Toward a More Just System of Care in Molecular Pathology

Abstract

Policy Points.

• American health care policy must be critically assessed to establish the role it plays in sustaining and alleviating the health disparities that currently exist in molecular genetic testing.

• It is critical to understand the economic and sociocultural influences that drive patients to undergo or forgo molecular testing, especially in marginalized patient populations.

• A multipronged solution with actions necessary from multiple stakeholders is required to reduce the cost of health care, rebalance regional disparities, encourage physician engagement, reduce data bias, and earn patients’ trust.

Context

The health status of a population is greatly influenced by both biological processes and external factors. For years, minority and low socioeconomic patient populations have faced worse outcomes and poorer health in the United States. Experts have worked extensively to understand the issues and find solutions to alleviate this disproportionate burden of disease. As a result, there have been some improvements and successes, but wide gaps still exist. Diagnostic molecular genetic testing and so‐called personalized medicine are just now being integrated into the current American health care system. The way in which these tests are integrated can either exacerbate or reduce health disparities.

Methods

We provide case scenarios—loosely based on real‐life patients—so that nonexperts can see the impacts of complex policy decisions and unintentional biases in technology without needing to understand all the intricacies. We use data to explain these findings from an extensive literature search examining both peer‐reviewed and gray literature.

Findings

Access to diagnostic molecular genetic testing is not equitable or sufficient, owing to at least five major factors: (1) cost to the patient, (2) location, (3) lack of provider buy‐in, (4) data‐set bias, and (5) lack of public trust.

Conclusions

Molecular genetic pathology can be made more equitable with the concerted efforts of multiple stakeholders. Confronting the five major factors identified here may help us usher in a new era of precision medicine without its discriminatory counterpart.

Pathology, laboratory medicine, and radiology serve as the foundation of the modern medical diagnosis, although these fields are often neglected in policy conversations about access to care. ¹ , ² When considering the role of diagnostic medicine in patient care, interactions with patients abound, but because laboratory workers and pathologists do not interact with patients face to face, their role may often be diminished. ³ The pathology laboratory handles every piece of tissue and every vial of blood or other bodily fluid and performs all tests and analyses on specimens that enable a clinical diagnosis. ⁴ The pathology laboratory thus is the window into, and the final verdict of, a patient's condition, and without it, medicine would cease to have its current level of precision and quality. ⁴

As pathologic diagnoses become defined on multiple diagnostic planes (i.e., symptomological, gross, histological, cytological, chemical, cytogenetic, and molecular), correlations can be drawn between pharmacological interventions, pathological subtypes, and clinical outcomes. ⁵ This level of diagnosis has led to the development of drugs targeting specific steps in pathogenesis, disease progression, and cellular immune response. ⁶ One of pathology's growing areas that has made this shift toward more personalized medical care, colloquially referred to as “precision medicine,” is molecular genetic pathology, often referred to as “molecular pathology.” ⁶

Molecular pathology is common in oncology, in which many solid organ tumors and also myeloproliferative and lymphoproliferative neoplasms are defined at the molecular level, and tailored cancer treatments have been designed for specific cancer subtypes or genetic alterations. ⁷ In addition, following recent advances in pharmacogenomics, many of the genes responsible for differences in drug metabolism among patients have been classified and characterized so that patients can be screened for genetic changes that may result in drug toxicity or futility. ⁸ Precision or “personalized” medicine has also been used for infectious disease testing and treatment, such as in HIV primary care settings in which human leukocyte antigen (HLA) sequencing and viral genome sequencing have shown clinical utility. ⁹ Finally, and most obviously, molecular testing is essential to diagnosing children with inherited or congenital disorders and diagnosing adults with genetic predispositions for illnesses. ¹⁰ These diagnoses direct primary care, alter disease screening schedules, and help guide long‐term planning and family planning. ¹⁰

To illustrate the growth in molecular genetic pathology, Medicare, the largest health care provider in America, paid for 2.1 million genetic tests in 2019, up fourfold from 2016 when it paid for approximately 627,000 genetic tests. According to the American Medical Association, as of 2020, one in ten Americans will have taken a genetic test ¹¹ , ¹²

With such substantive advancements in using molecular technologies to bring medical diagnostics and therapeutics into the “precision era,” molecular pathology has become more commonplace in clinics. But to avoid further widening the health care divide in the United States, the medical and policy communities must identify and address those features of molecular pathology that make it exceedingly susceptible to unjust distribution. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ Diagnostic molecular genetic testing is of particular interest as a cause of health care inequality for the following reasons: (1) molecular genetic testing is one of the most expensive diagnostic tests outside of radiology while simultaneously receiving some of the lowest third‐party payer coverage; (2) the test results can be adequately interpreted only through a large diverse data set that includes persons of differing ethnicities; and (3) genetic testing has a complex history with links to eugenics and discrimination. ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶

This article discusses these reasons, starting with patient cases highlighting the issue, identifying contributing factors, and defining actionable items for stakeholders who play critical roles in ordering, providing, funding, and/or developing molecular diagnostic technologies. Because the fields of pathology and laboratory medicine are not often explored in regard to bioethics, we earlier looked at the role of the laboratory, the personnel in the laboratory, and definitions unique to this field and then defined and organized them into a table for reference. ¹ Table 1 offers definitions more specific to the field of molecular pathology and the regulatory bodies involved in determining the reimbursement of molecular pathological testing.

