Objective : The Pharmacy and Therapeutic Committee generally selects formulary drugs by evaluating the comparative data from non-clinical and clinical studies of pharmacologically similar type drugs. Usually there are few data from directly comparative studies among all similar type drugs, and there are only two types of studies, standard drug or placebo-control study. In this situation, it is very difficult to select the superior drug.
Therefore, we conducted a retrospective cohort study to compare the efficiency among three 5-HT₃ antagonists and to investigate the possibility of data from this study, as a reference for decision making.
Design : Retrospective cohort study.
Methods : A retrospective cohort was assembled comprising in-patients of to Surgery, Obstetrics and Gynecology, Ophthalmology, Orthopedics, Dermatology, Psychiatry, or Thoracic Surgery Departments, who received chemotherapy and a 5-HT₃ receptor antagonist from January to June in 1996. We compared the side effects of nausea and vomiting among patients receiving various 5-HT₃ receptor antagonists. Physicians could freely prescribe three 5-HT₃ antagonists that were adopted in our hospital during the period. Pharmacists collected information regarding the efficacy of 5-HT₃ receptor antagonists on the first prescribed day by reviewing medical charts, nurses' reports and interviews with the nurses.
Statistical Analysis : ANOVA was used for testing the age differences among patients receiving 5-HT₃ receptor antagonists. Chi-square was used for testing difference of sex, administration route, number of chemotherapy treatment. Kruskal-Wallis rank test was used for testing the difference of efficacy. Probabilities less than 5% (P<0.05) were considered significant.
Results : Out of 68 patients, 35% received cisplatin and 65% received non-cisplatin anti-tumor drugs. The average age was 53.1±17.4 (49% male, 51% female). A total of 82% received intravenous drip injection.
Percentage of non-changed cases in ingestion ratio was 32% of ramosetron, 50% of ondansetron and 53% of azasetron. There was no significant difference in the ingestion ratio (p = 0.026).
Percentage of non-changed cases with regard to vomiting (complete control) was 80% of ramosetron, 50% of ondansetron, 79% of azasetron. There was a significant difference in vomiting (completecontrol) (p = 0.044). Ondansetron was less effective as an antiemetic function.
Discussion : When no data from a directly comparative study are available to determine the adoption of a new medicine from among numerous candidate drugs, an observational study such as this appears useful.