Physical and psychosocial work demands associated with pain and disability among skilled male Arab waiters

2.1 Study design and sample

This was a cross-sectional observational study conducted amongst a convenience sample of skilled male Arab waiters from three restaurants in the Tel Aviv-Jaffa area, in the center of Israel. Data collection was performed between January 2018 and January 2019. We chose to study this population due to its specific characteristics. The Arab waiters were highly skilled professionals who had been working in the industry for many years and had the intention of continuing their careers for many more years to come. Unlike most waiters in Israel, who are students, these waiters have solely focused on their profession and have not pursued any other career paths. Furthermore, the three selected restaurants were of considerable size, which facilitated the collection of data from a larger sample.

Our research included a study population of 100 Arab male waiters, aged 18–60, with at least two years of experience in serving, who have been employed at their current location for > 6 months, had not experienced significant orthopedic, or neurologic accidents nor undergone surgeries (e.g., orthopedic surgeries including the spine, or neurological surgeries that caused weakness or pain) and were employed full-time. The study was approved in October 2018 by the institutional review board (Ethics Committee) of Tel-Aviv University, Tel-Aviv, Israel. Participation in the study was voluntary. Upon being informed of the study’s goals and methodology, each participant signed an informed consent form.

Each of the three restaurants provided full-service options, was open seven days a week, privately owned, and specialized in the same type of Arabic food. Once the customer was seated, a tray of ∼10 small salad dishes was served, and then the main dishes were ordered.

The sample size was estimated according to the assumption that waiters suffer up to 15% more back pain than the general population (73% versus 58%, respectively), and was calculated to 80 subjects. This would give a power of 81.0% at a significance level of 0.05. We considered a dropout of up to 20%, thus, our sample included 100 participants.

2.2 Data collection

At the start of their shifts, each waiter completed a basic demographics and working conditions questionnaire, five physical disability questionnaires, and a three-part psychosocial questionnaire. The basic demographics and working conditions questionnaire included: age, years of education, number of children, years of working, hours worked per week, smoking, and physical activity.

2.2.1 Physical disability questionnaires

1. The Neck Disability Index (NDI) questionnaire consists of 10 items: pain intensity, personal care, lifting, sleeping, driving, recreation, headaches, concentration, reading, and work. Responses range from 0 (no disability response) to 5 (highest disability response). Higher scores represent greater disability. The maximum total score was 50. It was proved that the NDI questionnaire is a valid and reliable instrument [8]. Shashua et al. confirmed the Hebrew version’s validity and reliability [9].

2. The short form of the Disabilities of the Arms, Shoulders and Hands questionnaire (QuickDASH) includes 15 items in six domains: daily activities, symptoms, social function, work function, sleep, and confidence. Responses range from 1 (no difficulty/not at all/not limited/none/strongly disagree) to 5 (unable/extremely/so much difficulty that I can’t sleep/strongly agree). A higher score indicates more disability. The maximum total score was 50. The Swedish version of the QuickDASH has proven to be a valid and reliable instrument [10].

3. The Modified Oswestry Low Back Pain Disability Questionnaire (OswestryQ) consists of 10 items: pain intensity, personal care (washing, dressing, etc.), lifting, walking, sitting, standing, sleeping, sex life (if applicable), social life, and traveling. The 10 items are scored 0 (no activity limitations caused by the pain) to 5 (major activity limitations caused by the pain), summed up out of 50. The OswestryQ has proven to be a valid and reliable instrument in Hebrew [11], and the Israeli Arabic version [12].

4. The adjusted Nordic questionnaire (NordicQ) was used to collect information regarding pain, aches, or discomfort in the spine (neck and lower back), and upper limbs (shoulders, elbows, arms, wrists, and fingers), lasting for a day or longer during the past 12 months. Furthermore, participants were asked if they were unable to either work (e.g., functional impact on work tasks/conditions) at home or travel to work due to pain, aches, or discomfort, during the past 12 months. The questionnaire is repeatable, reliable, and valid [13], and has been found appropriate for use in an Israeli population [14].

5. The questionnaire used in this study is based on the premise that a large group of people’s collective knowledge can lead to an optimal and meaningful solution, the Delphi technique [15]. To develop the questionnaire, eight full-time waiters with more than two years of experience were contacted and divided into four pairs. After a thorough explanation of task demands, each pair generated a list of the ten most hazardous tasks they encountered in their work environment. A total of 22 tasks were identified, and each pair chose ten tasks from the list. Ten tasks that were rated by at least two pairs were ultimately selected by mutual agreement. The tasks include loading plates in both hands, standing for long periods, dealing with emotional overload (i.e., customer dissatisfaction with the service, hostile behavior of customers, etc.), serving dishes at large tables, removing plates and glasses while the diners were still seated, dealing with cognitive problems at work (i.e., remembering customer orders, customer requirements, etc.), working under pressure during a busy shift, serving dishes at crowded tables, working in the non-air-conditioned area and lifting more than 5 kg per serving. The 10 items are scored 1 (no activity at all) to 6 (regular activity), summed up out of 60.

