No cross-cultural adaptation of the ORFQ existed at the time that this study was conducted. Studies in low and middle income countries have either used the original measure [22, 64], or developed new measures based on the content of the ORFQ [65]. The difficulty in its cultural adaptation for rural Nigeria was related to the use of technical words such as ‘percentage’, ‘degrees’ and ‘pounds’ which did not have Igbo equivalents. Face and content validity could not be established using the items as they were in the original ORFQ because people with low literacy found it challenging to conceptualise the items. Content validity is an estimate of the validity of an outcome tool based on a detailed examination of the contents of the test items by subject experts and people with lived relevant experience; whereas face validity is an estimate by an experienced panel reviewing the content of an assessment or tool to see if it seems appropriate and relevant to the concept it purports to measure [66]. Face and content validity of the Igbo-ORFQ was hence established by using concrete numbers such as the number of workdays per week in place of percentage of exposure, and the weights of objects commonly used in this environment in place of pounds and kilograms; whilst retaining the original items for use amongst literate people in this population. Face and content validity of the Igbo-ORFQ was further enhanced by the demonstration of the activities in the items during the interviewer-administration of the questionnaire by the CHWs. These enhanced understanding and acceptability while retaining conceptual equivalence.
The Igbo-ORFQ demonstrated good reliability with an internal consistency of 0.84 and intraclass correlation coefficient of 0.83. Unweighted and linear weighted kappa showed agreement for most Igbo-ORFQ items. In contrast to the original ORFQ, items 19 (operating powered hand tools) and 24 (lifting 10–30 pounds’ objects), did not show agreement, suggesting that these activities were not consistently performed in this population. Similar median test and retest scores for the total scoring of Igbo-ORFQ suggest the stability of the occupational construct and the Igbo-ORFQ in this population [67, 68].
The findings from the construct validity investigation were unexpected and contradicted the a priori proposed hypotheses. The results suggested that there were no major differences between the threshold and incremental total scoring methods, particularly in the direction of relationships. The strength of associations appeared to be stronger with the threshold total scoring for some outcomes and stronger with the incremental total scoring method for some other outcomes. Notably, the results suggested that the greater the exposure to occupational biomechanical risk factors, the less the work-related disability, functional disability, performance-based mobility related disability, and pain intensity. In contrast, that the higher the exposure to occupational biomechanical risk factors, the more the social support.
Many factors could be considered in explaining these unexpected results. The positive associations between exposure to occupational biomechanical risk factors and all types of disability including functional and work-related disability could be related to the healthy worker effect. The rural Nigerian population that validated the Igbo-ORFQ were either farmers or involved in other informal manual jobs. This might have implied that higher exposure to occupational biomechanical factors were reflecting those in work, and lower exposures were reflecting those who were no longer working as there were limited non-manual jobs in this population which explain the lower work-related disability. The construct of functional disability captured by the three disability outcome measures (Igbo-WHODAS, Igbo-RMDQ and BPS) reflects mobility related activities that are also commonly performed in informal manual jobs, aligning with qualitative research evidence from this population [18, 19]. The positive association between the Igbo-ORFQ and pain intensity could be because pain intensity is a predictor of disability in this population [23]. In view of these findings, disability and pain intensity outcome measures may not be the best measures to validate the Igbo-ORFQ in this population. Instead, outcome tools that measure current pain episode(s) without acknowledging pain intensity or duration of pain in the construct definition may be more useful in validating the Igbo-ORFQ in this population as suggested by recent evidence [26]. This could be because people with greater pain intensity or longer pain duration may have left their manually driven jobs, again obscuring the relationship between exposure to occupational biomechanical risk factors and LBP related disability and pain intensity. This aligns with prospective research evidence in high income countries suggesting that the longer people stayed in physically demanding jobs, the less likely it was for them to develop LBP because people who developed LBP had left the job earlier [12]. This may be associated with the finding that exposure to biomechanical risk factors is more likely to predict new onset LBP in newly employed workers [7]. This finding could also be related to the fact that psychosocial factors may be more important than biomechanical factors in predicting chronic LBP disability [23] and other spinal pain intensity outcomes such as neck pain intensity [25]. Moreover, recent evidence suggests that exposure to biomechanical risk factors may be better measured in this population by using simplified questionnaires that combine objective and subjective procedures such as objective assessment of weight carried, and the frequency and duration that it is carried as well as description of posture predominantly adopted, and the frequency and duration that it is adopted without referring to specific occupational activities which may not be applicable in this population [25].
