The Jellinek Alcoholism Estimation Formula and its Application to Canadian Data

Publisher Website
Google Scholar
Abstract
The Jellinek formula represents an attempt to meet the need for a more versatile and economical method of alcoholism estimation than the field survey and one not subject to the severe deficiencies of most deductive statistical approaches. The assumption of a relationship between alcoholism prevalence and reported liver cirrhosis mortality, upon which the formula is based, was examined and concluded to be sound at least in the case of N. American and some other areas. However, the general reliability and methods of derivation of the factors in the formula were felt to be questionable and it was pointed out that their values might be expected to vary both in time and space. A comparison of estimates based on the Jellinek formula with others derived independently for the same populations indicated excellent or reasonably good agreement in 7 of 10 cases. In view of this evidence and other considerations, it was felt that Jellinek formula estimates for Canada or the U. S. based on data reported in recent years, would probably provide reasonably reliable indications of the magnitude of the problem of alcoholism in these countries. But it was concluded that application of the formula to N. American data for purposes of trend analysis, or to obtain estimates for years in which the liver cirrhosis mortality rate differed greatly from that attained in 1916, remained questionable. It was also suggested that the formula might overestimate in provinces or states where the true prevalence was lower than the national average and underestimate where it was higher. Attempts to estimate the values of the factors in the formula more directly were recommended. Some of the possible reasons for differences in the Jellinek formula estimates of the prevalence of alcoholism among the provinces of Canada were examined, including error variation, the proportion of drinkers in the population, personal income per capita, median age, rural-urban dis-tribution and comprehensiveness of reported liver cirrhosis mortality. It was not felt that the differences could be accounted for on the basis of error variation but that the degree of urbanization of the population was probably one of the most important factors.