Dementia in observational studies using electronic health records : opportunities and challenges with pharmacoepidemiology

Abstract

Dementia is a global health challenge that currently affects more than 50 million people internationally with increasing prevalence. Evidence to support safe medication use in people living with dementia and therapies that may reduce the risk its development can help to facilitate healthy ageing and address the growing dementia burden.

Studies of dementia face challenges such as requiring a long period of follow-up due to the gradual nature of its development, a lack of consistent dementia diagnosis definitions, and potential confounding bias when making comparisons between different groups of older people with different degrees of frailty and disease states. Although randomised controlled trials are the gold standard of investigating the effectiveness of interventions, their interpretation in studies of dementia are limited due to their restricted study samples in terms of size, age, and follow-up time. Therefore, alternative study types that can circumvent the weaknesses of clinical trials, such as observational studies, are crucial in bridging the knowledge gap in dementia research.

The aim of the thesis is to investigate several topics surrounding safe medication use related to dementia and highlight opportunities and challenges in using observational study designs within this disease area. The data used in the studies included in this thesis were electronic health records extracted from the population-wide database managed by the Hong Kong Hospital Authority. Two of the studies were collaboration projects with the addition of electronic health records from the United Kingdom, Sweden, and Australia.

Observational studies applying pharmacoepidemiologic methods were conducted in populations of people with dementia or with dementia as an outcome, with COVID-19 vaccines and antihypertensive drugs being medicines of interest. Studies included were an analysis of COVID-19 vaccine adverse events in people with dementia, multinational studies on prescribing trends of antihypertensive drugs in people with dementia and the association of the use of different antihypertensive drug classes and the risk of incident dementia, and the application of high-dimensional propensity scores as a method to mitigate confounding bias.

The findings of these studies comprised of evidence supporting the safe use of COVID-19 vaccines in people with dementia, descriptions and analyses of antihypertensive prescribing practices, and head-to-head comparisons between the use of different antihypertensive drug classes and the risk of incident dementia. Angiotensin-II receptor blockers were found to have a reduced risk of incident dementia compared to angiotensin-converting enzyme inhibitors in a large, multinational cohort study setting with a relatively long follow-up of more than five years. The high-dimensional propensity score method was also successfully implemented using electronic health records in Hong Kong in a proof of concept study to improve confounding control between the antihypertensive class comparison groups in older people.

To build on the findings presented in this thesis, future studies are needed, such as studies aiming to elucidate the mechanisms of antihypertensives in dementia risk reduction, and advances in study methods to reliably evaluate the relationships between these medicines and dementia.