The unexpected finding that the higher the exposure to occupational biomechanical risk factors, the more the social support could be reflecting the informal self-employed nature of most of the jobs undertaken by the people in this population. These manual jobs which included subsistence farming and family businesses often involved family members undertaking some manual activities entailed in the job as was suggested by previous qualitative research evidence in this population [18, 19]. Therefore, jobs requiring more manual activities were more likely to involve family members and friends to help with these activities explaining the positive association between exposure to occupational biomechanical risk factors and social support. This concurs with the Lancet LBP series suggesting that disability from LBP is greatest in the working population in low- and middle-income countries because the predominant informal employment in those countries preclude the feasibility of occupational modification [69–71].
Another possible but unlikely reason for these unexpected associations could be the difficulty in administering the ORFQ to people with low literacy, as well as reference to specific occupational activities which are not common in this population such as ‘operating powered hand tools’, ‘driving or riding motor vehicles’, ‘working on elevated surfaces, e.g., scaffold’. However, attempts were made to ameliorate this during the cross-cultural adaptation processes by concretising items in the questionnaire as previously described.
However, the Igbo-ORFQ may not be clinically useful because it was originally developed as an epidemiological outcome measure. As the test-retest median scores from this study suggest, the Igbo-ORFQ is stable and may not be helpful in determining the impact of interventions by detecting changes in exposure arising from clinical, rehabilitation or public health interventions. This is more so for changes that arise from individual personal changes rather than job modifications which may be limited in this lower middle-income country with minimal enforcement of occupational health regulations. Calls have been made to develop and deliver pragmatic interventions that target factors associated with LBP outcomes in low and middle income countries [71]. Therefore, a simple outcome tool that measures exposure to biomechanical risk factors within occupational and non-occupational settings and which acknowledge the findings from this study may be needed for this population.
The Igbo-ORFQ (Appendix 1) is the complete questionnaire including the first five (1–5) items that measure work organisational and psychosocial factors, and whose psychometrical soundness have been demonstrated elsewhere [25, 26]. This will enable future studies to improve the understanding of the importance of occupational biomechanical factors relative to occupational psychosocial factors for specific LBP outcomes and overcome the limitations of previous studies.
Considerable evidence from previous research supports the relevance of work-related psychosocial factors in the onset of LBP and the perpetuation of symptoms in high income countries. The strength of association between occupational psychosocial factors and first onset LBP is reported to be modest and smaller than in the chronic stages of LBP [12, 72]. Systematic reviews have found that negative affect, high psychological demands and emotional effort at work, stress, poor social support and work relations, low level of job control, work pace, monotonous tasks, perceived ability to work, high work dissatisfaction, and the belief that work is dangerous were associated with first onset and future episodes of LBP [6, 73, 74]. However, the inclusion of studies that only reported significant associations between psychosocial factors and first onset LBP, limits the establishment of the importance of psychosocial factors in early onset LBP [6] in these previous studies. Not concurrently accounting for the effects of physical factors does not expose the importance of psychosocial factors relative to physical factors in early onset LBP and other LBP outcomes [74, 75]. The use of non-validated psychological measures, and the lack of distinction between discrete LBP outcomes such as pain outcomes, sick leave, medical consultation, treatment and disability in individual studies [76–79], limit definite causal inferences in systematic reviews [9, 73, 74]. The inclusion of prospective cohort studies that only compared exposed with non-exposed cases, and the exclusion of studies of low methodological quality in another systematic review showed no evidence that job related psychosocial factors were causes or consequences of LBP [80]. Three systematic reviews suggest that job satisfaction and work organisational factors do not predict return to work/sick leave [9, 24, 81, 82]. Social support at work was not found to predict the transition of LBP to chronic LBP [81], but social dysfunction/isolation predicted duration of sick leave [9]. Recovery expectation regarding return to work is predictive of positive work outcomes [24, 82], but stress, anxiety and type of occupation (blue collar versus white collar) were not associated with work outcomes [9, 82]. The number and extent of working hours did not influence duration of sick leave [9, 24].
Previous studies have either focused on the role of psychosocial or biomechanical factors rather than on both factors, on a few LBP outcomes, particularly the transition to chronic LBP and the associated disability. Therefore, the adapted Igbo-ORFQ can support the investigation of the importance of both factors on several more relevant LBP outcomes in this population including pain episodes based on the findings from this study.
This study is limited by the lack of bilingual investigation of item-by-item agreement between the original ORFQ and the Igbo-ORFQ because of the limited literacy of the participants. Furthermore, interviewer-administration of the measure by several CHWs, again due to limited literacy, could have introduced random or systematic bias in the measurements. Future studies involving the Igbo-ORFQ could be designed addressing these limitations.