Study population

Cross-sectional data collected from the baseline survey of the Yoseikai Study, an occupational cohort study conducted in Japan, were used in the current study.

The research ethics committee for the Graduate School of Medicine and Faculty of Medicine at the University of Tokyo (No. 1264) and the Incorporated Medical Institution Yoseikai reviewed and approved the study’s aim and procedure. Written informed consent was obtained from all participants prior to the initiation of the study.

In March 2015, all of the employees at a Japanese hospital (N = 280) were recruited via an invitation letter from the authors. The survey was conducted during May 2015. During the survey period, occupational health staff at the hospital distributed a nonanonymous, self-administered questionnaire to each employee. Once the employees had completed the questionnaires, they placed them in sealed envelopes, which occupational health staff collected and forwarded to the authors. All employees were assured that their participation was voluntary, and supervisors and occupational health staff were not authorized to open the sealed envelopes. In total, 203 employees completed the self-administered questionnaire.

Study measures

The questionnaire included questions regarding the following: individual characteristics including sex, age, body mass index (BMI), and occupation type; LBP severity; and individual and work-related psychosocial factors. LBP severity was evaluated by the respondents, who were asked to indicate the severity of their LBP according to four grades (0 = no LBP, 1 = LBP that did not interfere with work, 2 = LBP that interfered with work, and 3 = LBP that interfered with work and required sick leave). The grades were determined with reference to Von Korff’s grading method [25]. LBP was defined as pain in the lower back lasting for more than 1 day and experienced during the preceding 4 weeks, according to the standard definition of LBP proposed by Dionne et al. [26]. Pain associated with menstruation or pregnancy or experienced during a feverish illness was excluded. A diagram showing the lower back area (between the inferior costal margin and gluteal folds [11]) was included in the questionnaire. LBP with disability was defined as LBP that interfered with work, regardless of work attendance (Grade 2 or 3), because presenteeism, or working while unwell, can lead to productivity loss and poor health.

The questions concerning current occupation type pertained to job satisfaction, job demand, job control, interpersonal stress at work, and social support. Work-related stress was assessed using the Brief Job Stress Questionnaire [27,28], which was developed by a research working group established by the Japan Labour, Health and Welfare Organization. The scale contains 57 items measuring psychosocial work environments, stress reactions, and buffering factors, with responses provided using a four-point Likert scale ranging from 1 to 4, with reverse scoring applied to some items. Total scores range from 57 to 228, and higher scores indicate greater work-related stress. The five original responses were reclassified as “not stressed,” which included low, slightly low, and moderate stress, and “stressed,” which included slightly high and high stress [20,23].

We evaluated mental health problems using the Kessler Screening Scale for Psychological Distress (K6), which was developed in 2002 as a short-form version of the K10 [29]. The scale measures psychological distress experienced during the preceding 30 days, using six items, with responses provided using a five–point scale ranging from 0 (all of the time) to 4 (none of the time) or ranging from 1 (all of the time) to 5 (none of the time). The Japanese version of the scale was developed by Furukawa et al. in 2008 and demonstrated reliability and validity [30]. Respondents were classified into three groups according to their total scores (≥10 = high, 5–9 = moderate, and ≤4 = low).

Somatic symptom burden was measured using the Somatic Symptoms Scale-8 (SSS-8), which is an abbreviated eight-item version of the Patient Health Questionnaire-15. We used the linguistically validated Japanese version of the SSS-8, which was developed in our previous study [31]. The scale measures the extent to which respondents have been bothered by somatic symptoms during the preceding 7 days, with responses provided using a five-point Likert scale ranging from 0 (not at all) to 4 (very much). Total scores range from 0 to 32 and represent somatic symptom severity, with ≥16 points indicating very severe symptoms [32].

Participants’ beliefs and fears were measured using the Fear-Avoidance Beliefs Questionnaire (FABQ), which was developed by Waddell et al. [33] and consists of 16 self-reported items. We used the Japanese version of the FABQ, which was developed and validated recently by Matsudaira et al. [34]. The study used the FABQ’s four-item physical activity subscale (FABQ-PA), which measures respondents’ beliefs about the effects of physical activity on their LBP. Responses are provided using a seven-point Likert scale ranging from 0 (completely disagree) to 6 (completely agree). Total scores range from 0 to 24, and higher scores represent higher FAB levels. Participants’ scores were classified into two categories (≤14 = low, ≥15 = high) [35].

Workaholism, which has been associated with psychological health, was measured using the Dutch Workaholism Scale [36], which consists of two subscales: working excessively and working compulsively. Each subscale consists of five items, with responses provided using a four-point Likert scale ranging from 1 (totally disagree) to 4 (totally agree). Respondents were classified into three groups according to their total scores (high, moderate, and low) [21].

Statistical analysis

We performed logistic regression analysis, as our dependent variable “presence or absence of chronic pain” was dichotomous. One set of guidelines suggested that accurate estimation of discriminant function parameters requires a sample size of at least 20 for each independent variable in logistic regression [37]. In addition, the prevalence rates for chronic pain reportedly range from 10% to 55% [38]. Therefore, we calculated an overall sample size of 200 to ensure that there were 20 participants, even with a minimum prevalence rate of 10% for chronic pain.

