In Vitro Fertilization Utilization Rates and Outcomes in States With and Without Insurance Coverage Mandates for Male Infertility Care.

We obtained nationwide insurance claims from over 91 million insured individuals from 2003 through 2021 from the Optum’s de-identified Clinformatics ® Data Mart Database (Optum CDM) (Optum, Inc, Eden Praire, MN) for US states. 13 Optum CDM provides a database of administrative medical claims, prescription claims, and insurance eligibility data for members of commercial and Medicare Advantage plans across the nation. Patients are enrolled in Optum CDM at the start of their insurance coverage, and their profiles are updated with administrative claims submitted by providers. Each patient is de-identified under the “expert determination method consistent with HIPAA and managed under data use agreements.” 13 Our study included US women aged 21 – 44 within Optum CDM from 2003 to 2020. We only used 2021 Optum CDM data as follow up data to assess outcomes from IVF cycles performed in 2020. We identified IVF cycles using the Current Procedure Terminology (CPT) code for oocyte retrievals (CPT 58970) and included oocyte retrieval claims that were at least 21 days apart. We only included claims which were completed for the first embryo transfer. We excluded oocyte retrievals with a diagnosis code for fertility preservation using the International Classification of Diseases (ICD) Ninth and Tenth Revision diagnosis codes, V26.42 / V26.82 and Z31.62 / Z31.84 respectively. To adequately capture patients’ medical conditions, we limited our cohort to patients with at least 1 year of continuous enrollment prior to their first oocyte retrieval. Gynecologic and infertility diagnoses were measured within 3 months prior to the first oocyte retrieval. Identification of comorbidities and gynecologic and infertility diagnoses were based on ICD codes ( Supplemental Tables 1 – 2 ). We then identified patients’ embryo transfers (CPT 58974) that occurred within 6 months of their oocyte retrieval and excluded mock embryo transfers using the CPT modifier 52 code, thus, effectively only assessing outcomes from a patient’s first embryo transfer. Our exposure of interest was the type of infertility coverage mandate in each state. States were categorized into three groups based on their laws as of 12/31/2020: Group 1 included states mandating IVF and male infertility care coverage; Group 2 included states mandating IVF coverage but not for male infertility care; and Group 3 included all remaining states ( Table 1 , Figure 1 ). Our primary outcome was IVF utilization rate, defined as the number of IVF cycles per 100,000 women aged 21 – 44 in Optum CDM in each group of states. Our secondary outcomes were fertility outcomes following IVF, including rates of pregnancy, live birth, multiple birth, and preterm birth after a woman’s first embryo transfer, in each group of states. Additionally, to measure the total number of children born from IVF in each group of states, we examined the number of live births from all embryo transfers per 100,000 women aged 21–44 in Optum CDM in each group of states. Fertility outcomes were identified among IVF cycles that had at least 6 months of continuous enrollment after the oocyte retrieval and an embryo transfer within 6 months after the oocyte retrieval. We assessed for a pregnancy starting from 7 days up to 100 days after the first embryo transfer date because approximately 85% of pregnancies were identified in this time window. We assessed for a live birth from 24 – 44 weeks and 6 days after the first day of the last menstrual period (LMP), which we estimated to be 19 days prior to the embryo transfer. We excluded deliveries less than 24 weeks because they are considered periviable by the American College of Obstetricians and Gynecologists. 14 Pregnancies, live births, and multiple births were identified using CPT, ICD, and Diagnosis Related Group (DRG) codes ( Supplemental Tables 3 – 8 ) according to our previously published methods. 15 Preterm births were defined as live births that occurred less than 37 weeks from the first date of the LMP. For each fertility outcome, observation time started on the day of the first embryo transfer and ended on the day of the outcome, a subsequent oocyte retrieval, a subsequent embryo transfer, loss of enrollment in Optum CDM, or the end of the assessment period for the outcome, whichever occurred first. We defined the number of live births as the maximum number of newborns based on either the live birth type described on the claims or the number of new newborns added to the mother’s Optum CDM family identifier in the two months around the birth. First, we compared women’s demographics at the time of the first oocyte retrieval, including age, race, comorbidities, and gynecologic and infertility diagnoses across state groups using chi-squared tests. For our primary outcome of utilization of IVF per 100,000 women 21–44 years old across state groups and 95% confidence intervals (CIs), we used a Poisson model adjusted for age and race, with an offset term for the population of each state, age, and race group. To compare IVF outcomes across state groups, we used cumulative incidence function regression models to estimate the adjusted cumulative incidences for pregnancy, live birth, multiple birth, and preterm birth with 95% CIs. For each model, we adjusted for age, race, and the comorbidities and gynecologic and infertility diagnoses listed in Table 2 . We estimated the percentage of singleton, twin, and triplet or higher order live births across state groups using generalized logit models. Due to the small sample size of triplet or higher order live births, we only adjusted for age, and we combined women who were 41 years old and older into one group. Finally, we used a Poisson model adjusted for age and race with an offset term for the population of each state, age, and race group to estimate the total number of live births after IVF per 100,000 women for each state group. We performed all analyses using SAS Version 9.4 (Cary, NC). Tests were 2-tailed, and we set the probability of Type 1 error at 0.05. Our institution’s Institutional Review Board deemed this study to be exempt from its oversight.