Table 1.

Important Definitions in Molecular Pathology and Pathology Economics

Term or Abbreviation	Definition	Role in Molecular Pathology
CMS	Centers for Medicare & Medicaid Services	CMS plays a critical role in reimbursement to health care institutions as well as setting quality standards in the laboratory through the police powers granted to them by the 1988 Clinical Laboratory Improvement Amendments (CLIA).
CLIA	The Clinical Laboratory Improvement Amendments (CLIA) are the set of federal regulatory standards to which all diagnostic laboratories must adhere.	The CLIA prescribes the standards to which the laboratory must adhere. Although the CLIA regulations have been updated, these procedures and standards have largely remained unchanged since 1988.
Clinical equipoise	Clinical equipoise refers to allowable variations in providers’ medical decision‐making based on experience or preference.	Clinical equipoise is an allowable variation, which EBM often accounts for, but when new data suggest that one treatment or diagnostic test is found to be more beneficial or less harmful, a clinician's choice cannot be defended using clinical equipoise.
CPT	Current procedural terminology (CPT) is a set of numerical codes and modifiers serving as the language by which clinical, pharmacological, and laboratory protocols and expenses are translated uniformly to an insurer.	These codes describe clinical protocols and tests. Payment to a health care organization is typically tied to CPT codes. Molecular CPT codes have been revised several times since 1993.
DTC‐GT	Direct to consumer genetic tests (DTC‐GT) are tests that can be ordered and interpreted by the patient.	Many of these tests are reasonably priced but have varying levels of laboratory standards and may be misleading to the patient. Delivering the results directly to the patient does not always allow for counseling or appropriate follow‐up.
EBM	Evidence‐based medicine (EBM) is practiced when the most current and highest evidence research studies are examined and taken into account when considering patients’ management.	In pathology, EBM is critical to allow physicians to get the most accurate data on their patients’ conditions, but it is also important when considering test utilization and phasing out tests with limited clinical utility.
Moral pluralism	According to moral pluralism, a philosophical theory, each individual has their own moral framework and though individuals can have similar beliefs, there is no way to formulate one moral or ethical framework across multiple communities without compromise and agreement.	In policy, it is important to remember that the United States is morally pluralistic. When making policies, the moral priorities of some people are incompatible with the moral priorities of others. This theory explains why legislative movement has been slow and policy has changed little or not at all in recent years.
WGS	Whole‐genome sequencing (WGS) is the sequencing of all DNA in a sample in order to sequence and reassemble the entire genome to identify genetic changes in multiple genes.	Whole‐genome sequencing is not used widely in molecular pathology, but its use in certain situations has proved useful.
Race	Race is assigned a sociopolitical designation based on visible physical characteristics.	Although race is used as a physically differentiating factor, race often has little correlation with genetics.
Ethnicity	Ethnicity is also a social grouping of people. Different from race, ethnicity groups people by location, language, cultural traditions, religion, and the like, not physical characteristics.	Ethnicity is more genetically relevant than race but less genetically relevant than ancestry. Although ethnicity incorporates geography, it is often tied to cultural practices that are often more relevant clinically when analyzing risk factors, diet, and social support.
Ancestry	Humans inherit their DNA from their biological parents, who pass down their DNA to their biological offspring. Ancestry is the linkage of humans to one other through descendancy.	Ancestry is relevant to genomics because inherited genetic features can be linked to certain geographic regions owing to a limited gene pool or inbreeding.

Methods

Our initial literature review of PubMed used a MeSH term search (“Molecular Diagnostic Techniques” [Mesh] AND (“Ethics” [Mesh] OR “Policy” [Mesh]) AND “Humans”[Mesh] with English[Filter]), which yielded 49 articles published after January 1, 1990 (3 years before the first molecular CPT code was established). We removed twelve articles based on title alone because they either were duplicates or did not fit our area of inquiry. After analyzing the available remaining article abstracts, or occasionally (4 times) the complete article, we removed an additional 17 articles owing to their lack of relevance. The remaining 14 articles demonstrated a core set of themes that we further explored using an iterative, hermeneutic search approach. During the iterative literature search, we turned to Google Scholar for a more holistic review of extant literature. During these iterative searches of each of the identified domains, we explored specific areas of empirical research as more questions arose. These articles were limited to those written in English and published after January 1, 1990. We also excluded articles focusing on the discovery of molecular methods or methods specific to use in animals or if they were not related to access to molecular pathology or one of the related themes or topic areas. And finally, for financial reasons, eight articles located behind paywalls without Loyola University institutional access or interlibrary loan availability were excluded. This is a limitation of our study.

