Year: 2004
Type: article
Abstract: The costs of alcohol use to society are well known and numerous, including material and personal losses. Costs to social service and health-care systems are important to estimate, but the published literature on this topic is contradictory. Many studies have found associations between alcohol use and utilization or cost of health services. In the case of out-patient utilization, an inverse relationship has consistently been found. However, the magnitude and shape of the relationship to in-patient care has not been consistent. The following question has not been answered conclusively: ‘Is there a relationship between alcohol consumption and inpatient medical care utilization and cost and, if so, what is the shape of the relationship? As Anzai et al. (2005) point out, various investigators have characterized the alcohol–in-patient utilization association as linearly positive, linearly inverse, J-shaped, inverse-J-shaped and U-shaped. On the surface, these results are contradictory. However, is it reasonable to compare these studies? I maintain that comparison is hindered by the inconsistent definitions of drinking categories and the variability of drink size. For example, in Anzai et al. (2005), a U-shaped association was observed, using five categories of drinkers defined in terms of grams per week: 0 (life-long abstainers), 1–149, 150–299, 300–449 and ≥450. These categories are equivalent to approximately 0, <11, 11–21, 21–32 and >32 standard drinks per week, given the usual definition of a standard drink of about 14 g of ethanol. On the other hand, Armstrong et al. (1998) observed a somewhat inverse linear association for men, but we utilized four categories of drinkers, equivalent to 0 (current abstainer), <7, 7–13 and ≥14 drinks per week. Thus, the shapes of the associations found between alcohol consumption categories and utilization for these two studies is based on totally different alcohol categories and therefore the comparison is flawed. In addition, in the Armstrong study, the abstainer category includes ex-drinkers and in the Anzai study it does not. In another article cited by Anzai (Armstrong & Klatsky 1989), eight drinking categories were used, based on a question from a self-administered questionnaire which asked about number of drinks consumed in a specified time period (day or month), but not how big the drinks were nor how many days per week or month the individual drank. There is a reasonably large literature on standard drink size, including research that has shown that people have a tendency to underestimate the size of their drinks in terms of number of standard drinks, thus underestimating the amount of alcohol they consume. Gender is another important factor to consider, as the association appears to differ by gender. Is there a ‘right’ way to analyse the association between alcohol consumption and utilization? There may not be a right way, but there are certainly better ways that facilitate understanding and interpretation of multiple studies. Using gender-specific, standardized definitions of drinking categories, such as those for ‘light’, ‘moderate’, and ‘heavy’ drinking as defined by the CDC and NIAAA, separating life-long abstainers from ex-drinkers, stratifying analyses by gender, and specifying the quantity of alcohol consumed in terms of standard drinks of each beverage type would permit more precise comparisons of studies.
Source: Addiction
Author: MARY ANNE ARMSTRONG
Institution: Kaiser Permanente
Language: English
Subfield: Pathology and Forensic Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W1599203726