Association of state insurance coverage mandates with assisted reproductive technology care discontinuation.

Using a large national surveillance system, this study quantifies the rate of ART care discontinuation in the United States after an initial failed autologous ART cycle and demonstrates that individuals residing in states without mandated insurance coverage for ART were more likely to discontinue treatment after an initial unsuccessful cycle than those living in states with comprehensive mandated ART insurance coverage. This association remained after controlling for patient and cycle characteristics, including median zip code household income. Further, non-Hispanic Black, non-Hispanic Asian, and Hispanic patients had higher rates of care discontinuation than non-Hispanic White patients in states with and without mandated insurance coverage for any ART. In a retrospective cohort study by Miller et al 14 of 974 good-prognosis couples in New Zealand who qualified for a subsequent publicly funded ART cycle, authors found only a 10% discontinuation rate after the first unsuccessful cycle. Conversely, in our study including all patients regardless of prognosis, 1 in every 4 US ART patients did not have a repeat ART cycle 1 year after an initial failed cycle. When patients were divided by fertility coverage scope mandated in their state of residence, the highest rates of care discontinuation were noted in those living in states with fertility mandates not including ART and states with no mandate. Given the high cost of an ART cycle in the United States, economic barriers may be a factor that limits utilization of continued ART treatment cycles, despite the potential for success after an initial failed cycle. 22 , 38 A study of patients recruited from 8 US fertility clinics reported that the median cost for ARTover 18 months of treatment was $24,373 (based on 2006 dollars), with median out-of-pocket cost estimated at $19,234 for 1 cycle. 23 , 24 Furthermore, Chambers et al showed that 1 fresh ART cycle accounted for 52% of an individual’s average disposable income in states without mandated insurance coverage for ART compared with 13% for states with such coverage. 22 Although many factors contribute to care discontinuation, addressing financial barriers with ART insurance coverage has been shown to increase the proportion of individuals and couples that continue care after an unsuccessful initial ART cycle. 13 Notably, when comparing ART care discontinuation rates across mandate levels, the percentage of patients who discontinued care was highest among individuals living in states with mandated insurance coverage for fertility services not including ART. This group also had the highest relative effect size in the adjusted analysis than patients living in states with no mandated insurance coverage. We hypothesize that state-mandated insurance benefits for fertility testing and treatment excluding ART may increase the initiation of fertility services, including ART. However, following an initial failed cycle, these patients may discontinue care at higher rates owing to the financial burden of additional cycles. Conversely, in states with no mandated insurance coverage for fertility services, patients may not ever pursue infertility care or proceed to ART owing to the financial burden. It is important to recognize that mandated insurance coverage for fertility services does not result in coverage for all state residents, as coverage is limited to individuals who have employer-based insurance plans that fall under the state insurance mandates. Moreover, in states with mandated insurance coverage for fertility services, the eligibility criteria for coverage can be limited to specific populations, for example, heterosexual couples using autologous gametes who meet specific diagnostic criteria; this would exclude single individuals, same-sex couples, and patients who require donor gametes. 26 , 27 These mandates do not always include treatment for men with male factor infertility. 39 Many state laws do not require religious organizations, smaller employers, employers that self-insure, and public insurers such as Medicare and Medicaid to provide insurance coverage for fertility treatment. 26 , 27 Therefore, this analysis may underestimate the potential benefit of mandated insurance coverage for ART that included all patients, insurance types, and employer categories. Previous studies have shown that individuals who identify as African American or Hispanic are less likely to utilize infertility care, even in states where comprehensive mandates are in place, 35 and are more likely to discontinue care after an initial failed cycle regardless of insurance coverage level. 13 Similarly, in our study, non-Hispanic Black, non-Hispanic Asian, and Hispanic patients were more likely to discontinue care than non-Hispanic White patients in both states with and without mandated insurance coverage for ART, even after controlling for median household income in patient zip code, suggesting the influence of other structural or social determinants of care continuation. 13 Black and Hispanic women are more likely to have public insurance or no insurance than the non-Hispanic White population, and state mandates for ART coverage examined in this study do not apply to uninsured women or those covered by public plans. 