Sociodemographic variation in live birth incidence and access to assisted reproductive services among female cancer patients in the United States.

Abstract

1640 Background: Sociodemographic variation in fertility outcomes after cancer is not well described. We constructed a novel longitudinal dataset to capture sociodemographic variation in 1) live births and 2) use of assisted reproductive technology (ART) among female cancer patients across the US from 2004-2022. Methods: We linked population-based cancer registry data (2004–2018 in 12 states; 2004–2011 in California) with live birth certificates (2004–2022) and the Society for Assisted Reproductive Technology (2004-2022) database, which captures ~90% of ART cycles nationally, to create a dataset integrating oncologic characteristics, obstetric outcomes, and ART use. Patients aged 15-45 at diagnosis of cancer were included. The primary outcome was 5-year cumulative live birth incidence (CLBI) using the Fine and Gray method to account for all-cause death as a competing event and log-rank tests to compare by age, race/ethnicity, insurance, cancer site, cancer stage, and receipt of systemic therapy. Secondary outcome was use of ART after cancer diagnosis, defined as >1 autologous oocyte/embryo cryopreservation or embryo transfer cycles. Patients were categorized by expected fertility detriment, defined as pelvic malignancy and/or receipt of chemotherapy, hormone therapy, or abdominal or pelvic radiation, and covariates compared descriptively. Person-years were accrued from diagnosis date to first post-diagnosis birth/use of ART, death, age 51 (oldest age a post-cancer birth was observed), or December 31, 2022, whichever occurred first. The study was approved by the institutional review board. Results: 286,198 female cancer patients were included with 15 live births per 1,000 person-years. 5-year CLBI was 6.14% [6.05, 6.24]. Lower CLBI was observed in Hispanic (4.5% [4.31, 4.76]) and non-Hispanic Black (NHB) (5.23% [4.98, 5.50]) populations compared to non-Hispanic white (NHW) populations (6.75% [6.62, 6.88]) (p<0.001). Patients with public insurance also had significantly decreased CLBI (4.68% [4.47, 4.90]) than those with private insurance (6.08% [5.95, 6.20], p<0.001). NHB, Hispanic, and patients with public insurance comprise higher proportions of patients with expected fertility detriment versus those without detriment (12.6 vs 8.6%, 18.2 vs 13.1%, and 18.7 vs 13.3%, p<0.001) but lower proportion of patients that use ART versus those that do not (5.9 vs 11.4%, 7.6 vs 16.6%, and 4.1 vs 17.1%, p<0.001). Conclusions: National vital statistics data demonstrates high fertility rates among Hispanic and NHB populations; among cancer patients, however, these trends are flipped, with lower cumulative incidence of live births in these groups. Despite a higher burden of anticipated fertility detriment, NHB, Hispanic, and publicly insured patients were less likely to access fertility care, revealing stark inequities in survivorship care.