Background Leading causes of death for drug-treatment clients across Scotland, 1996–2006, were drug-related (1383 DRDs) and non-drug-related suicides (269). We investigate DRD-risk by time since most recent hospital stay.

Methods Drug-treatment records were linked to national registers of deaths, hepatitis C virus (HCV) diagnoses, and hospital/psychiatric episodes. We calculated DRD-rates (and suicide-rates): during hospitalisation, within 28 days, 29–90 days, 91 days -1 year, >1 year since discharge from most recent hospital stay vs never admitted. Proportional hazards analysis adjusted for demographic and other time-specific influences on DRD-risk.

Results The cohort comprised 69 457 individuals, 350 317 person-years (pys) and 90 314 hospital-stays. DRD-rate per 1000 person-years (pys) was: 87 (95% CI 72 to 103) during hospitalisation, 21 (18 to 25) within 28 days, 12 (10 to 15) during 29–90 days and 8.5 (7.5 to 9.5) during 91 days to 1 year after discharge vs 4.2 (3.7 to 4.7) when >1 year after most recent hospitalisation and 1.9 (1.7–2.1) for those never admitted. Adjusted HRs by time since hospital-discharge (vs never admitted) were: 10 (95% CI 8 to 12) within 28 days, 5.6 (4.6 to 6.8) during 29–90 days, and 4.0 (3.5 to 4.7) vs 2.3 (2.0 to 2.7) when >1 year after most recent hospital stay. Alcohol misuse increased HR (1.5, 1.3 to 1.7) and female, never injector, and no HCV diagnosis decreased it: 0.56 (0.49 to 0.64), 0.62 (0.52 to 0.73), 0.74 (0.65 to 0.85).

Conclusions Hospital discharge marks high DRD-risk periods. Doctors should consider prescribing Naloxone when discharging patients with opiate-dependency, and emailing discharge summary to alert the patients' general practitioner or drug treatment agency.