To assess the questionnaire’s repeatability, it was evaluated twice by 20 participants using the test-retest method, which yielded a correlation result of ICC = 0.865. The items included in the questionnaire can be seen in Fig. 1.

OPEN IN VIEWER

Fig. 1

Frequency of the ten different work-related tasks/challenges included in the TaskQ in %.

2.3 Statistical analyses

All statistical computations were performed using SPSS 23.0 for Windows [20]. Statistical analyses were conducted at a 95% confidence level. A p-value of p < 0.05 was considered significant. The bivariate (Spearman’s) correlation analysis evaluated the test-re-test reliability of the TaskQ items. The internal consistency of the psychosocial and TaskQ questionnaires was evaluated by Cronbach’s alpha. Means, standard deviations, and frequencies characterized the study sample including background and working conditions, ergonomic, psychosocial factors, and WRMSP. Pain frequency was converted to dichotomous variables. Answers of “one-time event”, or “seldom” were considered as low frequency, and answers of “sometimes”, “often”, or “always”, were considered as high frequency. The frequencies of tasks (TaskQ) were also converted to dichotomous variables. Answers of “never”, “seldom”, or “sometimes” were considered as low frequency. Answers of “often”, “very often”, or “regularly” were considered as high frequency. Associations between various personal characteristics, working conditions, and indicators of WRMSP were measured using bivariate analysis. Linear regression analyzed pain and disability as dependent variables (OswestryQ, QuickDASH, NDI) and psychosocial factors (burnout, job satisfaction and workload scores) as predicting variables, adjusted for physical activity, number of children, age, and number of hours worked per week. Logistic regression analyzed low back, neck, hands, and arm pain that had occurred during the last 12 months as dependent variables (the adjusted NordicQ), adjusted for physical activity, number of children, age, and number of hours worked per week.

3.1 Background and working conditions characteristics of the studied sample, task demands, and psychosocial factors

Table 1 illustrates the background and working conditions of the studied sample. One hundred male Arab waiters, possessing high-level skills who worked daily, were included in the study. Means and standard deviations (Table 1) were: age 33.91±7.28, years of education 10.1±2.52, number of children 1.68±1.76, working years 6.05±4.58, weekly hours worked 34.64±9.38. Almost half (49%) were smokers and 51% performed some type of physical activity.

Means and standard deviation for the task demands were: TaskQ 29.12 (out of 60)±5.73 (Cronbach’s alpha = 0.765), mean weight of one carried dish per hour was 1.78±0.45 kg, and mean steps walked per hour 701.75±136.96. On average, the means and standard deviations for the questionnaires’ scores were: OswestryQ –14.85±3.10 (out of 50), QuickDASH- 25.54±7.18 (out of 75), and NDI –13.74±8.81 (out of 50). Means and standard deviations for psychosocial factors were: workload questionnaire-31.16±6.11 (out of 36), burnout scores –56.65±20.87 (out of 112), and job satisfaction- 55.26±13.25 (out of 70) with a Cronbach’s alpha of 0.875, 0.971 and 0.946, respectively.

The 12-month pain prevalence in at least one site was 83%. Table 2 illustrates the 12-month pain prevalence, pain frequency, and work-avoidance effect (12-month work-avoidance prevalence due to WRMSP) by body area. 42% reported 12-month neck pain prevalence, 53% -shoulders, 46% -elbows, 37% -arms, 36% -wrists, 17% -fingers, and 45% -low back. The prevalence of high-frequency pain was between 44% (lower back) and 89.5% (arms). The 12-month work-avoidance effect due to WRMSP was between 55.3% (arms) and 94.4% (shoulders).

Table 3 presents the results of the linear regression analysis for pain and disability (QuickDASH, NDI, OswestryQ) as dependent variables, and psychosocial variables (burnout, job satisfaction, and work-load) and task demands (TaskQ) as independent predictors, adjusted for physical activity, number of children, age and number of hours worked per week. Only significant results are presented in Table 3. Out of the psychosocial factors, only burnout score was associated with pain and disability. As the burnout score was higher, waiters experienced more back pain and disability (OswestryQ) (β= 0.52, P < 0.001), more pain and disability in the upper extremities (QuickDASH) (β= 0.29, P = 0.004), and more pain and disability in the neck (NDI)f (β= 0.55, p < 0.001). The proportions of the variances that explained the models were 29%, 11% and 34%, respectively. Furthermore, low back pain (LBP) (OswestryQ) was positively associated with task demands (TaskQ) (P = 0.047). The proportion of the variance explained by the model was 10.5%.

