Randomized response technique for estimating alcohol use and noncompliance in hypertensives.

Publisher Website
Google Scholar
Abstract
The randomized response technique (RRT) was used in a 2-phase investigation of underreporting on use of alcohol and noncompliance behavior in hypertensive patients. In phase I, the proportion of patients who reported daily use of alcohol and the proportion at least 1 day of noncompliance in the preceding week were estimated by RRT and by direct response. The proportion of respondents who reported daily use of alcoholic beverages during the preceding week was 26% according to RRT vs. 20% by direct response, and the proportion of patients who reported at least 1 day of noncompliance during the preceding week was 23% by the RRT vs. direct response. In phase II, the mean number of alcoholic drinks consumed/week and the mean score on an index of compliance behavior were estimated by RRT and by direct response. The estimated mean number of drinks/week was significantly higher according to the RRT compared with the direct response estimate (8.7 vs. 3.9). Significant underreporting on use of alcohol was found among women, older respondents (> 52 yr), and more highly-educated respondents (> 12 yr of school). There was no difference in the estimated mean scores for compliance. Feasibility of the RRT for use with hypertensive patients was demonstrated, and the procedure is suggested as a tool for the study of sensitive behaviors in other patient populations.

This publication has 4 references indexed in Scilit: