This analysis of APCD data shows that genetic testing for hereditary cancer susceptibility increased moderately in a linear fashion from 2013 through 2018 in Arkansas. Applying the time series coefficients to our baseline rates of testing suggests that testing for HBOC and Lynch syndrome nearly doubled for those with commercial and state employee coverage. In comparison, Knerr et al. reported a 33% increase in BRCA testing from 2005 to 2015 in women over 18 without an incident breast or ovarian cancer within their hospital system in Washington state [13]. Pace et al. reported that in Massachusetts, the mean monthly number of tests per 100,000 women doubled to quadrupled from 2011 to 2015, depending on the type of insurance [14]. However, the absolute rates of testing for HCS in Arkansas remained low. Less than 1 out of every 10,000 persons received Lynch syndrome testing, while fewer than 5 out of every 10,000 received HBOC testing. Estimates for the prevalence of mutations in the general US population are 1/300 for Lynch syndrome and 1/400 for BRCA1/2, emphasizing that many mutation carriers remain undiagnosed [1,2,3].

Additionally, our data suggest that health plan coverage is an important determinant in accessing cancer genetic testing. State employees had the highest rates of HBOC and Lynch syndrome testing and equally high rates as Medicare enrollees for any cancer genetic testing. Medicare enrollees had the highest rate for tier 2 molecular pathology procedures. The testing rates were substantially lower for Medicaid enrollees compared to enrollees in commercial, state employee, and Medicare plans. The overall rates of HBOC and Lynch syndrome testing were highest among women 18–64 years old, while rates of tier 2 molecular pathology procedures and any cancer genetic testing were highest among men 18–64 years old. In multivariate exploratory models that controlled for key demographic differences, those with state employee coverage were 1.6 to 4.5 times more likely to receive genetic testing compared to commercial enrollees, while commercial enrollees were substantially more likely to receive any type of genetic testing than Medicaid enrollees. Type of coverage played a similarly important role in a study more narrowly focused on women receiving BRCA1/2 testing in Massachusetts; testing among privately insured women increased from 9.3 in 2011 to 18.4 in 2015, while among Medicaid enrollees, it increased from 3.7 in 2011 to 14.7 in 2015 [14].

The lower rates of genetic testing in Medicaid enrollees may be due to several factors. First, 51% of all Arkansas Medicaid recipients in 2017 were age 20 or younger, the age group that is the least likely to have cancer genetic testing [15]. However, multivariate models controlling for age found that Medicaid recipients remained the least likely to receive any of the genetic testing examined, which suggests that substantial disparities may exist in the diagnosis of hereditary cancers across economic classes. The Affordable Care Act (ACA) implemented in 2014 mandated that genetic testing for HBOC be considered a preventive service in qualified individuals and be covered without cost sharing [15]. Arkansas participated in ACA expansion of access to insurance, unlike many other Southern states; however, this participation did not yield increases in receipt of genetic testing for Medicaid enrollees comparable to those with state employee or commercial coverage. Many studies document that physicians are less likely to discuss and order genetic screening among racial/ethnic minorities and individuals with less than a college level education [9, 16,17,18].

When genetic testing is offered by the physician, the cost is often mentioned as a major hurdle, with 49% of persons considering out-of-pocket cost as a major deterrent [19]. Some companies providing testing have specific policies that preclude billing patients for testing if Medicaid does not cover the test, but not all providers offering testing know the details of these financial policies for each company. Our results echo other data that document only modest increases in genetic testing since the ACA was implemented, suggesting that this provision may have expanded access to these services but may not be sufficient to reach all those at risk for hereditary cancer [20].

Considerable variation exists among commercial payers regarding coverage for genetic testing, which could partially explain the low uptake of testing in our study. A 2015 study reported that 76% of the private payers had coverage policies for BRCA1/2 testing [21]. The coverage for multigene panel testing was low, with only 23% of the payers covering panel testing and only under the condition that all genes tested in the panel were medically necessary for the individuals [22, 23]. One explanation provided by many commercial payers is that they deem multigene testing an experimental diagnostic or treatment approach [21, 23, 24]. Additionally, 70% of private payers required prior authorization for both single and multigene testing in a 2018 study [23]. High-quality studies demonstrating penetrance and effectiveness of multigene panels in a clearly defined population could persuade private payers to cover panel testing and likely increase its utilization; obtaining such data, though, is complicated by the differences in the genes included in panels offered by different companies. In the meantime, companies providing genetic testing for HCS have coped with the lack of coverage of gene panels by billing for testing for specific genes, such as BRCA1/2, that are routinely covered by insurance rather than all the genes included on the panel.

Women had 18-fold higher odds of testing for HBOC and had nearly 2-fold higher odds of testing for Lynch syndrome, which confirms a consistent finding that men are less likely to receive genetic testing than women [9]. Anxiety disorders and miscellaneous mental health disorders (which included conditions such as sleep and dissociative disorders) had positive associations with general genetic testing and HBOC testing in women. A previous study in women from BRCA1/2 mutation-negative families reported that women who had a higher perceived lifetime risk of cancer and higher worry about cancer were more likely to show an interest in genetic testing [25]. Another study, however, reported no association between distress and participation in BRCA testing [26]. Given the cross-sectional approach in our study, we cannot comment on whether the positive association between anxiety and other mental health disorders and genetic testing is due to mental health diagnoses before or after genetic testing.

This study has several limitations. We could not evaluate genetic testing trends across racial and ethnic groups. Unfortunately, the claims data for commercial insurance lack information on race and ethnicity. We did not include enrollees with a personal history of cancer in the current analysis; our analysis of genetic testing trends is focused on those who had not had cancer. Further, claims data reflect that a genetic test was performed, not the result of the test; we know that an enrollee had a test for HBOC, for example, but not whether that test documented a mutation. Insurance plan types have different lags in data availability, limiting our ability to evaluate the trends over a consistent time period for all four plan types. The exploratory models used a combination of cross-sectional and longitudinal approaches, the results of which should be interpreted as exploratory associations. Finally, the structure of billing codes for genetic testing for HCS is complex, constantly evolving, and does not allow for specific designations of each gene included in a testing panel at this time. The tier 2 molecular molecular proceduresrefer to the complexity of the testing procedure, not the genes tested, resulting in codes that include both cancer and non-cancer testing. Consequently, our ‘any cancer genetic’ test category, which includes tier 2 molecular tests, likely overstates genetic testing specific to cancer.

This study documented a modest increase in genetic testing for HBOC and Lynch syndrome across Medicaid, commercial, state employee and Medicare plans in Arkansas, although rates were lower than that observed in other states. The exploratory models highlight the influence of sex and health care coverage as a key determinant in accessing genetic screening. Further studies are ongoing to understand the barriers to genetic testing in Arkansas.