Results

The 14 papers we found in the initial MeSH search revealed significant barriers to molecular genetic testing access: (1) problems with reimbursement and the cost of the test, including offering justification to payers for testing and local/regional differences in coverage; ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ (2) providers’ knowledge, education, and recommendation to test; ²⁷ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁶ , ³⁸ (3) the importance of accurate databases; ³² and (4) the lack of patients’ trust in the medical community and medical research. ²⁸ , ³⁸ , ³⁹ Subsequent iterative searches revealed that the literature addressing the first point was so large that it was best broken into two discrete but linked ideas: (1) the cost to the patient and (2) regional/geographic disparities in reimbursement and technology adoption.

What Makes Molecular Pathology Susceptible to Unjust Distribution?

Cost to the Patient

Patient Scenario. A 38‐year‐old male presents to his doctor with new fatigue, easy bruising, and nosebleeds. He works full time for minimum wage. He had no serious medical conditions prior to presentation and has a high‐deductible health care plan to keep his monthly premiums affordable. His blood is drawn for a complete blood count. This laboratory test shows pancytopenia with rare immature cells, concerning for blasts. A presumptive diagnosis is acute leukemia, and the patient is referred to the hematology‐oncology service for a bone marrow biopsy to be performed that same day. The risks of the procedure, such as infection, bleeding, and pain are discussed with him. The patient is scared and does not ask about the price of the procedure, and the clinician does not bring up cost.

The procedure requires local anesthesia, and a skilled person performing the bone marrow biopsy. The interpretation requires a pathologist as well as various processing and ancillary tests carried out by a team of highly skilled laboratorians. The patient is then given a new diagnosis of “acute myeloid leukemia (AML) with mutated NPM1” and was also found to have mutated FLT3.

The patient has private insurance with a high‐deductible plan and 20% coinsurance. His doctor decides that next‐generation sequencing (NGS) is medically necessary according to current guidelines, as several genes must be interrogated to appropriately subtype AML. In this case, mutations in the genes NPM1 and FLT3 guide treatment decisions on the intensity of induction chemotherapy and whether this young patient should have hematopoietic stem cell transplantation. Unfortunately, as we will discuss later, the insurer considered NGS to be “unnecessary” experimental diagnostics and so would not cover it. The patient is thus expected to pay $23,250.00—with 31% of that out‐of‐pocket cost being a direct charge from the claim denial for NGS. This large bill is despite the patient being insured and seeking care in‐network (Table 2).

Table 2.

An Example of a Patient's Bill Reflecting the Exclusionary Nature of Molecular Pathology Pricing

Test	CPT Code	Covered by Insurance?	Charge	Discounts and Reductions	Charge After Negotiation/ Discounts	Deductible	Copay	20% Coinsurance	Amount Not Covered	Health Insurance Responsibility	Patient's Costs
Diagnostic bone marrow biopsy and aspiration	CPT 38222	Covered	$26,503.22	$1,503.22	$25,000.00	$10,000.00	$50.00	$2,990.00	—	$11,960.00	$13,040.00
Pathology interpretation of biopsy and aspiration	CPT 88305	Covered	$9,949.06	$2,049.06	$7,900.00			$1,580.00	—	$6,320.00	$1,580.00
Bone decalcification	CPT 88311	Covered	$900.00	$100.00	$800.00			$160.00	—	$640.00	$160.00
Karyotype / Chromosome analysis	CPT 88264	Covered	$796.00	$96.00	$700.00			$140.00	—	$560.00	$140.00
AML FISH (x7)	CPT 88374	Covered	$400.00	$0.00	$400.00			$80.00	—	$320.00	$80.00
CD34 immunohistochemistry	CPT 88342	Covered	$4,823.00	$1,023.00	$3,800.00			$760.00	—	$3,040.00	$760.00
AML flow cytometry	CPT 88189	Covered	$529.00	$79.00	$450.00			$90.00	—	$360.00	$90.00
Reticulin Stain including interpretation	CPT 88313	Covered	$1,206.00	$206.00	$1,000.00			$200.00	—	$800.00	$200.00
NGS 51+ genes	CPT 81455	Not covered	$7,200.00	—	—			—	$7,200.00	$0.00	$7,200.00
Total			$52,306.28	$5,056.28	$40,050.00					$24,000.00	$23,250.00
			Total Amount Billed	Total Discounts	Insurable Total After Discounts					Total Insurance Responsibility	Total Patient Responsibility

Amount billed by the hospital for service is in green. Amount negotiated between the hospital and insurance and also final prices are in blue. Based on the patient's specific plan, the patient is expected to pay the costs in red, and the insurer is expected to pay the costs in yellow.

