Estimation of the Coronary Risk

The two traditional components of epidemi­ological research (descriptive/analytical) contribute to improve our knowledge of dis­eases. Analysis of incidence and mortality data, which provides a description of a given disease at the population level, and knowledge of general characteristics (age, sex, transmission group, etc.) that are asso­ciated with a higher risk, also contribute towards identifying areas for improvement in access to care and generate etiologic assumptions.

However, epidemiological studies have important implications for prevention, early detection, diagnosis, and access to care. The importance of a phenomenon can be charac­terized by the use of crude numbers (fre­quency) or rates and ratios. The numerator is included in the denominator of the inci­dence or prevalence rate; however, rates, as opposed to frequencies, imply an element of time. With incidence rates, for example, only new cases of the disease occurring dur­ing a defined time period (e.g., 1 year) are taken into account, being divided by the average size of the population exposed dur­ing the same period.

To quantify the association between a disease and a risk factor, one generally uses relative risk (RR), R1/R0, where R1 is the risk of disease or death among the popula­tion exposed to the risk factor and R0 is the risk among unexposed subjects. Usually, RR above 1 denotes a deleterious effect of the risk factor, while an RR below 1 suggests a beneficial effect. An RR of 1 suggests that there is no correlation. Another index is the odds ratio (OR), defined as [R1/(1-R1))/ (R0/(1-R0)].

When comparing morbidity rates in a cohort with those of the general population, for example, direct comparison of crude rates can lead to erroneous conclusions. Indeed, it is known, for example, that the risk of almost all diseases, and particularly cardiovascular diseases, increases with age. If the age distribution is different between the two populations, the comparison of the risk will be affected by this confounding bias. Standardization appears the best way to avoid this kind of bias, and is based on the crude disease rate that would be observed in the cohort if its age distribution were the same as that of the comparator population. Thus, the standardized morbidi­ty or mortality ratio compares the observed number of cases of disease or death in the cohort with an expected number of cases. The expected number is calculated by (1) classifying the study group in terms of demographic variables such as age and sex; (2) computing the expected number of cases or deaths for each class (by multiplying the number of individuals in the study group in that class by the class-specific death rate in a standard reference population); and (3) adding together the expected cases or deaths in all classes.