Alaujan, Shiekha
(2018)
A programme of feasibility studies to develop a randomised controlled trial to assess pain management services for people with fibromyalgia or chronic widespread pain.
PhD thesis, University of Nottingham.
Abstract
Background
Fibromyalgia (FM) and chronic widespread pain (CWP) are highly prevalent chronic painful conditions that have substantial impact on patients, health care systems, and society. Diagnosis is complex and management strategies are associated with various levels of evidence for effectiveness and cost-effectiveness. Pain management services (PMSs) have been shown to be effective in some settings and are therefore recommended by clinical practice guidelines as a rational treatment. As these services are resource intensive, evidence is needed to demonstrate their cost-effectiveness.
Quality-adjusted life-years (QALYs) are commonly used as an outcome for the economic evaluation of health interventions. To perform an economic evaluation, mapping from a disease-specific measure onto a generic preference-based measure is essential when utility values of health states are not available in the same sample. In the field of FM/CWP, mapping studies that have estimated the EuroQoL 5-dimensional 3-level instrument (EQ-5D-3L) from the Revised Fibromyalgia Impact Questionnaire (FIQR) are lacking.
Aims
The aim of the research was to conduct a programme of feasibility work in preparation for the conduct of a definitive randomised controlled trial (RCT) of PMSs in managing people with fibromyalgia or chronic widespread pain, compared with standard care (SC). The programme included a systematic review, a feasibility observational study, an economic modelling, and a statistical mapping of FIQR to EQ-5D-3L.
Methods
The study followed the Medical Research Council (MRC) recommendations for evaluation of complex interventions through an iterative process of observation work before evaluation by large scale studies. Therefore, this work comprised four phases. Firstly, a systematic review was performed, using standard methods of systematic review along with quality and reporting criteria for published economic evaluation studies of PMSs in people with FM/CWP. An electronic search was performed in clinical and economic databases, from their inception to April 2016. Full economic evaluations undertaken from any perspective were included. The Cochrane Risk of Bias tool and the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) checklist were used to assess the methodological quality.
Secondly, a prospective observational multicentre study, using patient-level data was conducted among newly referred adults with FM/CWP (≥18 years-old) at two community-based PMSs in Nottingham. The feasibility of the study and recruitment, retention and follow-up rates were assessed. Participants provided details of clinical outcomes and health resource-use, using a postal questionnaire and diaries, at baseline and then three and six months after recruitment. Outcome measures were the FIQR, the Brief Pain Inventory (BPI), and the EQ-5D-3L.
Thirdly, data from the observational study, systematic review and secondary data sources were then integrated in a de novo Markov decision analytic model to assess the cost-effectiveness of PMSs compared with SC. Data included transition probabilities, utilities, healthcare resource use and their costing tariff. Costs per extra QALY were calculated from the perspective of National Health Services (NHS) England, using a lifetime horizon. For validating the decision analytic model, the Assessment of the Validation Status of Health-Economic decision models (AdViSHE) criteria were followed.
Finally, access to an anonymised dataset of 160 Spanish adults (≥18 years-old) with a confirmed diagnosis of FM, according to the American College of Rheumatology (ACR) 1990 criteria, was obtained from a researcher, to perform the mapping from the FIQR to the EQ-5D-3L. The econometric models investigated were ordinary least square (OLS), censored least absolute deviations (CLAD) and multinomial logistic (MNL) models. The choice of model was based on the ‘Best’ performance, which was defined as: the lowest absolute difference (AD), mean absolute error (MAE), and root mean square error (RMSE), and the highest R2 statistics.
Results
In the systematic review, six economic evaluation studies of PMSs in people with FM/CWP were identified. However, due to methodological weakness and variability, the review was unable to draw a firm conclusion about the cost-effectiveness. In addition, all identified studies were performed alongside randomised controlled studies (RCTs) with a short follow-up period. As FM/CWP are chronic diseases, the effectiveness and cost-effectiveness over longer periods of time need further investigation, and it is likely that this will require the use of economic evaluation modelling approaches.
Recruitment to the observational study was a significant challenge. Despite modifications to the study protocol to enhance recruitment and retention, only 14 participants were identified, of whom 10 were recruited and eight completed the study (July 2015-September 2016). Hence, the effectiveness and healthcare resource use data were affected by the small sample size
Following the integration of data from the feasibility work and published studies, PMS generated a mean of 15.2 (2.5% percentile -1.7; 97.5% percentile 33) more QALYs per patient, at a mean reduced cost of -£10453 and -£11148.8 (2.5% percentile -£31420.2; 97.5% percentile £3494.5) for deterministic and probabilistic analysis, respectively. PMS has a 97% probability of cost-effectiveness compared with SC at a willingness-to-pay of £20000 per QALY. Sensitivity analysis demonstrated that targeting each disease severity with PMS has a probability over 0.96 of cost-effectiveness compared with SC, at a willingness-to-pay of £20000 per QALY.
In the mapping sample, it was possible to map from the FIQR to the EQ-5D-3L. The predicted mean utilities in all OLS models were identical (up to four decimals) to the observed means. Among the ten specifications used, the best performing model was an OLS model, which included the FIQR (collapsed items in a discrete form), age, and educational level as independent variables in the regression. This model produced lowest errors (AD = 0.0000; MAE = 0.1279; MSE = 0.1576; R2= 0.4675) compared with the more advanced CLAD and MNL models.
Conclusions
The programme of feasibility work performed in this thesis found that due to the low sample size in the feasibility observational study, it has been concluded that it is not feasible to perform a future RCT to evaluate the cost-effectiveness of PMS for people with FM/CWP. However, this study has highlighted key methodological issues which should be taken into consideration in future studies. Moreover, the systematic review highlighted has key design areas that researchers need to consider when conducting economic evaluations, including following standard design guidelines. The indicative economic evaluation of PMS for people with FM/CWP suggested that PMS can lead to a significant health gain and cost savings to the NHS. Finally, it was found that mapping from the FIQR to the EQ-5D-3L is possible. The OLS was the best performing model, which predicted the mean tariff score of the EQ-5D-3L with an accuracy of up to four decimal places. Findings from the statistical mapping will enable future researchers to perform economic evaluations in the absence of EQ-5D-3L data from RCTs of FM/CWP.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
![[img]](/style/images/fileicons/application_pdf.png)