How do GPs approach deprescribing of long-term medicines within the context of limited life expectancy?

Problem

Polypharmacy is a widespread, growing challenge for health care. Many medicines associated with polypharmacy have long-term benefits, but patients with limited life expectancy (e.g. in the last year or two of life) may not live long enough to benefit fully. In such situations, deprescribing (the supervised withdrawal of medicines) may be appropriate. However, evidence suggests many long-term medications are continued even up to the final few weeks of life, which may reflect difficulties in reliably determining future life expectancy or uncertainties in the risks or benefits of deprescribing. This study sought to better understand GP attitudes and approaches to deprescribing of long-term medicines within the context of limited life expectancy, including the role of individualised life expectancy and risk information in informing the deprescribing process.

Approach

Semi-structured interviews were conducted with practising GPs. Interviews addressed current approaches to ascertaining life expectancy and medication risks/benefits in routine practice, and the role of individualised patient information in informing decision making. During the 40–50-minute interviews, interviewees were presented with hypothetical clinical vignettes focused on deprescribing statins, representing patients with differing clinical characteristics (including life expectancy and risks of a cardiovascular event in the remaining lifetime) to stimulate discussion. Audio-recorded interviews were transcribed, anonymised, and analysed thematically utilising a data-driven inductive approach.

Findings

Fifteen GPs from 12 general practices within the Bristol area participated. Mean time as a practising GP was 14 years (range 6-31 years), and 9 participants were male. All participants perceived deprescribing of long-term medicines to be important for patients with limited life expectancy. In assessing eligibility for discontinuation of long-term medicines, GPs tended to consider factors such as frailty, recent admission to hospital or nursing home and history of falls as well as number of medicines. Alongside life expectancy estimates, individualised risk data was considered helpful for informing and guiding discussions with patients around stopping medicines. Many GPs, however, expressed a reluctance to evaluate, document and share estimates of life expectancy with patients unless this information has been specifically requested. Reasons for this were a lack of confidence in life expectancy estimates, uncertainty around how best to communicate this information and concern about the patient response.

Consequences

For patients with limited life expectancy, deprescribing is considered important by GPs, with individualised risk data having an important role to play in decision making. However, further research is required to understand how best to overcome barriers to the use of explicit estimates of life expectancy to support deprescribing of long-term medicines in these patients.