Demographic and clinical characteristics were compared using Student's t test for continuous variables and chi-square tests for categorical variables. Factors associated with LBP were assessed using multivariable logistic regression analysis. Risk factors included job satisfaction; job demand; job stress; job control; social support from a supervisor/manager, colleagues, or family; K6 score; SSS-8 score; workaholism score; and FABQ-PA score. Because of the relatively low number of participants with back pain in the study, propensity score adjustment was used for each of the risk factors in multivariate modeling.

Propensity score adjustment preserved statistical power by reducing covariates into a single variable. For example, when the adjusted effect of LBP was evaluated, a propensity score was created using binary logistic regression to predict the probability of LBP as a function of the important factors (sex, age, BMI, occupation type [medical or nonmedical]) included in the study. Data analysis was performed using SAS software (version 9.4, SAS Institute Inc., Cary, NC).

Relationship between FAB and LBP with disability

The fear-avoidance model has been proposed as a representative type of thought process involving the chronicity of LBP. FABs in LBP are negative beliefs and anxiety regarding LBP, which lead to a catastrophizing response in which the worst possible outcome is imagined, causing fear and avoidance of activity and resulting in functional restriction. Of the psychological factors examined as prognostic factors for the development of chronic LBP, FABs have been shown to constitute an important factor and exerted a strong effect on employment conditions and disability prognoses [39]. The introduction of a psychosocial flag system to manage musculoskeletal problems, including LBP, in healthcare and the workplace has been suggested in Western countries. Within this concept, FABs are represented by a yellow flag, and some researchers have recommended that clinical practitioners should judge the involvement of FABs during the early stages of pain and manage them in cooperation with the patient’s workplace [40]. Some studies have shown that FABs led to a tendency toward development of chronic LBP with disability [41–43]. In the present study, the results indicated that FABs were related to LBP in Japanese public health service workers. As a significant independent factor in multiple logistic analysis with certain variables controlled for, FABs were an important factor and should be considered in the management of LBP. In addition, early intervention is important in increasing awareness of FABs in patients with disabling LBP in Japan.

Relationship between somatic symptoms and LBP with disability

The tendency toward somatization is defined as a predisposition to excessive worry regarding common somatic symptoms, such as headaches, dizziness, and stomach or bowel problems, which could be triggered by mental distress. This tendency has been shown to affect various health conditions (via related behaviors), musculoskeletal pain (particularly multisite pain) [44], and absence from work [45]. The relationship between pain and the tendency toward somatization has been observed in both longitudinal and cross-sectional studies, indicating that this tendency is a predictor, rather than a consequence, of other aspects of health [46].

The tendency toward somatization has been associated with [15,46–48] and identified as a major risk factor for LBP[24]. The association between chronic LBP with disability and a combination of psychosocial factors could be explained by dysfunction in the mesolimbic dopamine system, which controls both pain and pleasure [49,50] When a person experiences painful stimuli, the mesolimbic dopamine system is activated to inhibit pain. However, exposure to chronic, rather than acute, stress, such as anxiety or distress, has recently been suggested to result in hyperalgesia because of the inhibition of mesolimbic dopamine mechanisms. For example, hyperalgesia resulting from chronic stress because of discontentment with life and work could lead to the development of chronic LBP with disability [51]. The results of the current study indicated that the tendency toward somatization was associated with LBP that interfered with work. This could indicate LBP should be managed as brain dysfunction as well as musculoskeletal disease.

Relationship between interpersonal stress at work and LBP with disability

The relationship between interpersonal stress at work and occupational LBP has attracted attention for some time [18]. In particular, stressful, monotonous work was identified as a predictive factor for new-onset LBP in a cohort study [52]. In addition, a recent 2-year prospective epidemiological study involving 5,310 workers in Japan suggested that work-related stress affected the onset and persistence of LBP with disability [20]. The way in which psychological factors cause LBP remained unclear in the current study; however, two biomechanical (ergonomic) studies showed that psychological stress increased low back compression force during lifting tasks [53,54], which could indicate that stress is associated with increased risk of LBP development. Our results suggested that interpersonal stress at work is an important factor in understanding occupational LBP.

The results showed that fear-avoidance beliefs regarding LBP, which has been recognized worldwide as an important risk factor; the tendency toward somatization, which is a type of stress response; and interpersonal stress at work were associated with LBP with disability in medical workers. According to the results, pain education based on the biopsychosocial model could be inadequate in Japan, even for medical personnel. A recent systematic review examining LBP prevention indicated that exercise combined with education was likely to reduce the risk of LBP [55]. Ergonomic factors and the psychosocial factors examined in the current study should be considered in education regarding LBP prevention.

The present study was subject to some limitations. First, it was conducted at a single hospital; therefore, the generalization of the results is limited. Second, the number of participants with LBP that interfered with their work was low; however, we dealt with this problem statistically using propensity score adjustment. Third, the study was cross-sectional in design; therefore, causation could not be inferred. We plan to conduct an additional cohort study to examine causal associations. Fourth, various chronic pain conditions interfere with work, but the current study considered only LBP. Future research should examine the effects of chronic pain in various parts of the body and compare them to those observed for LBP. Fifth, the results showed that LBP was affected by individuals’ personal relationships; however, the study did not consider factors examined in previous studies (e.g., family environment, nursing, and genetic predisposition).

In conclusion, the results of the present study suggested that psychosocial factors, such as FABs, the tendency toward somatization, and interpersonal stress at work were associated with LBP that interfered with work. Future preventive strategies for reducing LBP in the workplace should include not only biomechanical factors, which are already well understood, but also the management of psychosocial factors.