Objective: This project assessed the extent of agreement between drug-taking data obtained from an in-home assessment and that obtained from an automated outpatient pharmacy system and from a mail questionnaire.
Design: The underlying assumption is that an in-home assessment provides the most complete information (criterion standard) about the prescription medications being taken. The drug data from the mail questionnaire and automated outpatient pharmacies were compared with that from the in-home assessment by use of t tests and measures of sensitivity, specificity, and positive predictive value.
Setting: Kaiser Permanente, Northwest Region, a prepaid group practice HMO.
Participants: All elderly HMO members enrolled in the Social HMO demonstration project who had a comprehensive assessment within 30 days of returning a mail questionnaire between March 1986 and March 1988 (n = 83 frail elderly).
Main outcome measures: The number of medications currently being taken and the number of different therapeutic classes of drugs being taken per elderly HMO member.
Main results: The in-home assessment and the automated prescription system closely agreed on the number of different medications and the different therapeutic classes of prescription drugs being taken. The in-home assessment and mail questionnaire were in less agreement. Agreement was less for non-prescription drugs.
Conclusions: An automated prescription system is an adequate source of information about the prescription drugs currently being taken by frail elderly HMO enrollees with a prepaid prescription drug benefit.