Insurance and geographic variations in non-invasive prenatal testing

Non-invasive prenatal testing (NIPT) during pregnancy is a highly sensitive and specific screening tool for chromosomal aneuploidy. This screening has the highest detection rate for Down syndrome, and unlike invasive methods, carries no risk for miscarriage or other pregnancy complications.^1,2 In 2012, the American College of Obstetrician-Gynecologists (ACOG) recommended that pregnant people at increased risk for aneuploidy, including those aged 35 and older, be offered NIPT.³ Because this relatively new technology has advantages over both traditional serum marker screening and invasive diagnostics with improved detection rate and lower obstetric risk respectively, quantifying disparities in uptake is particularly important.

As with other prenatal services in the U.S., inequities in invasive prenatal testing are well-established.^4,5 Single-site studies have also demonstrated disparities in NIPT based on insurance type, race, and ethnicity, however findings on the role of insurance are inconclusive.^6,7 While important, single-site studies may not be generalizable. Given pervasive inequities in access to prenatal care in the U.S., understanding geographic and insurance coverage disparities in NIPT use at the population-level is critical.

Massachusetts (MA) has one of the highest insurance coverage rates in the country and, by 2015, NIPT was reimbursed by both public and commercial payers for pregnant residents aged 35 and older. The state is therefore a ‘best case scenario’ context from which to understand variations in NIPT uptake. Our objective was to estimate the population-level rate of NIPT uptake in Massachusetts and identify disparities based on insurance type and patient zip-code.

We used the 2015 Massachusetts All-Payer Claims Database, which represents nearly all healthcare utilization records of Massachusetts residents under the age of 65.⁸ These data capture service use, patient age, insurance type, and 5-digit zip-code at the time of billing. We used 2015 data as this was the first year a specific NIPT billing code (CPT 81420) was adopted. Using ICD-9 codes for delivery, we identified pregnant patients aged 35 years or older at their delivery between 7/1/2015 to 12/31/2015 who were eligible for NIPT use per ACOG guidelines.

Our outcome, NIPT uptake, was defined as the number of patients who received NIPT per 1000 eligible (i.e., pregnant people aged 35 and older). We estimated NIPT rates by insurance type, admission to a teaching hospital, and at the zip-code level, the proportion of residents living below 100% of the federal poverty level (FPL), and the proportion of Black or Hispanic residents. Focusing on delivery hospital allowed us to capture potential differences in practice patterns between affiliates of teaching and non-teaching hospitals. Because of documented racial/ethnic and income disparities in invasive diagnostic testing, and in the absence of individual sociodemographic information, we used census data to characterize patient zip-codes by the proportion of Black/Hispanic and low-income residents (high vs. low). Our goal was to assess the role of area-level sociodemographic characteristics and NIPT uptake. To distinguish between the high and low categories, we ranked all zip-codes in Massachusetts by the overall proportion of Black/Hispanic and low-income residents, and defined zip-codes in the top 25% of each category as a “high” zipcode. We initially examined the proportion of Black and Hispanic residents separately, however because there was substantial overlap between the two, we chose to combine these groups. Finally, we estimated adjusted odds of NIPT using logistic regression, accounting for insurance type, delivery at a teaching hospital, and zip-code sociodemographic composition, and including hospital random effects and clustered standard errors at the hospital-level. Individuals with missing data were dropped from the analysis (n = 403).

There were 4506 non-invasive prenatal tests performed on 22,393 pregnant patients aged 35+ (201.2 per 1000) in 2015 (Table 1). We observed considerable geographic variation. Patients living in zip-codes within and surrounding the Boston metropolitan area were more likely to receive NIPT, as were those living in Eastern MA (Figure 1). The observed NIPT rate (per 1000 pregnant individuals aged 35+) was 48.7 for Medicaid enrollees and 272.1 for commercial enrollees; 123.0 and 223.2 for patients living in a zip-code with a high versus low proportion of Black/Hispanic residents; and 107.7 and 218.4 for those in a zip-code with a high versus low proportion of low-income residents. In adjusted models, pregnant people living in a zip-code with a high Black/Hispanic population were significantly less likely to receive NIPT (aOR:0.80; 95% CI: 0.59–0.95) then people living outside these zip-codes, as were those living in low-income zipcodes (aOR:0.80; 95% CI: 0.70–0.93). Medicaid enrollees were less likely than commercial enrollees to receive NIPT (aOR:0.18; 95% CI: 0.16–0.20).

TABLE 1.

Rates (per 1000 eligible patients) and adjusted odds of Non-invasive prenatal testing (NIPT) use

Cohort	NIPT rate	Adjusted OR (95% CI)	P
Total patients (n = 22,393)	201.2
Insurance
Medicaid (n = 7106)	48.7	0.18 (0.16–0.20)	<0.001
Commercial (n = 15,287)	272.1	Ref
Teaching hospital
Yes (n = 5025)	182.1	1.15 (0.80–1.67)	0.453
No (n = 17,368)	206.8	Ref
Zip-code sociodemographicsa
Proportion of Black/Hispanic residents
High (n = 4886)	123.0	0.80 (0.70–0.90)	<0.001
Low (n = 14,586)	223.2	Ref
Proportion living below the FPL
High (n = 3584)	107.7	0.80 (0.70–0.93)	0.003
Low (n = 15,886)	218.4	Ref

Note: The model included hospital random effects and clustered standard errors at the hospital level.

