Introduction

Advances in knowledge and technology have led to an increased availability and awareness of genetic testing (Roberts and Ostergren 2013). There are two main access points for genetic testing: (1) clinician-facilitated medical testing and (2) direct-to-consumer (DTC) personal genetic testing (PGT). Despite the growing availability of both, research suggests that awareness, use, and interest in medical genetic testing are lower among American racial minority groups (Mai et al. 2014). Factors suggested as possible reasons for lower uptake among racial minorities include decreased referral from minority-serving physicians (Shields et al. 2008), lower rates of health insurance (Armstrong et al. 2006), medical distrust (Armstrong et al. 2008), and concerns about potential harm (Catz et al. 2005). Scarce literature exists on racial minorities’ engagement with DTC-PGT (Hensley Alford et al. 2011; Bloss et al. 2010); although a number of population surveys have reported lower levels of DTC-PGT awareness among racial minority groups (Ortiz et al. 2011; Kolor et al. 2012; Finney Rutten et al. 2012; Agurs-Collins et al. 2015; Langford et al. 2012). When compared to Whites, however, differences in awareness have not consistently been statistically significant (Finney Rutten et al. 2012; Agurs-Collins et al. 2015). One investigation reported that lower DTC-PGT awareness among Blacks was partially mediated by numeracy levels (Langford et al. 2012), which highlights the need to examine additional factors when considering racial differences in genetic testing participation.

Only two previous studies that offered DTC-PGT as part of study participation evaluated differences in uptake by racial group, and each reported that Whites may be more likely to undergo DTC-PGT than non-Whites (Hensley Alford et al. 2011). In the multiplex study, participants were offered free genetic risk information for eight common complex conditions (Hensley Alford et al. 2011). Blacks were less likely to agree to participate in the study and the authors proposed that this might be due to lesser trust in researchers. Bloss et al. conducted an assessment of behavioral responses to DTC-PGT results among individuals who were offered DTC-PGT for a reduced price (Bloss et al. 2010). When participant demographic variables were compared to non-responders, it was reported that non-White individuals were less likely to participate in the study. The authors concluded that non-White individuals may be less likely to purchase and undergo DTC-PGT. Although these studies suggest non-White groups are less likely to undergo DTC-PGT, no previous study has investigated individuals of racial minority background who have actually purchased DTC-PGT. This study is the first direct investigation of actual DTC-PGT customers with racial minority backgrounds compared with White customers.

Given the increased presence of genetics in health and society, particularly with the rise of DTC-PGT and the promise of precision medicine for Americans (Jaffe 2015), it is important to understand racial minorities’ interest in and use of genetic testing. To contribute to this research area, we conducted the first exploration of interest and decision-making factors among Black and Asian individuals who purchased DTC-PGT. Our results suggest that some specific interests for purchasing DTC-PGT may exist for certain racial groups, but overall racial minority DTC-PGT customers seek out the product for largely the same reasons as White customers.

Materials and methods

Study design and procedures

The present exploratory analyses utilize baseline data from the Impact of Personal Genomics (PGen) Study, a longitudinal investigation of DTC-PGT customers. Details of the PGen Study design and methodology have been reported previously (Carere et al. 2014). Briefly, new customers of two DTC-PGT companies, 23andMe, Inc. (23andMe) and Pathway Genomics (Pathway) were recruited after ordering PGT. Following online consent, participants were invited to complete a baseline web-based survey that assessed interest and motivations for undergoing PGT. In total, 1838 individuals consented to participate in the PGen Study and 1648 participants (90%) completed the baseline survey prior to viewing their PGT results. The PGen Study was approved by the Partners Human Research Committee and the University of Michigan School of Public Health Institutional Review Board.

Baseline survey

Participants reported demographic characteristics and were asked to select their race (all that applied) from a list of options: American Indian/Native Alaskan, Asian, Black, Hawaiian or Pacific Islander, White, or Other (please specify). These racial categories are used by the US Census Bureau (https://www.census.gov/2010census/) and have enabled comparisons to the US population (Carere et al. 2014). Survey items utilized 3- and 5-point Likert scales (full response items reported in the “Results” section) to assess interest in specific types of genetic information (health and non-health related); factors in the decision-making process; level of trust in the PGT company regarding privacy and use of data; and perceived medical utility of PGT results. Questions related to receiving genetic information about non-health-related “traits” only appeared to 23andMe customers, since Pathway Genomics did not include information about traits in their test results. An explanation of the term “traits” is provided on the 23andMe website where the test is purchased, and the specific traits that are included in the genetic test are also listed on the website. Customers view the same explanation of traits when they receive their test results.

