Alcohol consumption in Russia: is monitoring health conditions in the Russian Federation (RLMS) trustworthy?

Abstract

Sir—I have written previously (Nemtsov 2003) about the quality of alcohol consumption indicators in the Russian Longitudinal Monitoring Survey, 1992–2001, Health Conditions in the Russian Federation (RLMS). The above many-sided survey is based on interviews with members of more than 4000 households in many regions of the country (Zohoori et al. 1996, 2002). An examination showed that the survey's mean figures are unreliable estimates of the actual alcohol consumption in Russia. In some years they were lower than those of official statistics, which failed to account for illicit alcohol consumption (4.8 vv. 5.0 l; 1992). Furthermore, the RLMS estimate is lower than two other estimates of the actual consumption in 1992, namely 13.81 l (Treml 1997) and 13.23 l (Nemtsov 2000). To understand the trustworthiness of the RLMS mean alcohol consumption indicators it is important to note that in 1992 the alcohol-dependent variables (alcoholic psychoses, alcoholic poisoning mortality and others) were close to the 1984 level, when official consumption figures of the State Statistics Committee of the Russian Federation was 10.5 l per annum, and the mean estimation from three different sources was 14.2 l per capita (Nemtsov 2000). Still less reliable are the raw data presented on the RLMS site (http://www.cpc.unc.edu/rlms/home.html), which have already been cited as a source in a number of publications (e.g. Cockerhan 2000; Carlson, 2001; Cockerhan et al. 2002). To bear out the above, let us examine the main element, that is, the table of individual alcohol consumption (Round X, 2001), which presents the data of a poll of 10 078 people. The main result of this part of the survey presented by variable I7 (intake per day) claims that 42% of the subjects were absolute teetotallers. This is a substantially higher figure than those in European countries, ranging from 4% in Denmark to 36% in Spain. Indeed, the claim is considerably higher than the number of abstainers recorded in Moscow in 1994 (9%; Simpura et al. 1997). If RLMS data were trustworthy one could be proud of the sobriety of the Russian population. Living in Russia, however, it is hard to overlook the surrounding drunkenness. Hard, also, to believe the RLMS data in face of the official statistics of the consequences of alcohol consumption in the country: in 2001 as many as 41 091 people died of alcohol poisoning and 161 442 suffered from alcoholic psychoses (28.5 and 111.9, respectively, per 100 000 population). Some of the gross RLMS mistakes stem from reasons common in any interview (the psychology of the respondents, forgetfulness, and mechanisms of defence and compensation are possibly more distinct among the Russian population). RLMS, however, has the source of the mistakes of a particular nature as well. Above all, they derive from a formal and uncritical treatment of the results of the survey.