Our final cohort included 23,061 women. Twenty-seven percent of the cohort was in Group 1, 14% in Group 2, and 59% in Group 3. Table 2 compares the demographics of the women across the state groups. Women in Group 1 tended to be slightly older at the time of their first oocyte retrieval [median age in years (IQR), 36 (33–39) vs 35 (31–38) vs 35 (32–38) for Groups 1, 2, and 3 respectively, P<0.001]. Women in Group 2 were more likely to be Black (5.6% vs 10% vs 5.8%, respectively, p<0.001) and less likely to be Asian (15% vs 9.6% vs 17%, respectively, p<0.001). There are no clinically significant differences in specific comorbidities and infertility disorders in patients across state groups. Table 3 displays the utilization rates of IVF by state groups. The IVF utilization was highest in Group 1 states and lowest in Group 3 states (516 vs 94.8 IVF attempts per 100,000 women, p<0.001). Group 1 states also exhibited higher IVF utilization rates than Group 2 states (516 vs 323 IVF attempts per 100,000 women, p<0.001). Table 4 displays fertility outcomes after the first embryo transfer, by state group. Pregnancy and live birth rates were similar across state groups, with a relative difference of 6% or less for all group comparisons. Multiple birth rates were lowest in Group 2 (p=0.012). Preterm birth rates were lower in Group 1 compared to Group 3 (8.4% vs 9.6%, p=0.030). The difference in live birth types (e.g., singleton, twin) was not significant. Table 4 also displays that the absolute number of live births from any IVF attempt per 100,000 women. The number of live births following IVF was highest in Group 1 and lowest in Group 3 (154 vs 31.9 live births per 100,000 women, p<0.001). Group 1 states also exhibited more live births per 100,000 women than Group 2 states (154 vs 106 total live births per 100,000 women, p<0.001).

The National Center for Health statistics reported that all 50 states recorded declining fertility rates from 2015 to 2020 1 , and 43 states recorded their lowest fertility rate in three decades in 2020 2 . Infertility affects up to 15% of couples in the United States (US) 3 . Cost barriers prevent many patients from accessing infertility treatments, and one fifth of infertile couples report that the male partner never underwent an infertility evaluation 5 . One cycle of in vitro fertilization (IVF) costs more than $12,000 on average 6 , and a lack of insurance coverage for IVF forces couples pay the full cost themselves 6 . To improve access to infertility care, 12 states passed laws mandating insurance coverage for IVF, as of 2020 7 . The mandates vary in their requirements for coverage of male infertility care, with seven states mandating coverage of male infertility care in addition to IVF coverage 8 . The remaining five states only mandate coverage for IVF. 8 A lack of mandated insurance coverage for male infertility care may be problematic because male factor is the sole cause of infertility in one third of infertility cases, with an additional 20% of cases with male factor as a contributing cause 9 , 10 . While previous studies have evaluated associations between IVF coverage mandates and IVF utilization or outcomes, none have studied how the inclusion (or exclusion) of male infertility care coverage may affect IVF utilization or outcomes. In addition, since insurance mandates only apply to patients with employer-sponsored health insurance, clinical registries are not ideal data sources for studying coverage mandates because clinical registries do not identify if patients have insurance coverage for IVF or are paying out of pocket 11 , 12 . Thus, our aim was to assess whether the inclusion of male infertility care coverage mandates led to differences in IVF utilization or outcomes. We examined rates of IVF utilization and IVF outcomes in states with mandated IVF coverage that includes male infertility coverage, mandated IVF coverage without male infertility coverage, and without any mandated IVF coverage. Our findings will provide policymakers and advocates with insights regarding the association of male infertility care coverage mandates with IVF use and outcomes that can be used for future policy making and advocacy.