Data analyses

From the 1648 participants who completed the baseline survey, we excluded from analysis participants who did not report their race (n = 1), reported more than one race (n = 111), and those who were not Black, Asian, or White (n = 49) for a total sample of 1487 participants. Descriptive statistics were used to characterize demographic characteristics. Chi-square tests were used to compare survey responses across the three racial groups, except in cases where > 20% of cells had expected counts of less than five, in which case Fisher’s exact test was used. In addition, we conducted multinomial regression to compare survey responses overall across all racial groups (item-level comparison) as well as between the racial minority and White group (race group comparison e.g., Black vs. White; Asian vs. White), with adjustment for age, sex, education, PGT company, health insurance, and income. These covariates were selected because they could be associated with motivations for pursuing and perceptions of PGT. In the race group comparison models, White was selected as the reference race group (as it was the largest group), and the middle response option was set as the reference for the outcome (e.g., “Somewhat Interested” on a 3-point scale or “Neither Agree nor Disagree” on a 5-point scale) because this provided the most stable and readily interpretable coefficients. An overall p value for the item-level comparison was computed, followed by odds ratios (OR), 95% confidence intervals (CI), and p values for the race group comparisons. Analyses were conducted using SAS software (version 9.4; SAS Institute, Cary, NC) and multinomial regression models were fitted using PROC LOGISTIC. Statistical significance for all analyses was set at p < 0.05.

Results

Participant characteristics are presented in Table 1. Across racial groups, the PGen population had high levels of education and income, and ≥ 95% of participants had health insurance. The majority of participants were 23andMe customers (66% of Whites, 84% of Blacks, and 56% of Asians). A larger proportion of Asian participants were male (49 vs. 39% of Whites and 36% of Blacks) and single (35 vs. 17% of Whites and 18% of Blacks) with no children (74 vs. 45% of Whites and 43% of Blacks).

Table 1 Demographic characteristics of PGen Study participants
Full size table

A number of significant differences in survey responses were observed in unadjusted analyses (Table 2). These included interest in pursuing PGT to obtain information related to “risk of disease or health condition” (p = 0.025), “carrier status” (p = 0.035), “ancestry” (p = 0.001), and “traits” (p = 0.0001). Among these particular items, Asian participants reported the greatest interest in disease-risk information (77.8% “very interested”), while Blacks reported the greatest interest in information related to ancestry (95.4% “very interested”) and traits (91.9% “very interested”). In addition, Blacks reported the least interest in carrier status information (18.2% “very interested”). Related to the potential trend of greater interest in non-health-related information, only 13.5% of Blacks participants (compared to 35.2% of Asians and 31.0% of Whites) indicated that the item “how well the results predict whether I’m going to get a particular disease” was something that they “considered a lot” when deciding to pursue DTC-PGT (p = 0.034). Moreover, fewer Black participants (40.9%) compared to Whites (61.2%) and Asians (74.1%) indicated that “interest in finding out about my personal risk for specific diseases” was a “very important” factor in their decision to seek DTC-PGT (p = 0.012), while a much larger proportion (91.0%) indicated that “curiosity about my genetic make-up” was a “very important” factor in their decision compared to Whites (77.3%) and Asians (72.2%) (p = 0.038). Asian participants were more likely to indicate that “…limited information about my family health history” was a “very important” factor for seeking DTC-PGT, and least likely to “strongly agree” (31.5%) with the item “I trust (PGT company) to use my genetic information only for the purposes to which I consented.” (p = 0.0004).

Table 2 Differences in responses to survey items on interests, considerations, importance, and trust between Whites, Blacks, and Asians in the PGen Study
Full size table

Results from the adjusted multinomial regression models indicated significant differences among item-level comparisons for only two survey items: interest in traits (p = 0.0006) and desire to learn about one’s genetics due to limited family health history information (p = 0.005) (Table 2). In the subsequent race group comparisons, a greater proportion of Black participants compared to Whites reported that they were “very interested” in receiving PGT information related to traits (91.9 vs. 70.8%; OR = 6.86 (95% CI: 1.62–29.1), p = 0.009). Conversely, a greater proportion of Asian participants compared to Whites reported they were “not at all interested” in receiving information related to traits (13.3 vs. 2.0%; 20.9 (3.89–112.03), p = 0.0004). In addition, a greater proportion of Asian participants reported that a “desire to learn more about my genetics because I have limited information about my family health history” was a “very important” factor for seeking DTC-PGT compared to Whites (58.0 vs. 37.5%; 3.53 (1.62–7.69), p = 0.002). Three survey items possessed too few cell counts in some response categories to fit a model appropriately and so we were unable to generate adjusted results for these items (“N/A” in Table 2).

