A similar male to female distribution was observed in the two drug groups, 32.6% to 67.4% for alprazolam and 33.1% to 66.9% for lorazepam. The lorazepam patients were slightly older than the alprazolam patients, with 25.9% and 20.9%, respectively, being 65 years of age or older.

Only about a third of both groups of patients continued treatment into a second 30-day period and, by the end of the 12th period, just 5% of the alprazolam patients and 7% of the lorazepam patients remained on treatment. Although the distributions of patients continuing with therapy varied little when calculated separately for the two sexes or by age group, for both drugs, females were more likely than males to continue taking the product, and the elderly were more likely to continue than younger patients.

Tables 1 and 2 give the numbers and rates of diagnoses and symptoms recorded in the physician service claims datafile for the alprazolam and lorazepam patients during the first 30-day period after starting their respective treatment. The numbers and rates recorded during the succeeding six 30-day periods are also shown, together with the excess rates during the first period over those for the succeeding six periods and 99% CIs for these rates.

TABLE 1 Rates of diagnostic events per 1000 patients experienced by the alprazolam patients recorded in the physician service claims datafile

Table1An acute adverse event signalling scheme usinge

 *P<0.01

 TABLE 2 Rates of diagnostic events per 1000 patients experienced by the lorazepam patients recorded in the physician service claims datafile

Table2An acute adverse event signalling scheme usinge

P<0.01 

Neurotic, depressive and sleep disorders stand out in Tables 1 and 2 as occurring significantly more often during the first 30-day period than in the subsequent six periods (identified by a lower 99% CI limit greater than zero). Because the data were supplied in aggregate form, it was not possible to relate an individual prescription to a physician service for the same patient that may have generated the prescription and, consequently, claims for services that may have resulted in alprazolam or lorazepam being prescribed could not be identified. However, the differences for these almost certainly point to the main indications for the treatment.