Estimating the total mortality among problem drug users

This paper's objective is to develop a method to estimate the total mortality among problem drug users. The total mortality is given by a base rate of mortality not related to drugs and the deaths that are directly and indirectly related to drugs. A fatal poisoning by drugs (overdose) is directly related to drugs, whereas a casualty due to a drug-related disease or a drug-related accident is indirectly related to drugs. As an example of a method to estimate the total mortality, the results from a cohort study among methadone patients in Amsterdam were projected on the whole population of problem drug users in The Netherlands. Due to differences between the problem drug users in Amsterdam and the rest of the country, adjustments were required. It was found that an initial estimation did not require adjustment for injection behavior and gender but did require adjustment for age and the percentage of HIV infection. In a first unadjusted estimation, the total number of deaths among problem drug users in The Netherlands in 2001 was estimated at 606 deaths. After adjustment for age, the estimated mortality decreased to 573 deaths, and after adjustment for HIV infection, this estimation again decreased to 479 deaths. From the ultimately estimated mortality, 11% was considered to be not related to drugs, 23% was attributed directly to drugs, and 66% was attributed indirectly to drugs. The number of direct deaths, as estimated by this method, falls in the same order of magnitude as the number extracted from the Causes of Death Statistics, when selecting cases according to the Drug-Related Deaths Standard as established by the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA). Further cross-validation with other measures will be needed to assess the accuracy of the method, the limitations of which are discussed with respect to stipulating directions for future research.