Optimizing the control group for evaluating ART outcomes: can outpatient claims data yield a better control group?

Purpose: We previously developed a subfertile comparison group with which to compare outcomes of assisted reproductive technology (ART) treatment. In this study, we evaluated whether insurance claims data in the Massachusetts All Payers Claims Database (APCD) defined a more appropriate comparison group. Methods: We used Massachusetts vital records of women who delivered between 2013 and 2017 on whom APCD data were available. ART deliveries were those linked to a national ART database. Deliveries were subfertile if fertility treatment was marked on the birth certificate, had prior hospitalization with ICD code for infertility, or prior fertility treatment. An infertile group included women with an APCD outpatient or inpatient ICD 9/10 infertility code prior to delivery. Fertile deliveries were none of the above. Demographics, health risks, and obstetric outcomes were compared among groups. Multivariable generalized estimating equations were used to calculate adjusted relative risk (aRR) and 95% confidence intervals (CI). Results: There were 70,726 fertile, 4,763 subfertile, 11,970 infertile, and 7,689 ART-treated deliveries. Only 3,297 deliveries were identified as both subfertile and infertile. Both subfertile and infertile were older, and had more education, chronic hypertension, and diabetes than the fertile group and less than the ART-treated group. Prematurity (aRR = 1.15–1.17) and birthweight (aRR = 1.10–1.21) were increased in all groups compared with the fertile group. Conclusion: Although the APCD allowed identification of more women than the previously defined subfertile categorization and allowed us to remove previously unidentified infertile women from the fertile group, it is not clear that it offered a clinically significantly improved comparison group.