3.4 Results of logistic regression analysis (final model): WRMSP (the adjusted NordicQ) low back, neck, hands and arms, shoulders and elbows as dependent variables and psychosocial factors scores as predicting variables

Table 4 presents the results of the logistic regression analysis for WRMSP (the adjusted NordicQ) for the previous year as a dependent variable and psychosocial factors as predictors, adjusted for physical activity, number of children, age, and number of hours worked per week. Only significant results are presented. Job satisfaction was negatively associated with LBP (OR [95% CI]: 1.12 [1.02–1.24]), hands (OR [95% CI]: 0.96 [0.92–0.99]), arms (OR [95% CI]: 0.97 [0.93–1.00]), shoulders (OR [95% CI]: 0.94 [0.90–0.98]) and elbows (OR [95% CI]: 0.96 [0.93–0.99]). The proportion of the variance explained by the models was 37.7%, 10.3%, 4.49%, 0.17%, and 0.13%, respectively. No significant association was found between neck pain and psychosocial factors.

It should be noted that logistic regression analysis was performed with WRMSP as the dependent variable and TaskQ as the predictor variable. However, the analysis did not yield a statistically significant result.

Figure 1 presents the results of the TaskQ, exhibiting the implementation of high-frequency task demands. The most frequent task demand was loading plates in both hands (67%), followed by standing for long periods (52%), dealing with emotional overload (46%), serving dishes at large tables (44%), removing plates and glasses while the diners were still seated (42%), dealing with cognitive problems at work (42%), working under pressure during a busy shift (34%), serving dishes at crowded tables (28%), working in the non-air-conditioned area (22%), and lifting more than 5 kg per serving (0%).

4.1 Prevalence of WRMSP in waiters

The 12-month pain prevalence in the current study was high for several body parts among the participants. This is an important finding as it highlights the prevalence of musculoskeletal pain among skilled male Arab waiters.

Pain prevalence in the upper extremities appears high, according to the adjusted NordicQ, especially in the neck, shoulders, elbows, arms, and wrists (42%, 53%, 46%, 37%, and 36%, respectively). Regarding shoulder pain prevalence during the past year, our study found it to be 53%, which is comparable to a Taiwanese research study conducted in 2004 among 905 restaurant workers, where the adjusted NordicQ was used to gather information regarding body site-specific WRMSP, pain intensity, and strategies for pain relief amongst waiters [23]. The study found that 84% of participants reported WRMSP during the previous month, with the highest prevalence rate found for the shoulders (58%).

Neck pain prevalence was 42%, which is comparable to a cross-sectional observational study of 20 waiters who described a 45% neck pain prevalence by observing and measuring ergonomic stressors at three different restaurants during an entire shift [3]. However, the magnitude of neck pain was lower among urban bus drivers in Israel, with only 21.2% in the 12 months [24]. It is not surprising that waiter’s service work, which is mainly manual, puts loads mostly on the upper extremities, in addition to loading the entire torso on the low back when bending.

The current study indicated a 45% prevalence of LBP during the past year, which is higher than a cross-sectional study of 100 wait staff from ten casual dining restaurants located in the eastern USA, where the 12-month pain prevalence most frequently occurred in the low back area (18%) [25]. Moreover, only 42% of the participants in the US study reported musculoskeletal symptoms during the past year, whereas we reported a one-year WRMSP prevalence of 83%.

Chefs in the food industry also reported pain in the past 12 months in the lower back (32.9%) and upper back (30.3%), most probably as a result of fatigue after long work hours [26]. In a recent study by Yalew et al. [27], 43.8% of participants reported LBP at some point in the past 12 months, with female participants having a higher prevalence of 70.6%.

It is important to note that women often report more pain compared to men describing the same task, but our male waiters reported higher pain and disability compared to the nurses and nursing assistants in an intervention study conducted in Israel, where the effects of sliding sheet usage on WRMSP and perceived psychosocial variables were examined among nurses and nursing assistants [28]. The scores obtained for pain and disability were at the highest (before the intervention), 8.61.

4.2 Results of the pain and disability questionnaires

We not only assessed the frequency of pain but also evaluated the combination of pain and increased disability in different body parts using validated questionnaires. Specifically, we used the OswestryQ for the low back, QuickDASH for the upper extremities (including hands, shoulders, and arms), and NDI for the neck. The pain and disability scores we obtained were 29.7/50, 25.54/75, and 13.74/50, respectively.

