Using electronic health record data to evaluate the epidemiology and management of inflammatory arthritis

In healthcare, there are opportunities to utilise the growth of routine data capture in developing real-world evidence of chronic disease. Inflammatory arthritis encompasses a number of chronic diseases including gout, rheumatoid arthritis (RA) and ankylosing spondylitis (AS), for which timely treatment is necessary to limit joint damage. The hypothesis underlying this thesis is that the epidemiology and management of inflammatory arthritis can be evaluated using routine electronic health record (EHR) data. This was investigated through literature reviews and retrospective studies using a population-based primary care dataset. Gout, AS and RA studies have used EHR data, and this thesis identified variation in methods that influenced reported trends in epidemiology and management. For future studies, considerations were raised for improving the reporting and assessment of EHR-pertinent biases. Incidence and prevalence are uncertain in AS, and have not been investigated in RA in recent years following the incentivisation of diagnostic recording. Between 1998 and 2017, this thesis identified that AS incidence declined for ten years before it stabilised, while RA incidence trends were unclear, and prevalence rose in older patients. In an ageing population, managing these diseases is important and studies should consider changes in coding practice in the study period. There have been efforts to reduce diagnostic delay in AS. This thesis found no improvement in time to diagnosis over two decades, largely driven by delay in rheumatology referral. This is concerning given the importance of treatment in early AS. In RA, shifts in management principles have increased DMARD prescribing over time. This thesis identified that the prescribing of potentially toxic corticosteroids and non-steroidal anti-inflammatories nonetheless persisted across the last 20 years with suboptimal prophylactic therapy. This thesis provides evidence of, and raises considerations for further improving, the use of EHR data in evaluating the epidemiology and management of inflammatory arthritis