Aldhobaie, Ghadah Yousf A
(2023)
Multimorbidity and Polypharmacy: A Health Informatics Approach.
Master of Philosophy thesis, University of Liverpool.
|
Text
201524350_Jun2023.pdf - Author Accepted Manuscript Download (3MB) | Preview |
Abstract
Introduction: Multimorbidity is increasing in prevalence, and is more common in older age groups. All bodily systems can be affected by multimorbidity (or multiple long-term conditions), and this is associated with increased healthcare utilisation and increased mortality. Additionally, people living with multiple long-term conditions are often on many drugs, which has been termed polypharmacy. This puts them at risk of adverse drug reactions, drug-drug interactions and poor adherence, all of which can increase healthcare costs. In this thesis, I have focused on cardiovascular and gastrointestinal drugs in order to understand the changes which have occurred in the usage of these drugs over the last two decades. Objective: The thesis aims to describe the changing patterns of medication prescription practice over the last two decades in older people with cardiovascular and/or gastrointestinal diseases using the Clinical Practice Research Datalink (CPRD). Methods: We extracted one million patient records from CPRD. Among these patients, we have included cardiovascular (CV) and gastrointestinal (GI) patients with two or more years of follow-up. These patients were then filtered further based on the second chronic condition and the patient’s age (whether they were 50 years and above). After adjusting for the range between the years 2001 and 2020, the number of eligible patients for this study was reduced further (used for the first group analysis). An association rule mining was applied to investigate the prescription pattern between 2001 and 2020. When this is filtered to those who had 20 years of follow-up appointments, it reduced the number and was analysed accordingly for the second group analysis to examine the change in the prescription patterns by applying specifically the Apriori algorithm Association Rules Mining. Results: For the first analysis, the extracted eligible CV and GI patients were 228,376 and 111,355, respectively. The eligible patients for the second analysis with continuous and constant 20 years follow-up were 17,075 and 3,110, respectively. In cardiovascular disease, the most commonly prescribed drug classes were statins followed by calcium channel blockers. The study also showed that there was a statistically significant increase (P-value <0.0001) in the mean number of total cardiovascular drugs prescribed in 2020 in comparison to 2001 (mean 2.201-1.581, standard deviation 1.153-0.860 respectively). In gastrointestinal disease, the most commonly prescribed drug classes were proton pump inhibitors followed by corticosteroids. In the same study period, there was a statistically significant increase (P-value <0.0001) in the mean number of total gastrointestinal drugs prescribed (mean 1.284-1.152, standard deviation 0.586-0.424 respectively). Evaluation of the 20 year follow-up data allowed association rule mining to be applied and the top 10 rules were identified. Conclusions: There was a statistically significant difference in prescription patterns during the study period. In particular, the total number of drugs increased significantly in 2020 when compared to 2001. A limitation of the studies is that only drugs associated with CV and GI diseases were evaluated, and further studies on all systems are needed. Researchers and clinicians need to better understand drug prescribing patterns overall in patients with multiple long-term conditions to develop strategies to overcome the possible adverse consequences of polypharmacy.
| Item Type: | Thesis (Master of Philosophy) |
|---|---|
| Divisions: | Faculty of Health and Life Sciences |
| Depositing User: | Symplectic Admin |
| Date Deposited: | 14 Nov 2023 15:38 |
| Last Modified: | 14 Nov 2023 15:39 |
| DOI: | 10.17638/03173717 |
| Supervisors: |
|
| URI: | https://livrepository.liverpool.ac.uk/id/eprint/3173717 |
Dimensions
Dimensions
