Feasibility of studying subfertility using retrospective self reports.

During an investigation of possible reproductive effects of environmental agents, 261 male and 155 female workers were interviewed concerning subfertility at some time in the past: the time taken to conceive, for all births; and the occurrence of one or more fertile phases lasting for 6 months or more. When these two variables were compared, the quality of reporting was acceptable in 89.7% of instances, and data editing enabled accuracy to be improved. Reporting was more reliable with shorter duration of recall, and female workers' reports were somewhat more reliable than those of male workers. The distribution of time taken to conceive was similar for male workers to that observed in previously published prospective series, though with a higher estimate of subfertility when infertile phases were also considered. Comparison with published estimates of reduced fertility appeared to be reassuring. As predicted, the equivalent comparisons for female workers showed the presence of a strong selection effect.