The US health care system is uniquely wasteful and inefficient. The United States spends about twice as much on medical care as other industrialized nations do and yet does not show substantially improved outcomes for that additional expenditure. ⁴⁰ , ⁴¹ , ⁴² Although some of this mismatch should be blamed on American “lifestyle factors,” such as poor diet and other social determinants of health, a substantial proportion of this inefficiency is attributed to the high prices of medical goods and services like diagnostic testing, pharmaceuticals, and administrative costs. ⁴⁰ , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ Indeed, the opacification of pricing in health care has contributed to runaway costs. ⁴⁹ Despite legislative efforts to make prices more transparent, the publication of “chargemaster” prices has been completely unproductive, as the listed chargemaster price has little relationship to what health care institutions are ultimately paid for that service, to the cost incurred by the health care institution to provide the service, or to what bill ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ , ⁵⁴ , ⁵⁵ many patients ultimately receive. Cost sharing between patients and insurers has only added to cost complexity, with a substantial proportion of Americans unable to understand the financial implications of their own health care plan and their rising out‐of‐pocket expenses. ⁵⁶

Diagnostic molecular testing is not exempt from the ills of the American health care system. Moreover, billing in diagnostic molecular testing is perhaps the most exaggerated example of this financial dysfunction: diagnostic molecular testing is simultaneously one of the more expensive laboratory diagnostic tests and also one of the most commonly denied diagnostic insurance claims. ²⁵ , ⁵⁷ One study found that third‐party payers denied 77% of all NGS tests ordered, even though 64% of the results had a clinically meaningful impact. ²⁵ Part of the reason for this discrepancy is simply that diagnostic molecular testing uses fairly new technology. ⁵⁸ Some test methods, such as next‐generation sequencing (NGS), have only recently been recommended as a standard of care for certain diseases or circumstances, ⁵⁹ , ⁶⁰ , ⁶¹ and even though NGS shows a clear benefit and likely will become the standard of care for many more diseases in the next five to ten years, that evidence is still being collected. ⁶² , ⁶³ , ⁶⁴ , ⁶⁵ , ⁶⁶ , ⁶⁷

Current procedural terminology (CPT) coding and CPT‐associated payer reimbursement lag behind evidence‐based medicine by design, so it is somewhat unsurprising that payer reimbursement is not occurring for the newest molecular diagnostic tests, even though they are necessary for patient care. ⁶⁸ , ⁶⁹ However, tests ordered and performed but not reimbursed by a third‐party payer can result in high bills sent directly to the patient, or they can result in the testing facility's absorbing the cost. In the latter case, these absorbed charges will later be passed on to other patients via vague chargemaster price increases across numerous fields, not just diagnostic testing, to keep hospitals’ balance sheets from turning red. ⁷⁰

Although the relative newness of molecular diagnostics is a large hurdle to reimbursement, even older molecular technologies are poorly reimbursed because of badly fitting CPT codes, resulting in charges passed on to patients directly or indirectly (see Figure 1). The first molecular CPT codes were created in 1993 to match the technology of that time and were somewhat well suited as methods‐based descriptors. ⁷¹ As molecular testing expanded and the number of assayable analytes grew, however, these original methods‐based descriptors did not appropriately convey what was being tested, and payers were unable to evaluate the appropriateness of this testing. ⁷¹

Figure 1.

Factors Negatively Influencing Patients’ Trust in Molecular Pathology [Colour figure can be viewed at wileyonlinelibrary.com]

In addition, because the original 1993 methods were more similar to clinical laboratory testing than to anatomic pathology diagnostics, all the molecular codes were given a low fee schedule, which has remained even though newer tests require a pathologist's interpretation. ⁷² In 2012, CPT codes were modified, but not overhauled, to address both physicians’ inappropriate compensation for interpretive tests and to incorporate both analytes and methods into CPT coding for transparency to payers. ⁷² Although well intentioned, this patchwork of changes resulted in a tiering of genes that has created confusion for third‐party payers and has often meant no reimbursement for infrequently tested genes, regardless of their clinical utility. ⁷¹ , ⁷² , ⁷³ It also resulted in underpayment for claims using the G0452 code for physician's interpretation costs. ⁷² Together, these 2012 modifications of the 1993 system have resulted in broad underpayment for molecular diagnostics, with the costs passed on to patients either directly or indirectly. In 2015, instead of these 2012 issues being resolved, additional CPT codes were created to describe multiplexed panels, including large next‐generation sequencing panels. These new genomic sequencing procedure codes again clarified the nature of the test but have very high rejection rates, from 46% to 79%, and thus this change also greatly contributes to the underpayment of diagnostic services. ⁷¹