Abbreviations: CI, confidence interval; NIPT, non-invasive prenatal testing; OR, odds ratio; Ref, referent.

The model included hospital random effects and clustered standard errors at the hospital level.

^aZip-code sociodemographic indicators were derived from the Census Bureau’s American Community Survey and defined based on patient zip-code at time of delivery and the population proportion of Black/Hispanic residents or residents living below 100% the FPL within each zip-code.

FIGURE 1.

Zip-code variation in Non-invasive prenatal testing (NIPT) rates (per 1000 eligible patients)

In this population-level analysis of NIPT uptake in Massachusetts, we found that birthing people covered by Medicaid were over five times less likely to receive NIPT than their counterparts with commercial coverage. Lower NIPT rates in zip-codes with a high proportion of low-income or Black/Hispanic residents also suggests that geographic variations in uptake may reflect racial/ethnic and income disparities independent of insurance coverage. We found no significant association between NIPT uptake and delivery at a teaching hospital.

Our findings on geographic variation are consistent with research from Australia and the Netherlands which found that people living in socioeconomically disadvantaged neighborhoods were less likely to receive NIPT.^9,10 The finding that people with Medicaid coverage were less likely to receive NIPT is inconsistent with a single-site study in Wisconsin which found that pregnant people with Medicaid were more likely to receive NIPT than those with commercial coverage.⁶ This discrepancy is likely due to the fact that commercial payers in Wisconsin were not required to reimburse for NIPT. As with the present study, a single-site study in Massachusetts also found lower NIPT uptake among Medicaid enrollees.⁷ A study performed in Colorado which evaluated all prenatal genetic testing found no difference in uptake between Medicaid and commercial enrollees.¹¹ Inconsistent findings on variation by insurance status highlight the fundamental role of payer coverage on financial barriers to NIPT, and how these may drive regional variations in use. In addition to within-state variation identified in the present study, there is likely substantial between-state variation due to differences in state policies governing commercial payers and Medicaid programs.

Research is needed to understand reasons for lower NIPT uptake by Medicaid enrollees in the state, despite coverage of this screening by the Massachusetts Medicaid program. It is possible that the facilities where patients with Medicaid receive prenatal care are less likely to offer NIPT. For example, federally qualified health centers which serve low-income populations may not have a lab on site or a relationship with an NIPT company, both of which are necessary to offer testing. It is also possible that pregnant Medicaid patients do not know about, prefer not to undergo, or are not offered NIPT. Interpersonal racism via clinician discrimination and bias perpetuates inequities in perinatal service delivery and outcomes.^12,13 Structural racism includes policies governing the distribution of healthcare resources and is another fundamental cause of perinatal health inequities.¹⁴ Medicaid typically reimburses services at lower rates than commercial payers, and clinical implementation of NIPT was predominantly led by industry.¹⁵ Consequently, profit-driven decisions in the early diffusion of this technology likely prioritized higherincome populations.¹⁶ Because Black, Hispanic, and low-income birthing people are more likely to be covered by Medicaid, discrepancies in NIPT uptake among Medicaid enrollees disproportionately affect these populations.

This study has limitations. First, we were unable to capture patient-level sociodemographic characteristics with insurance claims data. Zip-code level proxies are crude measures of patient characteristics and should be interpreted as such. We were also unable to capture NIPT use by enrollees with coverage through the Veterans Administration, Tricare, or Medicare, which may offer less generous coverage for pregnancy-related services. Given established variations in state Medicaid programs, our findings may not generalize outside of Massachusetts. This study also used data from the early days of NIPT, and increased awareness of the technology may have improved uptake since then. Finally, because we could not observe patient preferences for NIPT, we were unable to determine whether disparities are a consequence of patient preferences or access barriers. Black patients and those presenting for prenatal care at a later gestational age are less likely to receive genetic counseling, suggesting that discrepancies in NIPT may be due to provider counseling, not patient choice.¹¹ It is also important to note that people who did not receive NIPT may have received other serum screening tests, such as first trimester or quad screens, or undergone invasive procedures such as amniocentesis or chorionic villus sampling.

Our findings highlight substantial disparities in NIPT uptake based on insurance and zip-code of residence. These disparities likely reflect established inequities in prenatal care. Research is needed to identify barriers and facilitators to uptake and to evaluate interventions to address inequities in NIPT use. Specifically, survey research to determine which patients are and are not offered NIPT versus traditional serum screening will be critical moving forward. Ultimately, our findings suggest that the benefits of NIPT availability have not been realized in a substantial portion of the population-particularly among pregnant people who disproportionately experience barriers to care.

Key points.

What is already known about this topic?

• Single-site studies suggest racial, ethnic, and insurance disparities in use of Non-invasive prenatal testing (NIPT).

• Population-level research from outside the U.S. highlights significant geographic variations in NIPT uptake.

What does this study add?

• Enrollees living in zip-codes with a higher proportion of Black and Hispanic/Latino residents were significantly less likely to receive NIPT.

• Enrollees living in zip-codes with a higher proportion of people living below the federal poverty level (FPL) were significantly less likely to receive NIPT.

• Birthing people with Medicaid were five times less likely to receive NIPT than those with commercial coverage.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the Massachusetts Center for Health Information Analysis. Restrictions apply to the availability of these data, which were used under license for this study.