People with dementia are often prescribed many different medicines for symptoms of dementia or other health conditions. Some medicines may be inappropriate if they have been used for too long or their harms outweigh their benefits. Doctors are expected to manage medications and are paid to hold an annual dementia care review with every patient. We do not know what effect these opportunities to review medication have on quality of care, an important part of which is reducing the amount of medicines that may be harmful in people with dementia. Detailed medication recommendations have been created to help health professionals reduce harmful prescriptions. We have used these guides to select four groups of medications that may be inappropriate for people with dementia. They are antipsychotic medications, tricyclic antidepressants, anticholinergics (defined according to Anticholinergic Cognitive Burden scale score = 3) and proton pump inhibitors. We will look at how often these medicines are prescribed and the impact of dementia annual reviews and other medication reviews on their use, using GP records of all people with dementia available in the Clinical Practice Research Datalink in 2014 and 2015.
Objective: To estimate the (i) prevalence of potentially inappropriate medication use, and overall medication burden among people with dementia between 2015 and 2016 (ii) and the factors associated with potentially inappropriate medication use, in particular the effect of medication reviews and dementia annual review. Methods. A 2-year cohort study following all prevalent cases of dementia at the start date of 01/01/2015, and incident cases between the start date and the study end date of 31/12/2016. Prescriptions of four classes of potentially inappropriate medications (antipsychotics, tricyclic antidepressants, anticholinergics and proton pump inhibitors) will be extracted, as well as any medication review or dementia annual review (exposures) during that period and potentially confounding variables. Statistical analysis. We will estimate the prevalence of potentially inappropriate medication use in people with dementia across the study period. McNemar and Wilcoxon signed rank tests will test the change in each outcome before and after a medication review or dementia annual review. Multilevel logistic regression analyses will then be used to estimate the factors associated with each outcome, and in particular the effect of medication review or dementia annual review on subsequent prescriptions controlling for age, sex, time varying comorbidity, area-level deprivation, GP practice and exception from QOF dementia indicators.
We will extract prescription records for the selected potentially inappropriate medications, defined using World Health Organisation Anatomical Therapeutic Chemical Classification System (ATC) codes. The presence of each of the following PIMs will be coded within each of 12 2-month time periods between January 2015 and December 2016 for each patient conditional on their inclusion in the study for the whole of that period. 1. Antipsychotic medication will be defined using ATC code N0A5A. 2. Tricyclic antidepressants will be defined using ATC codes. Antidepressants (ATC N06A) are grouped into tricyclic antidepressant sub-type, according to ATC cross-referenced with the British National Formulary (BNF) to identify those used in UK practice (ATC N06AA02, N06AA03, N06AA04, N06AA06, N06AA07, N06AA09, N06AA10, N06AA12, N06AA16, N06AX03 and N06AX05). 3. Definite anticholinergic medication will be defined as any medication with an Anticholinergic Cognitive Burden (ACB) score of 3. An ACB=3 score indicates medications with definite anticholinergic activity and severe negative cognitive effects. 4. Proton pump inhibitors are implicated as potentially inappropriate, according to Beer's and STOPP criteria when they are used to treat peptic ulcer disease at maximum therapeutic dose for >8 weeks. Their use will be defined using ATC code A02BC, at maximum dose and the presence of a Read code indicating peptic ulcer disease (J13y200, J13z.00, 12E1.00, 14C1.00, J130200, J13y100, J13..00). 5. A binary indicator representing the use of any of the above PIMs 6. A continuous indicator of the total number of prescribed medicines within the 2-month time periods. This will include all prescribed medications but exclude devices, dressings and topical preparations. All health outcomes will be measured as the presence of a prescription at any point within the 2-month time periods.
George Savva - Chief Investigator - University of East Anglia
Bryony Porter - Corresponding Applicant - University of East Anglia
Antony Arthur - Collaborator - University of East Anglia
David Wright - Collaborator - University of East Anglia
Duncan Edwards - Collaborator - University of Cambridge
Kathryn Richardson - Collaborator - University of East Anglia
Nicholas Steel - Collaborator - University of East Anglia
Nicholas Steel - Collaborator - University of East Anglia
Practice Level Index of Multiple Deprivation