Background
Low back pain (LBP) is a common health condition worldwide [
1] and poses a significant burden on individuals’ quality of life [
2]. In primary care, 60% of patients presenting with LBP also report leg pain, a proportion of those patients will have sciatic pain [
3]. The presence of both, low back-related leg pain and sciatica is linked with increased pain, disability, poorer health outcomes, and work absence, compared to LBP without symptoms in the leg(s) [
4], and consequently contributes significantly to the economic burden on individuals and society in general. Patients with low back-related leg pain not attributed to sciatica are labelled as having ‘referred’ (non-specific) leg pain, we will use this term when we describe this group.
The economic burden of LBP has been investigated comprehensively, and found to be substantial [
5,
6]. To date, there is limited research on the economic burden of low back-related leg pain, including sciatica. A previous Dutch study estimated societal sciatica costs of about 13% of overall LBP related costs, equivalent to a current annual impact on the United Kingdom (UK) economy of over £500 million in healthcare costs and £3.8 billion in indirect costs related to sciatica [
7,
8]. Given the socioeconomic impact of sciatica, a number of trials have investigated the effectiveness of available treatments, from medications to surgery [
9‐
11], and national and international guidelines recommend timely assessment of patients presenting with sciatica in primary care [
12,
13]. There is however limited research on the determinants of quality of life and cost outcomes in LBP patients, and specifically few studies have investigated patients presenting with sciatica or with referred leg pain, in the primary care setting. A prospective description of costs and quality of life outcomes of this patient group will improve understanding of healthcare determinants, societal costs and health outcomes, and inform health care policy and planning.
The objective of this study is to provide data on the distribution of healthcare usage, costs and quality of life outcomes in primary care patients consulting with low back-related leg pain according to the presence/absence of sciatica, and explores factors associated with costs and quality of life outcomes for patients with symptoms of low back-related leg pain including sciatica.
Methods
Cohort study design and population
Full details of this prospective cohort study are published in the ATLAS (
Assessment and
Treatment of
Leg pain
Associated with the
Spine) study protocol [
14], and results papers [
15,
16]. The study’s aims and methods are briefly described here. Ethical Approval for this study was obtained from the South Birmingham Research Ethics Committee (REC ref. 10/H1207/82). The primary objective of the ATLAS study was to investigate overall prognosis of patients seeking care for symptoms of low back-related leg pain, including sciatica. The study population consisted of 609 adults aged 18 years and over with symptoms of low back-related leg pain, including sciatica, of any severity and duration, recruited from the National Health Service (NHS) general practices in North Staffordshire and Stoke-on-Trent. Patient recruitment was carried out between April 2011 and March 2013. Adults, visiting their General Practitioner (GP) with low back-related leg pain including sciatica, of any duration and intensity, who met pre-set study criteria, were invited to participate. At a research clinic, all study patients after giving written informed consent were assessed by a physiotherapist to confirm eligibility [
15,
16]. Patients also received a magnetic resonance imaging (MRI) scan (within 10 working days) as part of the research study, providing there were no contraindications to the procedure. Patients were clinically diagnosed as having sciatica or referred (non-specific) leg pain based on the examiner’s clinical opinion. Participation in the study did not infer any particular treatment advantage for the participants, treatments and treatment pathways were based on best clinical practice. In terms of care pathways, most patients received a course of physiotherapy, a small number of patients were referred to a specialist spinal service for consideration of spinal injections and/or surgery, and a small number of patients proceeded to have these procedures.
Overview of economic analysis
An economic regression-based analysis of costs and quality of life outcomes was conducted alongside the observational cohort study. The economic analysis used costs and QALYs as the main outcomes, with the analysis conducted from an NHS and societal perspective and designed to capture data at 4 and 12 months. The main focus of the analysis was to investigate the distribution of resource use and quality of life outcomes in these patients with referred leg pain or sciatica, and the predictors of costs and quality of life outcomes in the whole cohort, with a secondary analysis focusing on the subgroup clinically diagnosed with sciatica.
Quality of life outcome data
Preference-based health outcome data were collected at baseline, 4 and 12 months follow-up, using the patient completed EQ-5D-3 L questionnaire. The EQ-5D-3 L is a generic instrument measuring and valuing health related quality of life [
17]. Responses from individuals were converted to utility values obtained using the UK value set derived from a UK general population survey [
18] and expressed in QALYs using the area-under-the-curve approach linking utility scores at various time-points [
19]. Adjustment for differences between patients with or without sciatica in baseline EQ-5D-3 L scores was performed using a regression-based adjustment in order to avoid bias [
20]. Other secondary outcomes were also collected using self-completed questionnaires [
21‐
27] (see Additional file
1).
