Do quantity-frequency data underestimate drinking-related health risks?

Identifying health impairment related to ethanol consumption is one of the major objectives of public health research. The most frequently used method for assessing drinking behavior in public health surveys and related research has been estimation formulae, like the Quantity-Frequency (QF) method which derives an estimate of typical/average levels of daily consumption. In recent years, questions have arisen as to whether the QF method can accurately reflect actual drinking patterns. This study compares a QF method of assessing daily drinking behavior with a newer, more quantitative method (Time-Line, TL) of assessing daily drinking. The QF and TL methods yielded similar mean daily ethanol consumption levels; however, in contrast to the TL method, the QF method seriously masked subjects' actual drinking patterns by failing to identify certain types of ethanol consumption days, especially those thought to be associated with health risks. These findings, while provocative, were obtained with a small number of subjects (N = 40). Extrapolation to populations other than problem drinkers, while likely, awaits further empirical validation.