Application of Swedish quality register data for use in health economic analyses of chronic conditions

Abstract

Sweden has many registers to monitor and follow-up healthcare, and combined with the unique personal identification numbers, this represents vast opportunities for register based research. Part of these data sources are the Swedish quality registers, which are set up to monitor the quality of care of specific diseases. Two of these registers that both have good national coverage over time and are used for research purposes are the Swedish Rheumatology Quality Register (SRQ) and the Swedish Stroke Register (Riksstroke). These are set up to monitor the quality of care of patients with rheumatoid arthritis (RA) and stroke, respectively, both diseases being associated with a chronic condition of functional disability. Data from the two quality registers have been used for scientific research in various fields, but to a lesser extent in health economic analyses. Health economics addresses issues relating to the allocation of scarce resources to improve health. This includes resource allocation both within the economy to the healthcare system and within the healthcare system to different activities. Two common health economic approaches are economic evaluations of specific therapies and burden-of-illness studies, taking a broader analytical approach to a disease.

The overall aim of this thesis was to study the applicability of quality register data in health economic analyses of chronic conditions. The specific aims were: • To assess the applicability of data from a quality register in economic evaluations of anti-TNF treatment for RA in clinical practice; • To assess the applicability of data from a quality register in burden-of-illness studies where health outcomes, resource use and costs of stroke are put in relation to each other.

Data from the quality registers were used and linked to other relevant data sources to address the aims of this thesis. For the economic evaluations of treatments in RA, health economic models were constructed to enable the analyses. Statistical analyses were performed to allow for hospital comparisons of health outcomes and resource use for stroke, as well as estimating long-term transition probabilities for the health economic model in RA.

The papers concluded that: • Anti-TNF therapies have on macro level been used cost-effectively as first-line biological treatment for RA in Swedish clinical practice. The cost-effectiveness results did not differ depending on the source of effectiveness data (clinical practice or RCT trial). However, the results were sensitive to the underlying progression rate of the comparator and assumptions made in the model. • The impact of stroke on health outcomes, resources and costs were substantial. There were differences in performance between hospitals in these indicators that could not be explained by differences in patient mix. The results further indicated that the costs differed by level of functional disability and age, up to two years post stroke. Further, the papers of this thesis demonstrated the valuable contribution of quality register data in health economics in providing a valid base of data and opportunities to: • Assess real life effectiveness of treatments in economic evaluations; • Retrieve data on health outcomes and patient characteristics, which are essential in: • Measuring health outcomes and relating them to levels of resource use; • Enabling hospital comparisons of performance and performing case-mix adjustment of results; • Enabling stratification of cost estimates by level of health outcome; • Provide input parameters for future economic evaluations.

In order to assess the full health economic aspects of chronic conditions, quality registers play an important role, but there is a necessity to combine the quality register data with other registers or other data sources, published literature and potentially also conduct modelling to account for the long-term effects. Nevertheless, any quality register that wants to ensure that the data can be used for health economic analyses and provide valid data for such analyses, should consider to: • Ensure long-term follow-up of the patients (especially in chronic conditions); • Collect data on: • Patient characteristics, including the clinical markers important for the patient’s prognosis; • Treatments received; • Health outcomes that are common as measurements of treatment outcome (e.g in RCTs), as well as estimation of quality of life or utilities; • Resource use (quantifiable) outside the healthcare sector, e.g. informal care. The quality registers and other register data sources can be utilized to a greater extent in different assessments which share the aim of improving healthcare delivery and increasing its value – either by assessing level of health outcomes, processes and resources used; enabling comparisons between treatments or hospitals; or assessing determinants for different outcomes.