Epidemiology of health care utilisation for low back pain : a population-based observational study in Ethiopia

Introduction Low back pain (LBP) is a global public health problem. It is a highly prevalent and significant source of negative social, psychological, and economic burden. In Ethiopia, LBP ranked in the top ten causes of age standardised disability-adjusted life years (DALYs) in 2015. From 1990 to 2015, while DALYs caused by all other top 30 contributors (such as measles, malaria, and protein energy malnutrition) were shown to decrease, DALYs caused by LBP and sense organ diseases continued to increase (Misganaw et al., 2017b). This shows that combined with neglected tropical diseases, HIV/AIDS, tuberculosis, malaria, and anaemia, which are the common causes of DALYs in sub-Saharan Africa (Vos et al., 2013), LBP may pose a serious burden in Ethiopia. Primary prevention strategies have limited potential while timely and appropriate diagnosis and tailored treatment plans can reduce the burden of pain and improve patient outcomes. A better understanding of the epidemiology of health care utilisation for LBP in resource-limited communities like Ethiopia is therefore significantly important for future health care pathways development (Lentz et al., 2018). Objectives The aims of this study were 1) to develop and validate a measurement instrument used to measure determinants of health care utilisation for LBP, and 2) to investigate the epidemiology of health care utilisation for low back pain in Ethiopia. Methods To develop and validate the measurement instrument, a comprehensive review of the literature was undertaken and the relevant domains of potential determinants of health care utilisation for LBP were identified. Items relating to each domain were then generated, translated, and reviewed by an expert panel for content validity, clarity, and to suggest other items which may have been omitted. Factorial validity and internal consistency reliability were assessed by conducting principal component and parallel analyses, and Cronbach's alpha calculation, respectively, using a data from 1303 completed questions. The intraclass correlation coefficient (ICC) and Cohen Kappa statistic were calculated to evaluate the temporal stability of the instrument. The investigation of health care utilisation and hospital admission for LBP included a total of 1981 people with LBP. The calculation involved a single population proportion formula, with an expected prevalence of health care utilisation for LBP (p = 50%), 95% level of confidence, 4% margin of error, 3 design effect, and a 10% non-response. The study was conducted in June-November 2018 in South-West Shewa zone of Oromia regional state, Ethiopia. The study participants were selected using a multistage sampling technique with a systematic random sampling method. Data were collected using the Oromo language version of the instrument using the interview technique. Data entry was made using Epi-Info version 7.0, where it was exported to SPSS 23.0 and checked for accuracy. Finally, data analyses were carried out using R version 3.5.1. Health care utilisation and hospital admission for LBP were estimated as prevalence rates with 95% confidence intervals (CIs). The log-binomial regression model was fitted to determine prevalence ratios (PR) with 95% CIs in identifying factors associated with health care utilisation and hospital admission for LBP. Estimates of population parameters were also presented with 95% CIs and p-values. For all applications of inferential statistics, a p-value of < 0.05 was taken as the significance level. Results The content validity index of the items forming the newly developed measurement instrument ranged between 0.80 and 1.00 with the modified Kappa coefficient ranged between 0.79 and 1.00. The parallel analysis showed that there were six components with Eigenvalues exceeding the corresponding criterion values for a randomly generated data matrix of the same size. Cronbach's alpha for the internal consistency reliability ranged from 0.65 to 0.82. In assessing temporal stability, ICC ranged from 0.60, 95% CI: 0.23-0.98 to 0.95, 95% CI: 0.81-1.00 while Cohen Kappa ranged from 0.72, 95% CI: 0.49-0.94 to 0.93, 95% CI: 0.85-1.00. The lifetime prevalence of health care utilisation for LBP was 36.1%, 95% CI: 33.9-38.1 and the annual prevalence was 30%, 95% CI: 27.9-32.2. Of the total 543 individuals with a one-year history of presentation to health care facilities for LBP, 78 (14.4%, 95% CI: 11.6-17.3) were hospitalised for the pain, with an average length of stay (LOS) 7.4 days, 95% CI: 6.4-8.8. Several socio-demographic variables, modifiable health behaviours/lifestyle habits, pain interrelated factors, and specific factors, such as beliefs about the pain, depressive symptoms, and sleeping problem/insomnia were independently associated with health care utilisation for LBP. Hospital admission for LBP was also found to be associated with gender, age, living conditions, residential environment, alcohol consumption status, intensity of pain, and presence of additional spinal pain. Conclusions The newly developed measurement instrument has an overall good level of psychometric properties measured as content and factorial validity, internal consistency reliability, and temporal stability. The most decisive factors explaining variations in health care utilisation and hospital admission for LBP were also determined. There were potential inequalities between urban and rural populations in accessing the Ethiopian health care system with relatively better services. This study also highlighted the burden of LBP to individuals and the already overloaded and fragile Ethiopian health care system. It may be prudent that the Ethiopian health care policy makers develop the necessary strategies to meet the health needs of both urban and rural populations with LBP. Further research evidence is also needed on LBP patient referral procedures in the Ethiopian health care system to inform the health policy makers regarding appropriate management strategies capable of dealing with the increasing epidemiology of LBP and associated health needs of people experiencing the pain.