Background: In New Zealand, public funding for assisted reproductive technology (ART) is restricted to subfertile women who are unlikely to conceive spontaneously, based on clinical and social criteria known as the clinical priority access criteria (CPAC) score. The objective of this study was to compare this CPAC score with a prediction model for predicting spontaneous conception, developed in the Netherlands (the Hunault model).
Methods: We performed a cohort study and included couples with unexplained subfertility and assessed the measure of agreement and the performance of the CPAC score and the Hunault prediction score.
Results: Of 663 couples referred, 249 (38%) couples had unexplained subfertility. Of 246 women with full follow-up data, there were 143 women (58%) who had a live birth during the follow-up period, 65 (26%) after fertility treatment and 78 (32%) after natural conception. There were 100 couples (41%) who had a Hunault prediction score of <30%, which is the Dutch treatment threshold, and 36 couples (15%) who had a CPAC score of >65, which is the New Zealand threshold for publically funded treatment. There were 69 couples (28%) who meet the threshold for treatment in the Netherlands but did not meet the New Zealand threshold for public funding. The kappa coefficient as a measure of agreement of the two scores and their treatment thresholds was 0.30, suggesting a fair agreement. The area under the curve for the CPAC and Hunault scores were both 0.63, but the Hunault model performed better in calibration.
Conclusions: The CPAC score correlates fairly with the Hunault prediction score, although using the Hunault prediction model 26% more couples would be recommended for ART. The discriminative capacities of both scores were comparable, but the Hunault prediction score performed better in calibration. Funding models in New Zealand should consider treating those women with unexplained subfertility who are least likely to conceive spontaneously.