Background Anticipatory medications (AMs) are injectable drugs prescribed to a named patient, ahead of possible need, for administration if distressing symptoms arise in the final days of life. Little is known about how and when drugs are prescribed in primary care.

Aim To investigate the frequency, timing and recorded circumstances of injectable end-of-life AMs prescribing decisions for patients living at home.

Methods Retrospective mixed methods observational study using GP and community nursing records for 329 deceased adult patients registered with 11 GP practices in two UK counties (30 most recent deaths per practice). Patients died between 2017 and 2019. Multivariable logistic regression models were built to detect key factors associated with AMs prescribing.

Results AMs were prescribed for 167/329 (50.8%) of patients, between 0 and 1212 days before death. The median prescribing timing was 17 days before death across all GP practices, with a range of median 3 to 33 days in individual GP practices. The likelihood of AMs prescribing was significantly higher for patients with a recorded preferred place of death (odds ratio [OR] 34; 95% CI 15–77; p < 0.001) and specialist palliative care involvement (OR 7; 95% CI 3–19; p < 0.001). AMs were typically prescribed (114/167, 68.3% patients) as part of one main end of-life planning intervention when preferred place of death and/or Do Not attempt Cardiopulmonary Resuscitation (DNACPR) discussions were also first recorded. Standardised prescribing of four drugs and doses was commonplace and prompted by primary care electronic end-of-life templates.

Conclusion Standardised AM prescribing patterns and variability in the timing of prescriptions highlight challenges in diagnosing dying, and the risks involved in prescribing far in advance of possible need. Our findings warn about the dangers of electronic end of life templates and the bureaucratisation and standardisation of end-of-life care planning interventions.