expectancy at birth. The enormous reduction in the latter is due mostly to mortality increases at adult ages, as is the case in the period 1988-1992.
We now examine whether the most recent mortality increase can be attributed to a continuation of the alcohol-related increase of the period 1988-1992. Two reasons can be given for such an explanation: first, alcohol consumption increased between 1992 and 1993 (Table 8-1), and second, as noted above, cardiovascular mortality in 1992 was still lower than in 1984, before the anti-alcohol campaign.
To examine the question, we approximate the observed death rates by cause of death at ages 30-34 and 60-64 by a linear function of real alcohol consumption and calendar year (see Annex 8-2) and then compare the model estimates of age-specific death rates with the levels observed in 1993.
The trends of observed and model age-specific death rates for classes of causes of death are depicted in Figures 8-8 and 8-9. Clearly, the deviation (excess) observed from model death rates increases very substantially in 1993 as compared with the preceding years.
The approximate nature of the simple model applied here does not allow us