T W Tam1, L T Midanik. 1. Alcohol Research Group, Public Health Institute, Berkeley, California 94709-2176, USA.
Abstract
OBJECTIVE: Screener items are often used in surveys to identify individuals who are at high risk of experiencing alcohol dependence or alcohol-related problems. The purpose of this study is to determine the effect of two screener items (5 or more drinks ever and > or = 12 drinks last year) on prevalence estimates of alcohol dependence and negative social consequences. METHOD: The 1995 National Alcohol Survey data were analyzed. Differences in prevalence estimates between the unscreened current drinkers (n = 2,817) and the two screened samples: 5+ ever (n = 2,186) and > or = 12 drinks last year (n = 2,126) were compared. For each screened sample, prevalence estimates obtained from two base populations (screened positives only and all current drinkers assuming no alcohol-related problems among screened negatives) were examined. RESULTS: Comparisons of prevalence estimates of alcohol dependence and negative social consequences across the screened and unscreened samples revealed little bias in estimates with one exception. The only significant difference between the unscreened and screened samples was found among women when prevalence rates were obtained from data of screened positive respondents only. CONCLUSIONS: The effect of screening on national prevalence estimates of alcohol dependence and social consequences is small. Less bias in prevalence estimates is found when negatively screened light drinkers were assumed to report no alcohol-related problems rather than excluded from the estimation sample.
OBJECTIVE: Screener items are often used in surveys to identify individuals who are at high risk of experiencing alcohol dependence or alcohol-related problems. The purpose of this study is to determine the effect of two screener items (5 or more drinks ever and > or = 12 drinks last year) on prevalence estimates of alcohol dependence and negative social consequences. METHOD: The 1995 National Alcohol Survey data were analyzed. Differences in prevalence estimates between the unscreened current drinkers (n = 2,817) and the two screened samples: 5+ ever (n = 2,186) and > or = 12 drinks last year (n = 2,126) were compared. For each screened sample, prevalence estimates obtained from two base populations (screened positives only and all current drinkers assuming no alcohol-related problems among screened negatives) were examined. RESULTS: Comparisons of prevalence estimates of alcohol dependence and negative social consequences across the screened and unscreened samples revealed little bias in estimates with one exception. The only significant difference between the unscreened and screened samples was found among women when prevalence rates were obtained from data of screened positive respondents only. CONCLUSIONS: The effect of screening on national prevalence estimates of alcohol dependence and social consequences is small. Less bias in prevalence estimates is found when negatively screened light drinkers were assumed to report no alcohol-related problems rather than excluded from the estimation sample.