reported that these associations could be produced by a possible relation between prolonged sitting and continuous static load on the musculoskeletal system (Andersen et al. 2007).
Several attempts have been promoted as preventive strategies for decreasing the high prevalence rates of MP among healthcare professionals. However, single strategy ergonomic interventions, such as the implementation of assistive devices and aids has not resulted effective for reducing musculoskeletal complaints in this part of the workforce. Overall, keeping an appropriate level of physical condition is considered the most reported current strategy used by healthcare workers to enable them continue working (McPhail and Waite 2014). In fact, in workers with physically demanding jobs, high-intensity physical activity during leisure time is associated in a dose-response manner with work ability (Calatayud et al. 2015). Therefore, physical training may be an interesting tool for keeping PTs healthy and to enable them to perform their tasks efficiently. By increasing their physical capacity, the relative workload may decrease, reducing MP.
Strengths and limitations
To our knowledge, our study not only provides data about the association between different work-related factors with MP among PTs but also adds further insight about the relevance of such factors for this working group. However, the cross-sectional nature of this study is a limitation because the exposure and outcome were simultaneously assessed. Prospective cohort studies are needed to corroborate the associations between factors that cause MP among PTs. Furthermore, because the exact number of invited participants was unknown, we have not been able to provide the response percentage.
The findings of the present study suggest that several work-related factors are associated with MP among PTs. The lack of professional experience, working in private clinics, treating more patients at the same time, working in a seated position, treating more than 30 patients per week, and working more than 45 h per week were associated with MP among PTs, especially in specific body areas such as the low back, the shoulders or the neck. The results of this study might be considered for developing clinical guidelines and to develop effective interventions to prevent work-related MP and better working conditions among PTs.
Adegoke BO, Akodu AK, Oyeyemi AL (2008) Work-related musculoskeletal disorders among Nigerian Physiotherapists. BMC Musculoskelet Disord 9:112
Alrowayeh HN, Alshatti TA, Aljadi SH et al (2010) Prevalence, characteristics, and impacts of work-related musculoskeletal disorders: a survey among physical therapists in the State of Kuwait. BMC Musculoskelet Disord 11:116
Andersen JH, Haahr JP, Frost P (2007) Risk factors for more severe regional musculoskeletal symptoms: a two-year prospective study of a general working population. Arthritis Rheum 56:1355–1364
Bhattacharya A (2014) Costs of occupational musculoskeletal disorders (MSDs) in the United States. Int J Ind Ergon 3:448–454
Bork BE, Cook TM, Rosecrance JC et al (1996) Work-related musculoskeletal disorders among physical therapists. Phys Ther 76:827–835
Brodal P (2017) A neurobiologist’s attempt to understand persistent pain. Scand J Pain 15:140–147. https://doi.org/10.1016/j.sjpain.2017.03.001
Buckle PW, Devereux JJ (2002) The nature of work-related neck and upper limb musculoskeletal disorders. Appl Ergon 33:207–217
Calatayud J, Jakobsen MD, Sundstrup E et al (2015) Dose-response association between leisure time physical activity and work ability: cross-sectional study among 3000 workers. Scand J Public Health 43:819–824
Caragianis S (2002) The prevalence of occupational injuries among hand therapists in Australia and New Zealand. J Hand Ther Off J Am Soc Hand Ther 15:234–241
Chen H, Cohen P, Chen S (2010) How big is a big odds ratio? Interpreting the magnitudes of odds ratios in epidemiological studies. Commun Stat Simul Comput 39:860–864
Cromie JE, Robertson VJ, Best MO (2000) Work-related musculoskeletal disorders in physical therapists: prevalence, severity, risks, and responses. Phys Ther 80:336–351
Darragh AR, Huddleston W, King P (2009) Work-related musculoskeletal injuries and disorders among occupational and physical therapists. Am J Occup Ther 63:351–362
Devreux IC, Al-Awa B, Mamdouh K, Elsayed E (2012) Relation of work-related musculoskeletal disorders and over-commitment of rehabilitation staff in Saudi Arabia. Life Sci J 9:781
Grooten WJ, Wernstedt P, Campo M (2011) Work-related musculoskeletal disorders in female Swedish physical therapists with more than 15 years of job experience: prevalence and associations with work exposures. Physiother Theory Pract 27:213–222
Ji R-R, Nackley A, Huh Y et al (2018) Neuroinflammation and central sensitization in chronic and widespread pain. Anesthesiol J Am Soc Anesthesiol 129:343–366
Kuorinka I, Jonsson B, Kilbom A et al (1987) Standardised nordic questionnaires for the analysis of musculoskeletal symptoms. Appl Ergon 18:233–237
Liao J-C, Ho C-H, Chiu H-Y et al (2016) Physiotherapists working in clinics have increased risk for new-onset spine disorders: a 12-year population-based study. Medicine (Baltimore) 95:e4405
Luger T, Maher CG, Rieger MA, Steinhilber B (2017) Work‐break schedules for preventing musculoskeletal disorders in workers. The Cochrane library. Wiley, Hoboken
Luttmann A, Jäger M, Griefahn B et al (2003) Work-related musculoskeletal disorders—a definition. Preventing Musculoskeletal Disorders in the Workplace. World Health Organization, Geneva
Mansfield M, Thacker M, Smith T (2018) Physical activity participation and the association with work-related upper quadrant disorders (WRUQDs): a systematic review. Musculoskeletal Care 16:178–187. https://doi.org/10.1002/msc.1204
McMahon M, Stiller K, Trott P (2006) The prevalence of thumb problems in Australian physiotherapists is high: an observational study. Aust J Physiother 52:287–292
McPhail SM, Waite MC (2014) Physical activity and health-related quality of life among physiotherapists: a cross sectional survey in an Australian hospital and health service. J Occup Med Toxicol Lond 9:1
Nordin NAM, Leonard JH, Thye NC (2011) Work-related injuries among physiotherapists in public hospitals-a Southeast Asian picture. Clinics 66:373–378
Nyland LJ, Anne K (2003) Is undergraduate physiotherapy study a risk factor for low back pain? A prevalence study of LBP in physiotherapy students. BMC Musculoskelet Disord 4:22
Ohlsson K, Attewell R, Johnsson B et al (1994) An assessment of neck and upper extremity disorders by questionnaire and clinical examination. Ergonomics 37:891–897
Power H, Fleming H (2007) Work-related thumb pain in manipulative physiotherapists-an Irish survey. Physiother Irel 28:51
Rozenfeld V, Ribak J, Danziger J et al (2010) Prevalence, risk factors and preventive strategies in work-related musculoskeletal disorders among Israeli physical therapists. Physiother Res Int 15:176–184
Salik Y, Özcan A (2004) Work-related musculoskeletal disorders: a survey of physical therapists in Izmir-Turkey. BMC Musculoskelet Disord 5:27. https://doi.org/10.1186/1471-2474-5-27
Trinkoff AM, Lipscomb JA, Geiger-Brown J et al (2003) Perceived physical demands and reported musculoskeletal problems in registered nurses. Am J Prev Med 24:270–275
Vieira ER, Schneider P, Guidera C et al (2016) Work-related musculoskeletal disorders among physical therapists: a systematic review. J Back Musculoskelet Rehabil 29:417–428. https://doi.org/10.3233/BMR-150649
von Elm E, Altman DG, Egger M et al (2007) The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Med 4:e296
Yassi A (1997) Repetitive strain injuries. Lancet Lond Engl 349:943–947
No funding or grant from any commercial source was involved in this study. The authors thank the participants for their contribution to the study
Conflict of interest
The authors of this study declare that they have any conflict of interest
This study received ethical approval by the University of Valencia’s Ethical Committee (H1530736596718), and has been performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Ezzatvar, Y., Calatayud, J., Andersen, L.L. et al. Professional experience, work setting, work posture and workload influence the risk for musculoskeletal pain among physical therapists: a cross-sectional study. Int Arch Occup Environ Health 93, 189–196 (2020). https://doi.org/10.1007/s00420-019-01468-7
- Health care workers
- Physical work