Using enriched quality register data for health care evaluation : examples from rheumatoid arthritis

Abstract: Swedish registers have for decades successfully been used for medical research, enabling long-term follow-up of large patient cohorts using observational designs. With a tax funded health care system and the use of a personal identity number as a unique identifier, together with national health registers as well as national demographic registers, we have to a relatively low cost the possibility to answer an abundance of research questions in a real-world setting. This thesis describes how the Swedish Rheumatology Quality register (SRQ) can be enriched with national register data for estimating disease occurrence, assessment of health care resource use and work loss, and for potentiating randomized controlled trials (RCTs) conducted within the register framework. More specifically, by using the SRQ that collects disease specific data on patients with rheumatic diseases in routine clinical care together with objectively assessed data from national registers we could in a large sample of patients with rheumatoid arthritis (RA) estimate the incidence and burden of RA. Furthermore, with a randomized pragmatic clinical trial included in the register framework, we could provide information on health economic outcomes regarding a common clinical question in early RA, whether to continue treatment by adding an expensive biologic alternative or to continue with a conventional combination treatment strategy after insufficient response to methotrexate. We found an overall incidence of RA in Sweden on a par with previous local but detailed studies (41 per 100,000; paper I), with a substantial variation across age and sex. We observed that the incidence of RA peaked in the 7th decade in life in both sexes, and that the incidence in women was more than twice the incidence observed in men. Furthermore, we observed lower incidence estimates in individuals with higher education level and in densely population areas. With respect to burden of disease, we used general population comparators individually matched to register-identified subjects with RA and estimated the annual societal cost in prevalent as well as monthly societal cost in newly diagnosed patients to 2-3 times higher than in the general population (paper II). The randomized Swefot trial compared the addition of the biologic drug infliximab versus conventional combination therapy in patients with early RA who had failed initial methotrexate monotherapy. With the Swefot trial included in the SRQ, we could for the first time in a randomized register trial setting analyze work loss and cost-effectiveness for a strategy adding a biologic alternative as compared to conventional combination therapy in methotrexate-refractory early RA. We observed a substantial decrease in mean monthly work loss days in both treatment alternatives, with a reduction of 3 times to double that in the general population from randomization to 21 months of follow-up, but no difference between the strategies could be detected (paper III). This remaining gap to the general population indicates a need for earlier diagnosis as well as for more effective treatment strategies of RA. In the cost-effectiveness analysis we observed similar effects between the strategies over 21 months, while the infliximab strategy incurred higher costs (paper IV), suggesting that an attempt with conventional combination therapy appears reasonable before starting infliximab treatment in methotrexate-refractory early RA, both from a clinical and economic perspective.