Dementia covers multiple diseases that reduce brain function, affecting memory and the ability to perform daily tasks. People living with dementia have on average a greater number of other medical conditions, known as comorbidities, than those without. Therefore, they are more likely to receive several medications at the same time—known as polypharmacy—and may face greater risk of experiencing drug-drug interactions. Polypharmacy has been found to be associated with an increased risk of adverse outcomes in those already living with dementia.
This study will characterise common medication “clusters”—medications that can be grouped together based on similar usage—in the years prior to and following dementia diagnosis in England. Networks (a system of interconnected medications) will visualize how medication clusters change before and after dementia diagnosis. These medication clusters may highlight potentially inappropriate prescribing in the years surrounding dementia diagnosis, and the common drivers of polypharmacy. Understanding patterns in medication use before and after dementia diagnosis enables the implementation of safe treatment plans and could help focus the development of clinical guidance for treating comorbidities alongside dementia.
Psychotropic drugs affect how the brain works to change mood, cognition, or behaviour, and they are commonly used to manage psychological and behavioural difficulties in those living with dementia. Guidance states that prescribing of antipsychotics—a subset of psychotropic drugs used to treat hallucinations, delusions, or disordered thinking—should be temporary and reviewed regularly. Our proposed study will perform a secondary analysis that only includes psychotropics in the network and clusters.
Dementia diagnosis can complicate treatment plans for comborbidities, resulting in adverse outcomes. We aim to determine how medication clusters for comorbidities change before and following dementia diagnosis, including psychotropics. This study will help understand patterns in medications after dementia diagnosis, which is essential to implementing safe treatment plans and could help focus the development of clinical guidance for treating comorbidities with dementia. Our cohort will include individuals with a dementia diagnosis between 01/01/1998 and 31/12/2022 and are at least 65 years old. The outcomes will be drugs used for comorbidities (excluding vaccines or antibiotics) during the same period. Treatment episodes for each medication will be calculated using the Proportion of Days Covered approach with 20% added to the prescription duration to account for imperfect adherence. There will be a 14-day medication-free grace-period before a new treatment episode is assigned. Co-medication will be defined as overlapping treatment periods of two drugs. Networks will be generated in the 0-5 years, 5-10 years, 10-15 years prior to dementia diagnosis and in the first year, three years, and five years post dementia diagnosis. Nodes will represent medications and size proportional to the prevalence of that drug in the cohort. Edges will be weighted by the proportion of the cohort who were co-medicated with those drugs. Networks will be stratified by sex and dementia subtype. Modularity analysis will be applied to characterise clusters from the networks using clustering methods. Networks will be generated for associations between psychotropic drugs prescribed to investigate psychotropic prescribing overall. Edges will be weighted by the magnitude, adjusted using the Benjamini-Hochberg method, of the partial Pearson’s correlations between the two medications. An interactive tool will be created to enable exploration of the networks created, providing a tool for clinicians to explore patterns in medication in those who are diagnosed with dementia.
1) Patterns of medication clusters, which will be compared between time periods prior to and following dementia diagnosis. Medications included will be drugs used for comorbidities and will include drugs that impact the gastro-intestinal, cardiovascular, respiratory, central nervous, endocrine systems, as well as drugs for obstetrics, gynaecology, urinary-tract disorders, malignant disease, immunosuppressants, nutrition and blood disorders, and musculoskeletal and joint disorders.
2) Medication clusters of psychotropics, i.e. comprised of just psychotropic drugs used after dementia diagnosis.
In both cases, medications are identified by British National Formulary (BNF) chapters.
Wallis Lau - Chief Investigator - University College London ( UCL )
Olivia Bryant - Corresponding Applicant - University College London ( UCL )
Jan Chobanov - Collaborator - University College London ( UCL )
Kenneth Man - Collaborator - University College London ( UCL )
Li Wei - Collaborator - University College London ( UCL )
HES Accident and Emergency;HES Admitted Patient Care;HES Outpatient;ONS Death Registration Data