Prescriptions of Benzodiazepines in a University Teaching Hospital : Patient-based Data Analysis Using a Computer Ordering System

Objectives: Benzodiazepines are remarkably useful drugs for psychological illnesses. However, because of dependency and withdrawal symptoms, it is difficult for the patients to stop taking the medication once it is started. The purpose of the present study was to determine the tendency of benzodiazepine prescriptions at a university teaching hospital setting and clinical characteristics of patients in long-term prescription for benzodiazepines. Subjects: This cross-sectional study was performed by using data from the database of a computer ordering system at our university hospital. Subjects were 34422 outpatients (female ratio: 52%) who visited the university hospital and were prescribed some kind of drugs between July 2002 and June 2003. Method: The prevalence of patients receiving benzodiazepines was estimated and they were further categorized into two groups (a long-term group prescribed 4 months or longer, and a short-term group prescribed 3 months or shorter). Clinical characteristics were compared between the two groups including patients' sex and age, department clinics visited ('internal medicine', 'surgery', 'psychiatry', and 'others'), kind of benzodiazepines (anxiolytics and hypnotics), and their pharmacological half-life ('ultra-short', 'short', 'intermediate', 'long', and 'ultra-long'). Statistical analyses were performed based on chi-square test. In the long-term group, the patterns of prescription among department clinics from the perspective of pharmacological half-life were additionally investigated in anxiolytics and hypnotics. Results: Among the 34422 patients, 5959 patients (17%) were prescribed benzodiazepines during the study period: 4135 receiving anxiolytics and 3209 receiving hypnotics. The long-term group included 4470 patients (female 58%, mean age 59±17 years) and the short-term group included 1489 patients (female 61%, mean age 50±18 years). Compared to the short-term group, the proportion of male patients was higher (p=0.025) and patients' age was older (p<0.0001) in the long-term group. With regard to department clinics visited, patients prescribed at 'internal medicine' department and 'other' groups were more likely to have long-term prescription, compared with 'psychiatry' department group. Anxiolytics were found to be more frequently prescribed in the short-term group whereas hypnotics were more prescribed in the long-term group regardless of pharmacological half-life. In the long-term group, those in the 'psychiatry' group were more likely to prescribe 'intermediate' half-life agent, while those in the 'internal medicine', 'surgery', and 'other' groups were more likely to use short agent (p<0.0001 in both anxiolytics and hypnotics). Conclusion: Long-term use of benzodiazepines causes adverse effects which lead to serious health problem in the elderly. The future implications of the present study are that physicians should prescribe benzodiazepines based on appropriate mental assessment and that such rational prescription should be included in educational program at a university hospital setting and needs to be carefully monitored.