Confronting Genetic Testing Disparities: Title and subTitle BreakKnowledge Is Power

In 1597, when philosopher Francis Bacon first composed the phrase Nam et ipsa scientia potestas est, “For knowledge itself is power,” he could never have imagined its current applicability to reducing disparities in cancer care and prevention in the United States. The landmark Institute of Medicine report Unequal Treatment: Understanding Racial and Ethnic Disparities in Health Care¹ revealed that racial- and ethnic-minority patients receive inferior health care compared with white Americans, even when income and insurance inequalities are leveled. The promise of early cancer detection and prevention is within reach as newer technologies become incorporated into medical practice. Unfortunately, new technologies are likely to increase health care disparities even further as they widen the gap between those who already receive the best care and those who do not.

Specifically, the use of genetic counseling and testing for primary cancer prevention is yet another area of medicine in which disparities may be predicted to increase. Referral for counseling is often initiated by a primary care physician, an individual luxury more common among insured patients.² Referral also involves a careful review of family cancer history, the completeness of which may be inhibited by race- and ethnicity-specific practices and beliefs.³ Economic constraints secondary to the high cost of genetic testing and fears of discrimination (genetic, insurance, or otherwise) may also contribute to race/ethnicity-related disparities.⁴

Despite having a lower incidence of breast cancer, African American women have a higher breast cancer mortality rate than white women, due in part to diagnosis of the disease at a more advanced and less curable stage.^{5 - 7} In this issue of JAMA, Armstrong et al⁸ explore racial disparities in utilization of genetic counseling for primary prevention of breast or ovarian cancer in the University of Pennsylvania Health System.⁸ Their findings are revealing: among African Americans, the unadjusted odds ratio (OR) of undergoing genetic counseling was 0.22 (95% confidence interval [CI], 0.12-0.40). When potential confounding factors including probability of carrying a BRCA1/2 mutation (Frank model), socioeconomic factors (ie, age, marital status, educational attainment, annual household income, and health insurance), and psychological factors (risk perception and cancer worry) were added to the model, the resultant OR was attenuated at 0.40 (95% CI, 0.14-1.08) and no longer significant. However, when additional attitudinal factors and primary care physician factors were included, the OR was again strengthened with a final adjusted value of 0.28 (95% CI, 0.09-0.89). The authors conclude that African American women are much less likely to undergo genetic counseling for BRCA1/2 testing than white women, with the reason for most of the disparity left unexplained.

However, these results must be interpreted with some caution. The actual number of African American patients was quite small (n = 15), particularly in view of the number of patients excluded from the analysis because of lack of family history (n = 17) or primary diagnosis of breast cancer (n = 224). Moreover, because African American patients comprise less than 10% of the case population, logistic regression modeling to examine the influence of explanatory factors on “case-ness” will offer little information on this small subset. Second, the appreciation of the genetic counseling disparity highlighted by this study is to some degree founded on the assumption that the rate of primary prevention counseling visits in these 2 populations (white and African American) should be approximately the same. However, Ashkenazi Jewish women comprised 30% of the white case population. Because these individuals have a many-fold increased risk of harboring a germline BRCA mutation^{9 - 10} (the carrier rate is estimated at 1/40 compared with 1/1000 in the general population), some degree of racial disproportion in the women undergoing counseling would be expected.

Beyond simply demonstrating the disproportionate utilization of highly specialized primary prevention counseling, the authors expose a more basic health care disparity in access to care in the United States. In the study’s urban-based (Philadelphia) hospital setting, in a city that is 43% African American and 45% white,¹¹ the population receiving specialized primary prevention counseling in cancer was only 6.7% African American. The reason for this may be an unfortunate synergism of access barriers to preventive care that is compounded by other systemic barriers in cancer care. Clearly, access to preventive medical care in the United States is by no means equally distributed. The National Healthcare Quality Report, released in 2003, and the National Healthcare Disparities Report found that one of the greatest weaknesses in the US health care system is the under-use of general preventive care.¹² For example, rates of regular visits for prenatal care, dental care, and routine health screening are all significantly lower in African American and other minority populations compared with white Americans,¹³ leading to poorer scores on health status indices, as well as decreased awareness of advanced primary prevention services that may be available. For cancer care and prevention, access issues are magnified further. In one report, poor or absent health insurance coverage was identified as one of the most significant barriers to receiving comprehensive cancer care.¹⁴ Treatments, diagnostic testing, and screening modalities have seen rapid cost inflations in recent years, while the number of insured Americans has steadily declined, with nearly 50 million Americans (15% of all Americans) currently without health insurance coverage.¹⁵ Thus, poor access to basic preventive health care, not to mention specialized cancer prevention such as genetic counseling and testing, remains an important barrier that must be eliminated to ensure equal high-quality care in the US health care system.

