USA) for their advice regarding study design. The authors would also like to thank the physicians who participated in the Delphi consensus panel survey, and the residents, their relatives and care home staff who participated in the feasibility study.
AbstractBackground: No studies have been conducted in the United Kingdom (UK) context to date which categorise medications in terms of appropriateness for patients with advanced dementia, or which examine medication use in these vulnerable patients.Objectives: To categorise the appropriateness of a comprehensive list of medications and medication classes for use in patients with advanced dementia; to examine the feasibility of conducting a longitudinal prospective cohort study to collect clinical and medication use data;and to determine the appropriateness of prescribing for nursing home residents with advanced dementia in Northern Ireland (NI), using the categories developed.Methods: Three-round Delphi consensus panel survey of expert clinicians, to categorise the appropriateness of medications for patients with advanced dementia (defined as having Functional Assessment Staging [FAST] scores ranging from 6E to 7F). This was followed by a longitudinal prospective cohort feasibility study which was conducted in three nursing homes in NI. Clinical and medication use for participating residents with advanced dementia (FAST scores ranging from 6E to 7F) were collected and a short test of dementia severity administered.These data were collected at baseline and every three months for up to nine months or until death. For those residents who died during the study period, data were also collected within 14 days of death. The appropriateness ratings from the consensus panel survey were retrospectively applied to residents' medication data at each data collection timepoint to determine the appropriateness of medications prescribed for these residents.Results: Consensus was achieved for 87 (90%) of the 97 medications and medication classes included in the survey. Fifteen residents were recruited to participate in the longitudinal prospective cohort feasibility study, four of whom died during the data collection period. Mean numbers of medications prescribed per resident were 16.2 at baseline, 19.6 at three months, 17.4 at six months and 16.1 at nine months respectively. Fourteen residents at baseline were taking at least one medication considered by the consensus panel to be never appropriate, and approximately 25% of medications prescribed were considered to be never appropriate. Postdeath data collection indicated a decrease in the proportion of never appropriate medications and an increase in the proportion of always appropriate medications for those residents who died.
Conclusions