Resource use and cost data
Resource-use data due to low back-related leg pain/sciatica were collected from participants at 4 and 12 months from the time of recruitment into the study, using self-report postal questionnaires. The questionnaires specifically requested information on low back-related leg pain/sciatica healthcare resource utilisation including primary care consultations (e.g. GP, practice nurse, physiotherapist), prescribed medication, over-the-counter treatments and secondary care attendances including healthcare professionals (e.g. hospital consultants and physiotherapists), investigations (e.g. e.g. X-rays, MRI scans), and procedures such as surgery (injections, surgeries). Self-reported work-related data on time-off work were also collected in order to assess the impact of indirect costs of sickness absence due to low back related leg pain. Details of the number of study-related physiotherapy sessions attended by each participant were collected as part of the study through case report forms and costed separately from other physiotherapy visits. Study protocol-driven MRI scan costs were excluded from the analysis as these would not necessarily occur in usual practice.
Total 12-month costs per person were estimated by combining resource use data with unit costs. Unit costs were obtained from the British National Formulary (BNF) [
28] for drugs, and the NHS Reference costs [
29] and Unit Costs of Health and Social Care [
30] for other resource use items (see Additional file
2). Productivity costs were estimated using the human capital approach (wage cost per day multiplied by the number of absence days), salary costs were based on respondent job-specific wage estimates identified from annual earnings data and UK Standard Occupational Classification coding [
31‐
33]. Out-of-pocket treatment costs were based on patient reported costs. All costs were expressed in 2013/2014 UK (£) prices.
Data analysis
An analysis of costs and quality of life outcomes was conducted to determine the difference in costs and QALYs over a 12-month period between the group of patients with sciatica and those with referred leg pain. Multiple-imputation using chained equations was used to impute missing values for costs and EQ-5D-3 L scores for non-responders to the 4 and 12-month questionnaires [
34]. Confidence intervals for the mean differences in resource use and costs were obtained by bias corrected and accelerated non-parametric bootstrapping, using 1000 replications [
35]. Discounting was not performed because of the 12 month follow-up period. Descriptive statistics (mean, standard deviation) and 95% confidence intervals for resource use, costs and QALYs are presented for the whole cohort and separately for the sciatica and referred leg pain groups.
Separate GLMs with log link and gamma variance functions were fitted for the whole cohort, to identify factors that influence total costs and QALYs [
36]. GLM models account for non-normality in the outcome variables. Factors to be examined were selected a-priori based on evidence of their association with costs and health-related quality of life outcomes, building on evidence from previous studies, and expertise within the study team (Additional file
1). GLMs were used to examine the relationship of each factor with the cost (NHS and societal) and QALYs. Factors with
p < 0.25 were carried forward to the multivariable model for each cost and quality of life outcome variables. The final models reported significance at
p < 0.01,
p < 0.05 and
p < 0.1. All statistical analyses were performed using Stata V.13 analysis software [
37]. The base-case analysis was from a UK NHS perspective, using the imputed cost and QALY dataset while adjusting for baseline EQ-5D-3 L scores and a secondary analysis from a societal perspective. As part of sensitivity analysis, separate models were fitted to identify predictive cost and outcome factors for the sciatica group only.
Discussion
In this paper we report the total costs and quality of life outcomes of primary care patients seeking care for symptoms of back-related leg pain including sciatica, and we describe the factors associated with costs and quality of life outcomes in this patient group. To our knowledge, this the first study to provide information on costs (NHS and broader societal costs) and quality of life health outcomes, and on factors associated with these, in this population.
Our results from the overall cohort analyses demonstrate that healthcare utilisation from the NHS perspective is highest for health professional consultations such as physiotherapy and GPs. From a societal perspective, costs were highest for work-related productivity loss (for those in work). Similar findings were observed in the analyses for the sciatica and referred leg pain groups. An increase in the number of physiotherapy sessions (as part of the ATLAS study) and higher self-reported general health scores were significantly associated with reduced resource utilisation and costs at 12 months (approximately £1.5 and £1.2 respectively). The baseline factors associated with improvement of quality of life in this cohort were: higher scores of self-rated general health, lower pain intensity, lower depression scores, and lower disability scores. However, the disability score was not a significant predictor of quality of life in the sciatica group. A further finding was a higher number of days off work in the sciatica group within the 12 months follow-up, resulting in higher societal costs.