Perhaps a second important explanation for the racial disparities in genetic testing and the broader disparities of the American health care system can be appreciated through the findings of Armstrong et al⁸ and from results of the 2000 National Health Interview Survey.¹⁶ In the study, the authors show that in the 3 areas of perceived risk (BRCA mutation, breast cancer, and ovarian cancer), average-risk African American controls had consistently lower scores (a greater proportion reported “less than average” risk) than white counterparts. In the National Health Interview Survey, investigators found that while 49.9% of whites reported that they had heard of genetic testing for cancer risk, this proportion was only 32.9% among African Americans and 20.6% among Hispanics (men and women). The common thread in these 2 examples was the power of knowledge (personal and community-wide knowledge base of clinical cancer genetics) hailed by Bacon over 4 centuries ago. The average African American woman underestimates her risk of breast cancer and African Americans are on the whole less aware of genetic testing technology as a means of assessing personal risk. In focus group sessions conducted by our group at the University of Chicago, Matthews et al³ found that among a sample of African American participants with a strong family history of cancer, nearly half (48%) reported rarely discussing cancer-related issues with family members, and none had knowledge of breast cancer genetics, genetic counseling, or the BRCA genes.³ These findings of poor personal risk awareness as well as a limited knowledge base of cancer genetics together might then equate to lower utilization of genetic services by African Americans.

However, the function of knowledge base in genetic services utilization disparities is not limited to minority populations, or even to patients for that matter. The role of the physician in identifying and referring high-risk patients for genetic counseling cannot be underestimated. A carefully obtained family history remains among the most useful data in preventive medicine,¹⁷ but when important clues to familial cancer inheritance are missed because of poor knowledge of clinical cancer genetics, the chance that a high-risk patient will be referred for primary prevention counseling is already reduced. In a 1998 survey of 2250 internists, gynecologists, and oncologists from the Eastern United States, Doksum et al¹⁸ found that knowledge base in breast cancer genetics was relatively low overall, with only 25% of physicians answering all knowledge base questions in genetics correctly. Among oncologists, high knowledge base was found to be significantly associated with having discussed BRCA testing with a patient. In a more recent study (Physician Survey on Cancer Susceptibility Testing), Freedman et al¹⁹ found that only 51% of 1251 US physicians felt qualified to recommend genetic counseling to their patients. Among primary care physicians, the numbers were even lower (40%). Not surprisingly, physicians who felt confident and qualified to recommend counseling were much more likely (OR, 3.63; 95% CI, 2.17-4.86) to feel qualified to recommend testing.

For the American public and particularly for minority populations, the need for nationwide education to increase awareness of inherited cancer syndromes and the benefits of cancer risk evaluation is enormous in the post-genomic age. While increased awareness of risk may make individuals more likely to investigate their own family histories or to make those who are referred for counseling more likely to pursue these services, the true power of this knowledge may be even greater. In a follow-up analysis to the Physician Survey on Cancer Susceptibility Testing, Wideroff et al²⁰ reported that the factor most strongly associated with physician use of genetic services, more so than availability of services or physician self-assurance, is patient inquiry as to whether they can or should get tested (OR, 5.52; 95% CI, 3.97-7.67).²⁰ Thus, even if a physician fails to appropriately recommend genetic counseling, informed individuals may act as their own health care advocates.

Educational deficiencies in genetics among physicians and the American public are not restricted to breast cancer. In 2005, more than 150 000 Americans will be diagnosed with colorectal cancer.²¹ As many as 10% of these cancers are hereditary and may involve germline mutations in colorectal cancer susceptibility genes such as APC, the causative gene in familial adenomatous polyposis, and in the mismatch repair genes (MLH1/MSH2/MSH6) of hereditary nonpolyposis colon cancer. While genetic testing for these syndromes has been available for more than a decade, genetic counseling and testing for these individuals remain only a small fraction of the source of referrals to high-risk cancer clinics. Much of this disparity may be attributable to poorer physician knowledge of colorectal cancer genetics compared with breast cancer, although this hypothesis remains untested. The influence of women’s health initiatives and national advocacy campaigns on awareness of breast cancer should not be underestimated.

Ultimately, whether testing is performed for familial breast and ovarian cancer, familial colorectal cancer, or other familial cancer syndromes, the goals and benefits of testing are the same: to offer individuals a more precise estimate of future cancer risk so that risk may be modified appropriately. Prophylactic mastectomy and oophorectomy have clear risk-reducing benefits in mutation carriers.^{22 - 23} Intensive surveillance and chemoprevention with tamoxifen are also important elements of breast cancer prevention in women at high risk. For individuals with familial adenomatous polyposis, prophylactic colectomy is standard-of-care for management of the nearly 100% colorectal cancer risk in this disease, with additional screening for upper gastrointestinal malignancies also offered in affected individuals.²⁴ Even though colorectal cancer risk may be somewhat less in hereditary nonpolyposis colon cancer (70% by age 70 years), regular colonoscopy with polypectomy has been shown to reduce mortality and to be cost-effective.²⁵ In women with hereditary non-polyposis colon cancer, additional uterine and ovarian cancer screening also plays an important part in the aggressive cancer prevention offered to at-risk individuals.

The benefits gained from risk assessment, genetic counseling, intensive screening, as well as risk modifying behaviors, medications, and surgeries will remain unrealized for the majority of mutation carriers until efforts are increased to define and bridge the racial, ethnic, economic, and knowledge-based disparities that contribute to the unequal access and utilization of preventive medical services. Only by improving the ability to recognize and define cancer risk in all Americans can cancer control efforts be targeted effectively to reduce disparities in health outcomes.