40 Lower median household income ($41,361 and $51,450 for Black and Hispanic households, respectively vs $70,642 for non-Hispanic White households in 2018) 41 , 42 may also limit the ability to continue care after a failed cycle even with insurance coverage for ART, owing to the burden of uncovered costs including copays, missed work, travel, and transportation. In a survey of women accessing fertility treatment in Illinois, African American and Hispanic women reported having more difficulty taking time off work and paying for treatment than White respondents. 43 Thus, it is important to identify other factors associated with equitable utilization and continuation of ART care beyond state insurance coverage mandates. The study limitations include an inability to identify patients who switch fertility clinics to continue care and identify those with previous cycles completed at other clinics. In addition, this study does not account for individual-level insurance information that is not available in NASS. Thus, this study evaluated the effect of state-level laws mandating insurance coverage for fertility services and not the effect of insurance coverage itself. The availability of actual insurance coverage may be impacted by the employer mix within a state and the number of employers that are subject to existing mandates. Furthermore, the range of benefits covered by an employer’s plan could also be beyond what is required by insurance mandates for ART, if present. We had limited ability to account for differences in income and wealth in this study. The estimation of average household income was done using residential zip code, which may underestimate the true income of patients in this cohort seeking ART. In addition, the “distance to clinic” variable was calculated from the main clinic address provided by the reporting clinic and did account for potential clinic satellite locations used by patients in this cohort. Finally, the large amount of missing data for race or ethnicity is a limitation of this analysis despite the use of multiple imputation. This study demonstrates an association between categories of state-mandated insurance coverage for ART and ART care discontinuation after an initial unsuccessful cycle but does not imply that mandated ART insurance coverage is the cause for this relationship. State-mandated insurance coverage for ART has been correlated with increased rates of utilization; however, it is unclear if these mandates increase population demand and utilization or if increased utilization and therefore demand drives the development and implementation of mandates in these states. 44 Finally, this observational, retrospective cohort study may be subject to bias because of uncontrolled confounding. Specifically, if the rate of ART discontinuation was influenced by factors that are also associated with state of residence but not controlled for in our analysis, it may confound the correlation between mandated insurance coverage and care discontinuation. Economic barriers contribute to disparities in the utilization of effective infertility treatment, including ART, in the United States. 45 In this study, comprehensive mandated insurance coverage for ART at the state level was associated with lower rates of ART care discontinuation than states with no mandates for ART. Although state-mandated insurance coverage for ART was associated with a reduction in disparities in care discontinuation seen in the lowest category of median household income by zip code, racial and ethnic disparities persisted. Future collection of additional patient-level information in NASS may enhance research addressing the economic, geographic, social, cultural, and structural barriers that limit access to and utilization of this health service.

The data used for this study were derived from the US National ART Surveillance System (NASS)—a congressionallymandated reporting system that collects information on nearly all (98%) ART cycles performed in the United States. 4 In NASS, ART cycles, defined as fertility treatments in which oocytes or embryos are handled in the laboratory, include cycle-level information on patient characteristics, clinical characteristics of the ART procedure, and pregnancy outcomes. We selected all patients who began their first autologous ART cycle from January 1, 2016– December 31, 2017 and did not have a live birth as a result of this cycle (N=127,090). Patients were excluded if the cycle intention was long-term banking (N=27,916), the intended transfer was to a gestational carrier (N=768), the cycle outcome was unknown (N=365), or patient death occurred (N=9). Patients were also excluded if their primary residence was outside the United States, including US territories, or if state of residence was unknown (N=6708). These exclusion criteria applied to each patient’s initial cycle only. Subsequent patient cycles recorded in NASS during 2016 to 2017 were linked to each patient’s first cycle to identify those who continued care by undergoing an additional cycle—an oocyte retrieval or embryo transfer—within 12 months of the previous cycle’s date of failure and those who discontinued care (ie, did not undergo an additional cycle within 12 months). For patients who intended to undergo fresh transfer during their initial cycle, care discontinuation was determined within 1 year after the date of the initial cycle’s outcome or date of transfer when no pregnancy occurred or when the outcome was an ectopic or heterotopic pregnancy. For patients who cryopreserved all embryos made during their initial cycle, care discontinuation was determined 1 year after the date of cycle outcome (or date of transfer when no pregnancy occurred or for ectopic or heterotopic pregnancies) of the first intended frozen embryo transfer with embryo(s) derived from the initial cycle. Given the variability in state-mandated