Standardized mortality ratio

From Wikipedia, the free encyclopedia

The standardized mortality ratio or SMR in epidemiology is the ratio of observed deaths to expected deaths according to a specific health outcome in a population and serves as an indirect means of adjusting a rate. The figure for observed deaths is usually obtained for a particular sample of a population. The figure for expected deaths reflects the number of deaths for the larger population from which the study sample has been taken e.g. national level of mortality attributed to a particular health outcome. The calculation used to determine the SMR is simply: number of observed deaths/number of expected deaths.

The SMR may be quoted as either a ratio or, sometimes, a percentage. If the SMR is quoted as a ratio and is equal to 1.0, then this means the number of observed deaths equals that of expected cases. If higher than 1.0, then there is a higher number of deaths than is expected.

An example might be a cohort study into cumulative exposure to arsenic from drinking water, whereby the mortality rates due to a number of cancers in a highly exposed group (which drinks water with a mean arsenic concentration of, say 10mg) is compared with those in the general population. An SMR for bladder cancer of 1.70 in the exposed group would mean that there is 70% more cases of death due to bladder cancer in the cohort than in the reference population (in this case the national population, which is generally considered not to exhibit cumulative exposure to high arsenic levels).

The SMR may well be quoted with an indication of the uncertainty associated with its estimation, such as a confidence interval (CI) or p value, which allows it to be interpreted in terms of statistical significance. As a general convention, epidemiologists work at a confidence level of 95%. Therefore, if our SMR of 1.70 for bladder cancer among those exposed to arsenic in drinking water was quoted as 1.70 (1.20 - 2.18), then the CI in brackets tells us the there if we repeat the sampling in the sample population, only 5% of those repetitions might result in an SMR more extreme than the confidence limits. If the SMR of 1.70 in our example was quoted as 1.70 (0.90 - 2.18), then the CI includes 1.0. This means that at the 95% level of confidence, we cannot be certain that our result is different from 1.0 (i.e. it is no different from the reference population).