Direct‐to‐consumer genetic testing (DTC‐GT) has popped up as a response to the molecular diagnostic billing insanity. These tests are performed without a physician's orders but with transparent pricing that reflects the test's actual cost. The major benefit of DTC‐GT is the lack of surprise billing from the opacification and inflation because of the third‐party payer system. Physicians do not order DTC‐GTs, however, and patients may choose inadequate or inappropriate testing. ⁷⁴ For example, some DTC‐GCs evaluate the BRCA genes for an inherited predisposition to breast and ovarian cancer. Although these tests reliably pick up some mutations in BRCA1 and BRCA2 genes, they also miss many common risk‐increasing mutations, including mutations in the very BRCA genes examined. ⁷⁵ , ⁷⁶ Patients with a family history of breast or ovarian cancer would typically never have a physician recommend such DTC‐GTs because of the high number of false negatives. ⁷⁷ Instead, patients with a concerning family history are directed to more comprehensive genetic testing. Understanding results of genetic tests can be complicated, and a negative or positive test result for a genetic alteration requires complex interpretation from a physician and detailed discussions with a genetic counselor to explain risk. A physician or advanced practice provider (APP) should be available to discuss any necessary additional testing, screening programs, or treatments because of genetic test results, but DTC‐GT misses all these critical aspects. ⁷⁸ , ⁷⁹ We will come back to DTC‐GT in subsequent sections on database bias.

Another issue with the typical method of physician‐ordered genetic testing is that somatic genetic testing, which is testing for DNA alterations that were not inherited but acquired in tissues, is often ordered as an unexpected add‐on test to biopsy procedures or the surgical resections of tumors. This test is often not explicitly requested by patients or treating physicians/APPs. Although patients undergo a well‐established consent process for the physical risks of the biopsy or surgical procedure, the financial risks of that same procedure are not often explicitly discussed within the process of informed consent. ⁸⁰ This is because each patient has typically entered into a contract with a third‐party payer that has set patient cost sharing, which cannot be known by the health care facility performing the procedure.

Moreover, each different third‐party payer has negotiated down the health care facilities’ chargemaster price, and thus the consenting physician would have to have broad knowledge of every possible payer's annually changing contracts for each of thousands of line items on the chargemaster—contracts that the health care facility often intentionally hides from the provider. Finally, the consenting provider often does not yet know what chargemaster line items will be required. For instance, the patient could have a severe complication during surgery that requires blood transfusion, antibiotics, or other therapies. Or the procedure could go well, but a biopsy from the patient could lead to unclear results on histology alone, requiring the pathologist to order additional ancillary testing, such as immunohistochemistry and molecular testing, to make a diagnosis. In sum, even though informed financial consent should ethically be obtained before providing a service, in the American health care system it cannot be obtained.

All these factors together lead to many artificially expensive molecular diagnostic tests, which are either certainly necessary or likely necessary for care, but are reimbursed only partially or are rejected outright by third‐party payers. These inflated costs are then passed on to patients directly or indirectly and often by surprise. Assuming that the molecular assay was ordered by the treating physician/APP upfront and was not the result of a secondary diagnostic workup, the best attempt at an informed financial consent would be to point to the bloated chargemaster cost as the maximum out‐of‐pocket expense that a patient had to pay for a particular molecular test. Although wealthy patients can afford the potential expenditure and so are more likely to consent to the financial risk, patients with fewer resources will potentially forgo vital testing. ⁶⁸ , ⁸¹ , ⁸² , ⁸³ , ⁸⁴ , ⁸⁵ , ⁸⁶ Effectively, then, molecular diagnostic testing has become intrinsically tied to class in America: biomarker testing is available for those who can afford to lose thousands or tens of thousands of dollars, but not for those who cannot take this risk.

Regional Disparities

Patient Scenario. A medium‐sized hospital in a small city in middle America is seeing an increase in cancer patients. The city has an aging population, and the other nearby hospital has recently closed. The hospital's hematology‐oncology specialists are interested in improving the turn‐around‐time for biomarker testing in order to help treat their cancer patients earlier with targeted therapy, potentially reducing the cytotoxic chemotherapy requirements. Currently, all biomarker testing is being sent to a company on the East Coast, with results arriving about 3 to 4 weeks after the biopsy. The hospital's medical director has molecular genetic training and extensive experience designing laboratory‐developed tests/procedures in the CLIA‐certified lab, and is interested in validating an analogous test. Although the hospital may see a return on investment in 10 years—assuming third‐party payers cover the test—the upfront capital costs are nearly $1 million, and the hospital knows that reimbursement for NGS is suboptimal in its region. In addition, the East Coast commercial lab that is currently receiving patient samples has an NGS payment assistance program to help when bills are rejected by third‐party payers so that patients do not receive a large bill. For these reasons, the hospital administration has decided that the institution will not invest in molecular pathology and will continue to send NGS testing more than 900 miles away. Over the next year, several patients with advanced cancer will die while waiting for their biomarker testing to return or after their tissue was exhausted by attempting failed NGS testing.

