Methodological considerations in the study of frequency, risk factors and outcome of reduced fertility.

Infertility is a major health problem for many couples. Many clinicians believe that its prevalence is increasing. This may be attributable to greater concern about infertility in society and development of new treatment methods. Infertility is also a complex medical, social and individual problem. We do not know enough about the risk factors and their mechanisms of action. In the study of reduced fertility there is a need for clear definitions and careful selection of study populations. There are several sources of bias. One is selection bias in the ascertainment of infertility couples, others are recall bias and exposure assessment bias. Furthermore, medical classification of causes is difficult in population based studies. This paper will discuss problems in the interpretation of results of past studies and in methods already in use. Some new design options in future research are presented.