CORRELATIONS BETWEEN PRESCRIPTIONS AND DRUGS TAKEN IN SELF-POISONING - IMPLICATIONS FOR PRESCRIBERS ACID DRUG REGULATION

Objective: To compare prescription data for Australia with drugs inges ted in self-poisoning and suicide, to determine which drugs are over-r epresented. Design: Comparison of data on drugs taken in self-poisonin g admissions and suicides with Australian prescription drug dispensing data from the Drug Utilization SubCommittee (DUSC). Setting: Newcastl e and Lake Macquarie, Australia, 1989-1992. Subjects: Between July 198 9 and June 1992, 1159 prescription drugs were taken in overdose. Eight y-three drug-related deaths were investigated by the coroner between 1 987 and 1992. On 48 occasions a prescription drug was the primary caus e of death. Results: Drugs over-represented in self-poisoning (relativ e to Australian prescriptions) included not only those prescribed for psychiatric conditions (antidepressants, neuroleptics and lithium), bu t also benzodiazepines, barbiturates and other anticonvulsants. The hi ghest odds ratios for death when adjusted for prescription numbers wer e for short-acting barbiturates (523.7; 95% confidence interval [CI], 207-1322), chloral hydrate (58.1; 95% CI, 18.1-187), colchicine (27.9; 95% CI, 3.8-202), dextropropoxyphene (20.8; 95% CI, 8.8-48.9), tricyc lic antidepressants (13.3; 95% CI, 7.2-24.5) and anticonvulsants (11.6 ; 95% CI, 4.1-32.2). Conclusions: Short-acting barbiturates, chloral h ydrate and dextropropoxyphene have little or no clinical advantage ove r alternatives and excessive toxicity in overdose. They should be remo ved from the market. The toxicity of anticonvulsants and colchicine sh ould be considered when they are prescribed, and smaller amounts per p rescription may be advisable for high risk patients using these and ot her toxic drugs.