Patterns of General Practice prescribing in the last year of life

06 Jun 2018

Background: Little is known about patterns of prescribing in the last year of life in primary care. Methods: Recently completed analysis of UK Clinical Practice Research Datalink (CPRD), a large database extracted from GP records that is representative of the UK population. Prescribing data for the last 12 months of life were extracted from 118,571 patients’ GP records who died between September 2010 and August 2015. Two clinicians used British National Formulary codes to identify eight palliative care drug categories: antiemetics, anti-secretory, benzodiazepines, steroids, strong injectable opioids, strong patch opioids, strong enteral opioids, and weak oral opioids. The percentage of patients prescribed each medications was calculated at weekly intervals during the year before death, further stratified by cause of death and age at death. Random effects logistic regression models were run adjusted for age-group, sex, and cause of death. Results: Levels of prescribing significantly increased for all drugs at time of death compared with one year before death (with the exception of weak oral opioids and steroids where results were inconsistent across ages, genders, and causes of death). The largest increase was for 90+ year old male patients prescribed strong injectable opioids (OR: 21032, 95% CI: 3499-12641) and the greatest decrease for females prescribed steroids who died of external causes (OR: 0.52, 95% CI 0.34-0.70). Patients who died of cancer generally showed the most marked increase in prescribing towards the end of life compared with other causes of death. Conclusions: Primary care patterns of prescribing in the last year of life has not been previously investigated. The findings of this recently completed study of a large nationally-representative dataset reveal increasing GP prescribing in response to the rising symptom burden of all patient groups as death approached. The implications for clinical practice and future research will be discussed.