Discussion

Overall, we found few differences in interests in and motivations for purchasing DTC-PGT between Blacks, Asians, and Whites who were customers of 23andMe or Pathway Genomics. The significant results from our unadjusted tests were, for the most part, no longer significant after multivariable adjustment. However, we have presented them here to fully explore and describe the responses of our unique study population, since racial minority customers of DTC-PGT have not been directly investigated previously. Moreover, the overall narrative from the significant unadjusted results remained present in our significant adjusted findings—namely, an interest in non-health-related information among Black customers and a desire to learn about genetics due to limited family health history among Asian customers.

The data presented here represent a detailed evaluation of interest and considerations for participation in DTC-PGT from participants across racial groups who chose to have DTC-PGT testing. The engagement of racial minority groups in DTC-PGT specifically is under-described in the literature and, therefore, our data and reported findings contribute valuable new information to the study of the field of consumer genomics. Other studies related to this area of investigation have examined how minority groups perceive various types of genetic information. Rahm et al. examined perceptions of DTC-PGT advertising among various demographic groups and reported that individuals from minority groups may not consider genetic knowledge to be as empowering as Whites (Rahm et al. 2012). One vignette-based study found that minority groups were more likely to report reluctance to adhere to medical prescriptions based on genetics compared to conventional prescriptions (Butrick et al. 2011). While racial differences in knowledge/interest in genetic testing have been reported (Mai et al. 2014; Pagan et al. 2009), contradictory evidence exists. Armstrong et al. reported no statistically significant difference in willingness to undergo medical genetic testing between Whites and other racial groups (Armstrong et al. 2012) and Catz et al. reported no difference in either knowledge or interest in genetic testing between racial minorities and non-Hispanic Whites (Catz et al. 2005). Therefore, research surrounding interest and knowledge about genetic testing among minorities is conflicting and suggests that additional factors beyond race may contribute to reported differences.

A number of limitations should be noted when considering results from our exploratory analyses. The sample size of our minority racial groups was small, so the responses of these participants may not reflect those of individuals from these racial backgrounds. However, no previous study has directly examined customers of DTC-PGT and so our results provide the first documentation of the interests and decision-making factors of actual DTC-PGT customers with racial minority backgrounds. We also performed several statistical tests on a large number of survey items, so it is possible that the two significant findings reported from our adjusted models arose from chance. However, stringent corrections for multiple testing may not be appropriate to our investigation, as several survey items are related and so item-level responses are unlikely to be independent. Moreover, the magnitude of the difference in the response percentages is notably large between race groups (e.g., 91.9% Blacks vs. 70.8% Whites “very interested” in traits), which suggests that a difference in perceptions was likely present. In addition, we excluded individuals who reported more than one race which contributed to our limited sample size for this investigation. We felt that limiting the analysis to individuals of one racial background would minimize the potential for miscategorization of race, since several diverse combinations of race categories comprised the “multi-racial” group and examination of free text responses illustrated participants’ uncertainty in reporting race. Moreover, we did not account for Hispanic ethnicity in our analyses, but a sensitivity analysis with Hispanic participants omitted did not alter the results (data not shown). Finally, the generalizability of our findings to the US population may be limited because the PGen Study included customers from only two DTC-PGT companies, and these customers tended to be high-earning and highly educated, with high levels of health insurance coverage (Carere et al. 2014). Our findings should assist in generating hypotheses for future studies with a larger non-White sample. Future studies aimed at more rigorously exploring race and engagement with DTC-PGT should target racial minority individuals for recruitment to obtain larger sample sizes. The categorization of race groups may be expanded to more precisely characterize race.

Overall, our findings demonstrate few differences in interests and perceptions of DTC-PGT between racial minority customers and White DTC-PGT customers, although some particular interests may exist for specific groups. Persons who seek out DTC-PGT apparently do so with largely similar perceptions and for largely similar reasons regardless of their racial identity.