1 - Fertility and Infertility Statistics: Their Importance and Application

Abstract

To show clearly that an infertility treatment was responsible for the changes leading to pregnancy, comparison needs to be made with defined populations, both of normal subjects and of untreated patients. The comparisons should be made clearly and statistically evaluated. Attention in this discussion of fertility and infertility statistics, their importance and application, is directed to the following: defining the background population; diagnosis of cause of infertility; evaluation of results in infertility patients; some studies illustrating the use of life table analysis; and the design of a randomized control trial. Vessey et al. (1978) calculated the rate for outcome of pregnancy in parous women after stopping barrier contraception and found, using life table analysis, that 36% were undelivered at 12 months. It is possible to calculate the monthly probability of conception or fecundability from cumulative data using the mathematical model described by Potter and Parker (1964). To evaluate the results of treatment, an adequate investigation of the couple is essential. Each of the degrees of impairment of normality in the male can be assessed only when paired with the normal female so that no understatement of fertility may be attributable to a female disorder. Similarly, to be assessable each female abnormality must occur alone in any 1 patient and the partner must be normal. It is only then that the impairment of fertility resulting from a single condition can be determined. In assessing the results in infertility patients, there are 2 main problems: some patients become pregnant and others are lost to follow-up, which progressively depletes the numbers under observation. When numbers in any 1 category are sufficiently large, an effort should be made to subdivide them according to severity. Examples of studies using life table analysis are presented. A table shows the cumulative conception rates in couples with primary infertility who had been investigated and in whom no abnormality was found. They were stratified according to years of infertility rather than from the time they were initially seen in the clinic. This method has been criticized as antedating the period of medical observation, but no treatment was offered. Had the patients been zeroed around the time first seen medically, the findings would have been weighted according to the duration of previous infertility. The apparent effect of treatment on conception rates is also shown.