The vast majority (85%) of American patients with a cancer diagnosis do not travel to receive care but are treated in their home community, which makes the geographic distribution of cancer‐fighting resources, like diagnostic molecular testing, highly impactful. ⁸⁷ But just as only some patients are willing to risk receiving an expensive surprise bill, only some hospitals and institutions are willing to risk being underpaid or not paid. As stated earlier, NGS is the most expensive molecular technique and also the most poorly reimbursed. ⁸⁸ , ⁸⁹ , ⁹⁰ , ⁹¹ , ⁹² , ⁹³ , ⁹⁴ The Centers for Medicare & Medicaid (CMS) has tried to outline broad indications and limitations of coverage but has missed critical diseases and circumstances in its determination, such as patients with diseases not requiring traditional staging, including diffuse gliomas and acute leukemia. ⁹⁵ According to CMS's national coverage determination (NCD), the diagnostic NGS laboratory must have the approval of or clearance from the Food and Drug Administration (FDA), thereby eliminating the mandate for well‐regulated, CLIA‐certified laboratory‐developed tests and or laboratory‐developed procedures (LDPs) to be reimbursed unless local coverage determination approves reimbursement. ⁹⁵ Together, owing to the costs of obtaining FDA approval/clearance for laboratory‐developed procedures, these policies favor large, well‐funded, administratively endowed institutions. ⁹⁶ , ⁹⁷ , ⁹⁸ Smaller, less‐resourced hospitals and academic centers with high‐quality LDPs, with oversight by the federal government via CLIA, are unable to meet the duplicative and burdensome FDA government regulatory requirements. ⁹⁸ , ⁹⁹ All American FDA‐approved companion‐diagnostic commercial institutions are located in major cities and not in rural communities, towns, or small cities, with the singular exception of Marshfield, Wisconsin, home of PreventionGenetics, LLC. ¹⁰⁰ , ¹⁰¹ , ¹⁰² , ¹⁰³ Some academic centers have also committed to NGS sequencing, often despite insufficient revenue, through institutional support for research, philanthropic efforts, and clinical trials. Likewise, these academic centers are typically located in large cities.

In addition to being widely reimbursed through Medicare with the FDA's approval, large genetics companies are more likely than small hospitals to be reimbursed as well by all payers, including private insurance. Insurance preauthorization is commonly required by payers for genetic tests to help eliminate excessive spending and misordering or outright fraud. Preauthorization enables insurance companies to decide how necessary a medical treatment is so they can deny the procedure/test/care before the service is provided, by reviewing the paperwork from the health care facility. ¹⁰⁴ , ¹⁰⁵ Industry partners on the coasts and well‐funded academic centers also have support staff to maximize payer reimbursement by means of the onerous preauthorization approval process, whereas small community health care facilities do not have sufficient staff or resources. ¹⁰⁶ , ¹⁰⁷ , ¹⁰⁸ Additionally, nearly all commercial laboratories offer some form of NGS payment assistance program for rejected insurance claims, so that patients will not be surprised by huge bills. Although the funds for these programs vary, some commercial laboratories subsidize payers’ nonpayment by selling deidentified patient data to researchers or other companies, a policy that hospitals have not adopted. ⁶⁸ , ¹⁰⁹ In sum, the financial incentives and policies driving reimbursement have resulted in consolidating diagnostic genetic testing in commercial labs and large academic centers, excluding small towns and much of middle America.

Delayed testing can result in suboptimal treatment. In an American Medical Association study, 28% of doctors state that waiting for insurance preauthorization approval has led to a serious adverse outcome for at least one patient (death, hospitalization, or disability). ¹¹⁰ Assuming that preauthorization is not required or moves quickly, most commercial labs have a two‐ to three‐week turnaround time from receipt of the specimen, which can be lengthened when the specimens sent are not suitable for testing and additional material needs to be shipped. This partially explains why 37% of oncologists reported usually waiting more than three weeks to receive test results. ¹¹¹ Recently, strong evidence suggested that oncologists should wait for EGFR mutation results, because EGFR‐specific mutation treatment (tyrosine kinase inhibitors) was associated with superior outcomes compared with conventional chemotherapy, which in the past might have been started if the biomarker testing were delayed. ¹¹² Indeed, delays can be deadly: 25% to 30% of stage IV non‐small cell lung cancer patients die within three months. ¹¹³ This explains why the American Society of Clinical Oncology (ASCO) recommends that all lung biomarker testing (a type of molecular diagnostic testing) be returned within one to two weeks, a feat easiest achieved with local laboratories and the elimination or reduction of preauthorization insurance policies. ¹¹¹

The democratization of diagnostic molecular testing across multiple regions not only would improve turnaround times but also would allow better control of specimen handling, prioritization of assays, more accuracy in variant calls, and the ability to revise pathology diagnoses when results come through the laboratory instead of going directly to the clinic, as is often the case for tests sent directly to commercial labs.

