Exploring cardiovascular risk and outcomes in primary care consulters for osteoarthritis using longitudinal electronic primary care data

Abstract

For adults aged 45 and over in the United Kingdom (UK), osteoarthritis is a common reason for consultation with primary care. Cardiovascular disease (CVD) is a common comorbidity with osteoarthritis; two in five people with osteoarthritis also have CVD. In the UK, primary care is often the first point of contact for people in need of osteoarthritis management and CVD preventive services.

This thesis aims to assess the prevalence of CVD risk factors, the level of CVD risk and outcomes in primary care consulters with osteoarthritis in comparison to those without osteoarthritis. To address this question, the thesis includes a series of retrospective cohort studies using CPRD GOLD, a large representative primary care EHR database, and linked data in the UK between 1992-2017. The consistently higher prevalence of modifiable cardiovascular risk factors (CVRFs), variables that are associated with a higher risk of CVD events, in the osteoarthritis population was revealed in chapter 2 and the socioeconomic inequality in these CVRFs was found in chapter 3. The thesis could then try to understand that if ruling out the impact of CVRFs, what is the excess risk of CVD due to osteoarthritis. The thesis then tried to use the established risk score (sex-specific Framingham risk score) to predict the CVD risk and to assess the following preventative treatments in the osteoarthritis and non-osteoarthritis populations in Chapter 4. If the score performs well, we could then estimate the risk difference (excess risk) based on the established risk score. Unfortunately, its overall prediction performance was not good, as found in chapter 5. We then applied the frailty Cox model (conditional Cox model) based on the matched cohorts, controlling demographical variables, period effect, and coding variation at the practice level, adjusting lifestyles, body measurements, clinical measurements, and treatments to estimate the excess risk of CVD due to osteoarthritis. As the finding in Chapter 6, there was a significant excess risk of CVD for osteoarthritis, and higher in men, the elder and the deprived population. Importantly, the excess risk was also observed in the youngest age group (35-44 years). There could be potential unmeasured confounders, but the findings consistently suggest that clinical effectiveness, cost-effectiveness, and acceptability of potential preventive care strategies under the current screening criteria should be further addressed in the osteoarthritis population in the UK.