The reliability of personal alcohol consumption estimates in a working population.

Health and safety problems related to alcohol consumption represent a major concern in many businesses and consequently pre-employment questionnaires and workplace health promotion packages frequently contain questions seeking basic estimates of alcohol consumption. Although individuals with very heavy drinking patterns often attract much attention, on a population basis most morbidity is likely to arise from drinkers consuming above the recommended limits but not sufficient to result in gross occupational or social effects. This study reviews the research on the reliability of questionnaire techniques used to quantify alcohol consumption and compares the most valid retrospective interview based estimate of consumption, the time line follow-back (TLFB), with a prospective daily diary (DD) method in a working population. The DD method was acceptable and produced significantly higher estimates of consumption and for 'normal' subjects gave consistent estimates of +/- 10 units/week over several weeks. Appropriate questionnaire design is discussed and the use of a DD method in health promotion activities recommended.