Investigating the relationship between beta-blocker and antidepressant use through linkage of the administrative databases of Saskatchewan Health.

The main objectives of this research were to confirm previous reports of an increased risk of antidepressant use subsequent to propranolol therapy, and to determine whether this higher relative risk (RR) is indicative of depressive symptoms as a side-effect of the drug's use. 'New' users of beta-blockers, other antihypertensives or diuretics in two separate years were identified from records of Saskatchewan Health. Incidence of concurrent antidepressants prescribing was determined in the year after initiating therapy with a cohort drug. Medical claims were examined to identify physician services and diagnoses associated with cohort drug and antidepressant prescriptions. Risk of concurrent antidepressant use in new propanolol users aged 20-39 years was approximately double that of new diuretic users [RR 2.2 (1.5-3.3) in 1984 and 2.0 (1.1-3.5) in 1990/91]. When cases with a diagnosis of migraine headache were excluded, the risk of concurrent propranolol/antidepressant use was age- and sex-dependent, but not consistent for the two study years. It was concluded that although the risk of concurrent antidepressant use was greater in younger propranolol users, the risk cannot be solely attributed to depressive side-effects of the drug. The purposes of this paper were to present the methods employed in linking the databases of Saskatchewan Health, describe the results of the analysis, and to highlight the methodologic problems that arose.