In another study conducted in Israel [28], the effects of sliding sheet usage on WRMSP and perceived psychosocial variables were examined among nurses and nursing assistants. The pain and disability scores obtained in that study were significantly lower than those in our present study, with the highest scores (before the intervention) being 8.61 for OswestryQ, 19.12 for QuickDASH, and 9.80 for NDI.

It is worth noting that there are sex differences in the prevalence and degree of disability [29]. Women often report more pain than men describing the same task, probably due to reduced tolerance and increased responses to painful stimuli. However, our male waiters reported higher pain and disability than the nurses and nursing assistants in the intervention study mentioned above [28].

4.3 Work avoidance

Due to WRMSP, our wait staff avoided tasks with high frequencies such as the neck (89.5%), shoulders (94.4%), and fingers (84.6%). A comparison between the 12-month work avoidance prevalence due to WRMSP in physical therapists and nurses in Israel [30], using the adjusted NordicQ, showed that both physical therapists and nurses reported much less work avoidance compared to the waiters in our study.

Specifically, physical therapists, nurses, and our waiters demonstrated a 7.7%, 15.8%, and 89.5% work avoidance for the neck, 7.7%, 14%, and 94.4% for the shoulders, 3.8%, 3.5%, and 83.3% for the elbows, and 34.6%, 24.6%, and 78.7% for the lower back, respectively. It is worth noting that both nurses and physical therapists, who are well-recognized in the literature as suffering from WRMSP as a result of their physical and psychosocial work environment, reported much less work avoidance compared to our male waiters.

Moreover, among our waiters, the pain and disability questionnaires (QuickDash, OswestryQ, and NDI) had moderate scores compared to their high work avoidance scores. After all, the pain and disability questionnaires reflect the reality of pain disability in waiters. It is possible that our professional waiters are trying to prevent more severe disability in the future by taking temporary rest and avoiding pain. These waiters, whose working conditions are difficult, must keep their jobs as the main and even exclusive breadwinners in the family.

It is important to note that there is a high prevalence of poor mental health among waiters [31], making them unable to work effectively resulting in high turnover intention. In our study, we did not examine the mental health of the waiters, however, they seem to try to prevent themselves from pain and illness by staying home at their own pace.

4.4 TaskQ

The TaskQ using the Delphi technique [15] was specially developed for our study. It included tasks such as carrying plates with both hands, standing for long periods, dealing with emotional stressors (e.g., dissatisfied customers, hostile behavior), serving large tables, clearing plates while diners are still seated, coping with cognitive challenges (e.g., remembering orders), working under pressure during busy shifts, serving crowded tables, working in non-air-conditioned areas, and lifting over 5 kg per serving. The questionnaire revealed that these tasks comprise approximately 50% of the waiters’ job demands. Furthermore, the TaskQ scores were significantly associated with LBP and disability (OswestryQ), indicating that demanding job tasks may exacerbate LBP.

Although we found no association between TaskQ and WRMSP, using the adjusted NordicQ, we did find an association between TaskQ and LBP and disability (OswestryQ). We believe that multi-question questionnaires, such as the OswestryQ, are more precise and sensitive for assessing pain with its disabilities.

4.5 Psychosocial factors

In this study, we investigated three psychosocial factors among wait staff: job satisfaction, workload, and burnout. Our findings revealed a positive association between our burnout scores and WRMSP, which is consistent with previous studies demonstrating the professional impact of burnout [23, 33]. It is possible that the fast-paced and monotonous routine work of waiters may contribute to burnout scores. In contrast, we found a negative association between job satisfaction and WRMSP. Job satisfaction may act as a moderating factor in reducing WRMSP, or pain may decrease due to job satisfaction.

Few studies have investigated the relationship between work-related psychosocial factors and WRMSP in wait staff. Petree et al. [34] found that stressors such as work overload and busy shifts negatively impact workers of all ages. Our study is the first to investigate the association between burnout score, pain, and disability in waiters. The burnout variable in wait staff ergonomic studies should be further investigated to identify and address workforce problems that could impact the standard of services, emotional labor, and burnout.

Overall, the study’s findings highlight the need for interventions that address both physical and psychosocial risk factors to reduce the risk of WRMSP among waiters. The scarcity of literature on waiters’ physical and psychosocial risk factors underscores the importance of further research in this area. We recommend adapting the work environment to the worker, to minimize carrying heavy loads and awkward postures, improving work-life balance, and job control.

4.6 Limitations

Our study consisted of a convenience sample rather than a representative one of 100 participants. The strategies used in the current research uncover all the work hazards which might cause WRMSP and related disabilities in the waiter population. Furthermore, when employing questionnaires, some doubt exists as to the participant’s memory which may lead to recall bias. Lastly, the study population encompassed only Arab men waiters, hence, the results may not necessarily represent the general population.