LBP national and international guidelines have recommended early assessment, diagnosis and management of back-related leg pain in patients presenting with back-related problems, with the view to prioritise timely and appropriate management [
12,
38]. However, information on how costs and health related quality of life outcomes vary between patients with sciatica and referred leg pain remains limited, and patterns of how costs and outcomes change over time, have not been explored. No previous studies have assessed the impact of patient demographic and clinical characteristics on the costs and quality of life outcomes for this primary care population.
We did not find any significant association between costs and age, gender, body mass Index, comorbidities, disability, pain intensity, and depression in our sample. This contrasts with more general literature which suggests that costs of musculoskeletal conditions are often associated with demographic factors such as age, and disease specific factors such as depression and disability [
39‐
43]. Factors significantly associated with quality of life outcomes in this cohort included general health, disability, pain intensity, and depression.
The strengths of our study lie in the analysis approaches used, presentation of disaggregated results and comprehensive assessment of factors associated with costs and quality of life in patients consulting with symptoms of low back-related leg pain including sciatica. The analyses are also performed from a broader societal perspective, and therefore report important work-related outcomes that are particularly relevant for this patient group, as most are of a working age. The analysis considers a comprehensive set of potential prognostic factors influencing cost and quality of life, underpinned by previous research and clinical experience. However, there are limitations to our analysis. Resource use information was collected using self-reported data from study participants at 4 and 12 months questionnaires. A limitation of using self-report data only, is that it is subject to recall and information bias (e.g patients not able to distinguish between LBP-related resource use and others) and therefore respondents could potentially over-report or under-report resource utilisation, particularly over longer periods of recall [
44]. Nevertheless self-reported data provide an efficient means of obtaining economic evaluation data in the absence of routine data sources and has been widely used in similar studies [
45‐
47]. Response rates for the main cost (60% at 12 months) and EQ-5D 3 L outcome data (97% at baseline, 64% at 4 months and 68% at 12 months) were low raising some concerns about the validity of the findings. However, appropriate and robust multiple imputation approaches were used to address potential biases resulting from incomplete data. In addition, data for the ATLAS study clinic sessions (which was significantly associated with total costs) was observed from routine physiotherapy databases, and could have influenced the overall results of the model. Lastly, although we evaluated costs from a societal perspective, we were not able to include all components of the indirect costs. Our study therefore presents conservative estimates of the total societal costs since presenteeism related costs were not included in the cost analysis. These costs have been shown to be greater than absenteeism in some cases [
48].
The analysis reported here presents important information on costs for treating patients with back-related leg pain including sciatica. Within the ATLAS cohort, patients with sciatica pain and referred leg pain appeared to be similar in terms of quality-of-life outcomes although patients in the sciatica group had a higher proportion of consultation visits, medication and time-off work. The productivity costs incurred by individuals in this cohort as a result of back-related leg pain were substantial. Therefore it is important that the cost-effectiveness of interventions used for patients with low back-related leg pain and sciatica, are investigated from a perspective of both, the health provider and society.
The regression analyses showed that patient self-rated general health and the number of physiotherapy sessions received (as part of the ATLAS study) were important predictors of resource utilisation and costs, possibly indicating that initial comprehensive treatment seems to reduce future need for health care use and costs. However, the study’s observational design precludes any causal inferences. Baseline low levels of disability, pain intensity and depression were associated with changes in quality of life as measured by the EQ-5D-3 L. The generalizability of the findings in this research is supported by the inclusion of eligible patients with any level of intensity and duration of pain, as this broadens the population beyond patients presenting with only the most severe symptoms.
Conclusion
In conclusion, in 1 year, resource utilisation for primary care patients with low back-related leg pain and sciatica showed that health care and societal costs were highest for visits to physiotherapists and GPs and in relation to work-related productivity loss. Health-related quality of life was low at baseline for the overall cohort and for both the sciatica and referred leg pain groups, but improved across all time points. The type of care received and patient self-reported general health were important variables for predicting future costs. Similarly to other musculoskeletal conditions, quality of life was significantly influenced by characteristics such as levels of disability, depression and pain intensity. Our study contributes to understanding the economics of low back-related leg pain and sciatica care and has important implications for health care policy resource allocation in this patient group.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.