insurance coverage for fertility services in the United States and the lack of an established classification standard, 26 we stratified states into 4 groups based on the scope of covered fertility benefits required by the state mandate in effect during 2016 to 2017 ( Figure ). All laws considered in this stratification were in effect during the entire study period. The first group included states with a comprehensive mandate, defined as covering at least 3 ART cycles. The second group included states with a limited mandate, defined as covering 1, 2, or an unspecified number of required ART cycles. The third group included states with mandates for fertility care not including ART. This group includes states with mandates for insurance coverage for fertility services not specifically including or excluding ART coverage or mandates that insurers offer policies covering infertility treatments. The remainder of states were categorized as having no mandate. 26 – 28 Stratification of states was performed by the authors’ review of the existing fertility coverage laws in place during the years studied and were developed with guidance from previously published work on insurance mandates for ART in the United States. 37 Patients were separated into these 4 groups based on their provided state of residence. Patient demographic characteristics, medical history, and ART cycle information were extracted from NASS. Information on race or ethnicity was missing for 38% of included patients. Procedures used to address missing patient race or ethnicity, calculate a patient’s distance from their ART clinic, and determine median household income by zip code can be found in supplementary materials . We used Pearson chi-squared tests to evaluate associations between patient demographics, reproductive history, cycle characteristics and outcomes, and treatment discontinuation. Multivariable analyses were conducted to evaluate factors independently associated with care discontinuation. Variables included in the final multivariable model were chosen a priori based on clinical expertise ( Supplemental Materials ). We used multilevel logistic regression models (PROC RLOGIST in SUDAAN) to analyze the multiply imputed data and account for clustering by clinic. Predicted marginal proportions were calculated to estimate unadjusted and adjusted relative risks (aRR) and accompanying 95% confidence intervals (CIs) for each covariate. To further evaluate factors associated with care discontinuation in states with mandated insurance coverage for any ART compared with those without mandated insurance coverage for ART, we stratified analyses into the following 2 categories: patients residing in states with a comprehensive or limited ART insurance mandate and those residing in states with no mandate or a mandate that did not include ART. Multivariate analyses were conducted controlling for the same variables listed above. Statistical analyses were conducted using SAS version 9.4 (SAS Institute) and SUDAAN 11.0.3 (RTI International). All P values <.05 were considered statistically significant. Epidemiologic research using NASS data was reviewed and approved by the institutional review board at the Centers for Disease Control and Prevention.

In the United States, during 2016 and 2017, 91,324 patients underwent their first autologous ART cycle and did not have a live birth. Of these individuals, 41.6% had their first cycle cancelled before oocyte retrieval or embryo transfer, and the remaining 58.4% did not achieve live birth after embryo transfer. Of this cohort, 26.4% did not pursue additional treatment within 1 year of their first unsuccessful cycle outcome. More than half (56.7%) of the patients in this cohort lived in a state where there was no mandated insurance coverage for fertility services ( Table 1 ). The remaining patients lived in a state that was classified as comprehensive mandate (15.1%), limited mandate (2.9%), or a state with mandated insurance coverage for fertility services not including ART (25.3%). The mean age of this cohort was 35.3 years. Overall, 65.0% of patients identified as non-Hispanic White with non-Hispanic Asian, Hispanic, and non-Hispanic Black patients representing 15.9%, 9.1%, and 8.8%, respectively. The most common infertility diagnoses were male factor (31.3%) and diminished ovarian reserve (29.1%) ( Table 1 ). In this cohort of patients who did not have a live birth after their first ART cycle, patients who lived in states categorized as having comprehensive mandates discontinued care at the lowest rate (21.1%) than those in states with limited mandates (25.4%), mandates for fertility coverage not including ART (29.4%), and no mandated insurance coverage for fertility benefits (26.4%) ( Table 2 ). Patients living in states with insurance mandates not including ART or states with no mandates were 46% (aRR, 1.46; 95% CI, 1.31–1.63) and 26% (aRR, 1.26; 95% CI, 1.15–1.39) more likely to discontinue care, respectively, than patients who lived in states with comprehensive ART insurance mandates after controlling for patient and cycle characteristics ( Table 2 ). Other demographic characteristics were also associated with ART care discontinuation in this population after adjusting for potential confounding factors ( Table 2 ). Increasing female age, distance from clinic ≥50 miles, median zip code household income <$100,000 per year, and previous live births were all associated with increased risk of care discontinuation. Patients who identified as non-Hispanic Black, non-Hispanic Asian, or Hispanic were 1.13–1.29 times more likely to discontinue care