Patterns of filled prescriptions and the association with risk of drug-induced death. A population-based nested case-control register study.

Abstract

Opioid analgesics (OA) and other pharmaceuticals have been associated with drug-induced deaths. However, there is a lack of knowledge regarding patterns of use of these pharmaceuticals in the population and regarding such associations. We identify and describe subgroups of people with different patterns of filled prescriptions of OA and other relevant pharmaceuticals and examine associations with drug-induced deaths. In addition, we estimate the proportion of drug-induced deaths with a filled OA prescription and OA as cause of death. A Norwegian population-based nested case-control register study with cases (drug-induced deaths 2010-2018, N = 2388) and population controls matched for age, gender and year of inclusion (N = 21 465). Patterns of filled prescriptions for opioid analgesics (OA), benzodiazepines and benzodiazepine-related drugs, gabapentinoids, ADHD medication and antidepressants/antipsychotics were explored by k-means cluster analysis. Associations with drug-induced deaths were estimated by conditional logistic regression adjusted for sociodemographic characteristics. Overlap of filled OA prescriptions and OA as cause of death was estimated. Five clusters were identified: 'few prescriptions', 'weak OA', 'ADHD medication', 'sedative/psychiatric morbidity' and 'strong OA'. The 'strong OA' cluster had higher socioeconomic status compared to the other groupings. The risk of drug-induced death was also highest in this cluster (OR = 35.5; CI 25.6-49.3) and, for 68% (CI 64-73) of cases, filled prescriptions for OA was indicated as the underlying cause of death. The cluster analysis identified a subgroup with filled prescriptions of OA and other pharmaceuticals and a higher socioeconomic status than other subgroups. This subgroup had a high risk of drug-induced death that needs to be addressed.