Again, diagnoses are increasingly relying on genetic alterations found by molecular testing; results from comprehensive genetic testing can modify or completely change a diagnosis. As just one example, at a single institution over three years, nine patient diagnoses were completely revised after NGS testing: a formerly diagnosed malignant peripheral nerve sheath tumor was determined to be a melanoma; a lung adenocarcinoma (non‐small cell lung cancer) was found to be a prostate adenocarcinoma metastasis; and an epithelioid sarcoma was diagnosed by genetic alterations after previously being deemed a likely non‐small cell lung cancer. ¹¹⁴ Memorial Sloan Kettering Cancer Center (MSKCC), a hospital that consistently ranks in the top three cancer centers in America, performs molecular genetic testing on‐site for numerous reasons, including the ability to better control specimens and the ability to review cases in the appropriate clinical context. ¹¹⁵ The molecular department does not sign out NGS cancer cases without first reviewing clinical history through electronic chart review, as low‐level alterations are likely to be missed in low tumor samples, and molecular diagnostic errors are likely to arise without the appropriate clinical context, a component that commercial labs struggle to integrate without access to patients’ complete charts. ¹¹⁶ Additionally, at MSKCC scant tissue is triaged by molecular pathologists in order to maximize the chance of finding actionable biomarkers. ¹¹⁷ A sample that could be sent to a commercial NGS lab and rejected as insufficient or, worse, exhausted on a failed run, can be salvaged by on‐site triaging to other in‐house molecular assays, such as PCR, with very low tissue requirements. ¹¹⁷ High‐quality molecular diagnostic testing, as seen at MSKCC, should not be available only to those living in large cities, those near a cancer center, or those able to travel for medical care.

The disparity between rural and urban populations in cancer care has already been detailed extensively elsewhere, with mortality rates higher in community hospitals and nonmetropolitan areas when compared with cancer centers and cities. ⁸⁷ , ¹¹⁸ , ¹¹⁹ , ¹²⁰ , ¹²¹ , ¹²² The cause of the widening disparity and poor outcomes for non‐city dwellers is multifactorial with a high proportion of high‐stage and screening preventable cancers, a higher proportion of underinsured and uninsured people, an increase in tobacco‐related cancer types, and fewer available diagnostic services and treatments, including a marked decrease in clinical trial enrollment. ⁸⁷ , ¹¹⁸ , ¹¹⁹ , ¹²² Limited access to diagnostic services, including pathology ancillary testing and new technology, such as described here, has already been cited as a contributing factor. ¹²⁰

As mentioned, for some NGS tests and circumstances, regional/local Medicare administrative contractors (MACs) still determine whether NGS will be covered. Medicare hospital and outpatient claims, including laboratory testing, are processed through contracted companies (MACs), which hold regional territories. For example, Novitas is a company that oversees a region on the East Coast (Delaware, Maryland, New Jersey, Pennsylvania, and Washington, DC) and a southern region (Arkansas, Colorado, Louisiana, Mississippi, New Mexico, Oklahoma, Texas), and the company Palmetto oversees Tennessee, Georgia, Mississippi, South Carolina, North Carolina, West Virginia, and Virginia. ⁷² MACs determine whether to approve or deny claims without national coverage determination, including all LDPs and diagnoses not covered by the NCD. Some MACs are more likely to reject molecular test claims than other MACs are. ⁷¹ , ⁷² This variation in molecular test payment approval results in some regions of the country with more affordable genetic testing and correspondingly more accessibility to genetic testing and ultimately higher‐quality personalized medical care. Although many MACs are adopting Palmetto's system of coverage determination for genetic testing, some differences in access and affordability by LCD may contribute to the known regional disparities in health outcomes. ¹²³ , ¹²⁴

Lack of Provider Buy‐in

Patient Scenario. A 57‐year‐old male patient with a history of localized non‐small cell lung cancer (NSCLC) was treated by resection alone ten years ago in a small community hospital. Unfortunately, after receiving a PET scan at his most recent oncology follow‐up, the patient was found to have non‐resectable, metastatic NSCLC. The treating physician told the patient that the only available option was systemic, cytotoxic chemotherapy. If the physician could obtain a biopsy for the molecular subtype of the tumor, she might be able to recommend safe, less toxic, targeted chemotherapeutic agents that could treat the metastatic NSCLC.