than those identified as non-Hispanic White. Patients with “other” included as a reason for ART were significantly more likely to discontinue care (aRR, 1.12; 95% CI, 1.08–1.16), whereas those with polycystic ovary syndrome (aRR, 0.91; 95% CI, 0.87–0.95) and unexplained infertility (aRR, 0.89; 95% CI, 0.84–0.94) diagnoses were less likely to discontinue care than those without those diagnoses. Cycle outcomes associated with discontinuation of care included having fewer oocytes retrieved (none, 1–10, and 11–20 oocytes compared with ≥21 oocytes) and not having embryos cryopreserved (aRR, 2.24; 95% CI, 2.12–2.37). Patients whose first cycle resulted in spontaneous abortion (aRR, 1.13; 95% CI, 1.09–1.17), stillbirth (aRR, 1.64; 95% CI, 1.46–1.84), or induced abortion (aRR, 1.32; 95% CI, 1.18–1.48) were more likely to discontinue care than those who did not achieve pregnancy after embryo transfer. In contrast, cancellation before retrieval (aRR, 0.53; 95% CI, 0.45–0.62) or before transfer (aRR, 0.69; 95% CI, 0.65–0.74) were both associated with a lower risk of care discontinuation than patients with no pregnancy after embryo transfer. When our cohort was stratified into 2 groups based on presence (comprehensive or limited mandate) or absence (fertility mandate not including ART or no mandate) of mandated insurance coverage for any ART, patients residing in zip codes with median household income <$50,000 per year remained at an increased risk of care discontinuation in states without mandated insurance coverage for ART (aRR, 1.17; 95% CI, 1.01e1.36) but the association was no longer significant in states with mandated insurance coverage for ART (aRR, 1.23; 95% CI, 1.00–1.50) ( Table 3 ). The risk of care discontinuation remained elevated among non-Hispanic Black, Hispanic, and Asian patients than in non-Hispanic White patients regardless of whether they resided in a state with or without mandated insurance coverage for any ART ( Table 3 ).

According to the World Health Organization, the American Medical Association, and the American Society for Reproductive Medicine, infertility is a disease. 1 – 3 Since the first infant was delivered after in vitro fertilization (IVF) in the United States in 1981, there has been a consistent increase in the availability and utilization of assisted reproductive technology (ART) such that in 2019, 2.1% of all infants born in the United States were the result of ART. 4 Given that the cumulative likelihood of live birth increases with subsequent ART cycles, successive use of ART after an initial failed cycle can improve cumulative success rates for patients with infertility. 5 – 8 Despite this, 10% to 37% of patients pursuing ART discontinue care after their first unsuccessful cycle. 9 – 14 Involuntary childlessness is associated with symptoms of anxiety, depression, and complicated grief; patients who have unmet parenthood goals after seeking fertility evaluation or treatment have been shown to have negative long-term health consequences. 15 – 17 Existing data describing ART care discontinuation report that emotional distress, perception of poor prognosis, increasing female age, and lower oocyte or embryo yield are associated with early care discontinuation. 9 – 12 , 18 – 20 These studies primarily involved patients in countries where national health plans or other public funding systems cover costs associated with ART cycles for eligible patients, limiting generalizability to countries where widespread financial coverage is not available and many patients pay out-of-pocket. 9 – 11 , 18 , 20 In the United States, where insurance coverage for ART is not required, a single IVF cycle has been estimated to cost between $12,400 and $24,000, and evidence suggests that cost may be a treatment barrier. 21 – 25 Since the 1980s, 19 states have passed laws to improve access to fertility treatment that vary greatly in scope. 26 – 28 More specifically, some states have enacted “mandates” that require certain categories of employers to include coverage for fertility-related services. 26 – 28 Studies show that these insurance mandates are associated with increased ARTutilization, fewer embryos per transfer, and lower rates of multiple gestations. 29 – 33 In 1 retrospective study, patients with insurance coverage for IVF (both mandated and nonmandated) were more likely to attempt IVF again if the first cycle was unsuccessful and had a higher cumulative probability of live birth than those who self-paid for IVF. 34 A secondary analysis of the same data looking at factors associated with discontinuation showed that women without IVF insurance coverage were more likely to discontinue treatment than women with insurance coverage, 13 Both studies used data obtained from a single center. Despite these reported benefits of insurance coverage for ART, racial, ethnic, and socioeconomic disparities in utilization have been found to persist even in states with mandates. 35 , 36 Therefore, it is important to better understand how insurance coverage for infertility services impacts the patterns of ART care in states with mandates to further understand factors that contribute to disparities in fertility care utilization nationally. This population-based retrospective cohort study aims to quantify the rate of discontinuation in ART care after an initial failed cycle in the United States, determine if patients living in states without comprehensive mandated insurance coverage for fertility services are more likely to discontinue ART after an initial failed cycle, and evaluate which patient and cycle outcomes are associated with care discontinuation.