The use of home ovulation testing kits in donor insemination (DI) has been proposed to increase patient and clinic convenience while not compromising fecundity rates. Such a system was introduced into our DI service in December 1994, and we here report an audit of experience over 6 months. Patients were offered home or laboratory luteinizing hormone (LH) testing, and those requesting home testing were asked to store an aliquot of tested urine for subsequent assay in the laboratory allowing retrospective analysis of the accuracy of cycle timing. Insemination using cryopreserved semen was performed on the day home testing predicted ovulation, or on the day an LH surge was detected in the laboratory, and on the following day. Pregnancy rates were significantly reduced in home testers: 3.4% per cycle (174 cycles, 64 women) versus 12.7% (110 cycles, 53 women) over the same time period (P < 0.005, 95% confidence interval 6.5-18.9). Urine samples from 140 cycles from 51 women using home testing were analysed. There were insufficient data in nine to allocate the cycle. Of home tested cycles, 37 (28%) were inseminated on a day other than the first day of the LH surge. In 13 of these insemination was performed after the first day of the LH surge. Incorrect treatment was associated with high baseline LH, but those with 'late' treatment had low basal LH concentrations, similar to those correctly treated. Analysis of individual urine samples showed that the positive predictive value of home testing was 72%. These results suggest that home ovulation testing results in reduced chance of pregnancy, with increased frustration for both patients and clinic staff. This may be particularly so in women with high baseline LH concentrations.