States with IVF coverage mandates that include coverage for male care display higher IVF utilization rates and higher absolute number of live births. We believe the inclusion of male infertility coverage has the potential to expand family building options for infertile patients. In the setting of nationally declining fertility rates, policymakers considering new legislation mandating insurance coverage for infertility care should consider including coverage for male infertility care to promote access to care.

To our knowledge, this is the first study to use nationwide claims data to find differences in IVF utilization between states with and without insurance coverage mandates that include male infertility coverage. Notably, we found that IVF utilization and the absolute number of live births per 100,000 women were highest in states with an insurance coverage mandate that includes male infertility coverage. Among people enrolled in Optum CDM, we found that IVF utilization was highest in states with mandated insurance coverage for IVF compared to states with no mandated IVF insurance coverage 16 . Previous investigators have shown that states with IVF coverage mandates had higher use of IVF when compared to states without such mandates 17 . However, no previous studies have specifically evaluated IVF coverage mandates that include male infertility care. Our current study adds to the existing literature by revealing that there is even higher IVF utilization among insured individuals in states that have an IVF coverage mandate that includes coverage for male infertility care, when compared to states with mandated coverage for IVF without male infertility care. While this is not a causal finding, we reveal a statistically significant association which was not previously known. We also found that IVF outcomes were similar between state groups regardless of mandated insurance coverage. Other studies have shown that multiple birth rates were lower in states with mandates for IVF coverage compared to those without mandated coverage 11 , 12 . However, we did not find that multiple birth rates were clinically different between state groups when controlled for age, race, comorbidities, and gynecologic and infertility diagnoses. Previous studies utilized important IVF clinical registries such as the Society for Assisted Reproductive Technology (SART) Clinical Outcome Reporting System. 15 However, these do not distinguish between IVF cycles covered or not covered by insurance. State insurance mandates only apply to patients with health insurance, and Optum CDM allowed us to specifically identify patients with insurance coverage for their IVF. Our finding that absolute live birth numbers from IVF are highest in states with mandated infertility coverage for both females and males is noteworthy. Our results are consistent with findings reported by other investigators that the total number of live births is higher in states with mandated insurance coverage for IVF 18 . However, we add to the current literature by revealing that the total number of live births is even higher when IVF coverage mandates include coverage for male infertility care. Since IVF outcomes were similar across states groups, it is likely that the higher number of babies born in states with IVF mandates that also cover male infertility care is explained by increased IVF utilization in those states, not necessarily by better outcomes per IVF attempt. Our study has limitations. First, our study used administrative claims data with diagnostic and billing codes, which may result in misclassification through coding errors from providers. However, our team’s previous studies confirm that our methods for identifying pregnancies, births, and multiple births are reliable when compared with SART data. 15 Second, given the nature of claims data, we are unable to measure the number of IVF cycles completed specifically for the indication of male factor infertility. Finally, Optum CDM only includes patients with private health insurance coverage. Thus, our findings may not be generalizable to patients who do not have health insurance or have insurance that is not captured by Optum CDM. However, because we were interested in studying the impact of state insurance mandates, which only apply to people with employer-sponsored health insurance, Optum CDM was an appropriate data source for our study. These limitations notwithstanding, our findings have important implications for many stakeholders. For example, in response to public health concerns surrounding infertility, the Centers for Disease Control and Prevention has advocated for increased insurance coverage for IVF to eliminate disparities in access to infertility care. 19 Our results suggest that state mandates for IVF coverage, particularly for both female and male infertility care, may be tools to address public health concerns regarding population growth. Further, as physicians and patients advocate for expanded insurance coverage for IVF, our results suggest that patients may benefit from IVF coverage mandates that include coverage for male infertility care. Moving forward, researchers should also investigate differences in IVF costs for families between states with and without mandates for male infertility care coverage, including among different socioeconomic groups 20 .