Patients across the United States receive vastly different levels of care when being treated for cancer. Although a laboratory's level of diagnostic capabilities and a person's insurance coverage play roles, the oncology physician's/APP's awareness of new research, clinical trials, and new standards of care also is important. ¹²⁵ , ¹²⁶ In one study of disparities in germline (inherited) and somatic (acquired) genetic testing in ovarian cancer patients, the physician's/APP's lack of awareness and lack of recommendations for testing were substantial barriers to optimal care. ¹²⁷ A physician's or APP's lack of knowledge of current guidelines may not reflect a deficiency in their skills or abilities but may point to underlying systemic issues in the communication, delivery, and support of these topics.

Medicine today is centered on the concept of evidence‐based medicine (EBM). ¹²⁸ This type of medical care utilizes the newest research to deliver the highest‐quality medical care and thus is always changing, requiring health care providers to stay abreast of changes after they complete their formal training. ¹²⁸ This is why continuing medical education (CME) and/or ongoing examinations are necessary for them to maintain their licensure. Even so, these CME licensure requirements may not be sufficient. Workplace interactions with colleagues, on‐the‐job database “research,” clinical teaching, and attendance at conferences were strongly valued by physicians as methods of “lifelong learning” and are not always reflected in all certification requirements for maintaining licensure. ¹²⁹ The barriers in radiology to continuing medical education include insufficient time or financial support, lack of interest or perceived benefit, and lack of awareness of CME opportunities. ¹³⁰

Maintaining current knowledge of oncology care is even more challenging, with 30 new cancer drugs, 45 new adjunct therapies, and 1 biosimilar being approved in 2020 alone. It is understandable why some physicians/APPs stick to the treatment methods with which they are familiar. ³⁴ Suggested methods for keeping oncologists up to date include flexible learning environments like taking online courses, testing physicians to demonstrate areas of weakness resulting in motivation for growth, and providing updated educational resources as soon as possible through less traditional learning media to prevent a lag in knowledge. ¹³⁰ , ¹³¹ Hospitals and health care institutions can also help eliminate general barriers by providing off‐service time for attending conferences, taking advantage of educational opportunities, funding CME, creating an environment of inquiry with structured peer education/peer teaching, and improving the marketing of CME opportunities. ¹³⁰

Interdepartmental communication can also help with continuing education. Multidisciplinary tumor boards allow for bilateral transfer of information between the pathologist and oncologist or patient‐facing physician/APP. Although unidirectional, the pathology report is also a useful source of updated information for treating physicians. ¹³² Since pathologists have a great level of detailed knowledge of pathologic subtypes and their meaningfulness, and because the pathology report is typically a free text, it can function as a space for the pathologist to explain the diagnosis and as a location to provide recommendations for ancillary testing, such as diagnostic molecular testing, and to integrate the results after testing. ¹³³ This is beneficial to clinicians and also patients when they receive their pathology reports directly, sometimes ahead of the explanatory call from their care team. ¹³⁴

There is evidence in regard to inherited genetic testing that genetics‐trained doctors and genetic counselors with a master's degree provide more appropriate medical care and improve access to testing. ¹³⁵ Although there has been a push with more genetic testing being ordered by other physicians and APPs, the role of staff with such medical expertise should not be diminished. ¹³⁵ Improving reimbursement for genetic counseling appointments and ensuring adequate training programs to produce expert providers is important as well. ¹³⁵ Health care facilities should take advantage of their providers’ expertise by supporting peer‐to‐peer education, as described earlier. ¹³⁵

Bias in Databases

Patient Scenario A. A white male of European descent has his stage IV tumor sequenced for targeted therapy. To decrease costs by half, the patient does not have matched normal tissue or blood sequenced. Many variants are identified in the patient's tumor (a variant is a change or difference between the sequenced DNA and the human reference genome). Using databases of common benign variants, such as single nucleotide polymorphisms (SNPs), only two differences from the reference genome are reported, and the rest are conclusively identified as benign, non‐disease‐causing changes. The two different mutations were found in two different genes. Both mutations are known pathological variants in this patient's tumor type. Targeted therapy, available for only one of the two variants, is initiated. No variants of uncertain significance (VUS) are identified.

Patient Scenario B. A black male of African descent has his stage IV tumor sequenced for targeted therapy. To decrease costs by half, the patient does not have matched normal tissue or blood sequenced. The sequenced DNA has many differences from the reference genome. Although databases of common benign variants identify two‐thirds of these alterations as SNPs, this proportion is not reported. Of the remaining third, which is five variants, two known pathological variants are present, and only one of these two can be targeted. The remaining three changes are considered to be VUS. One of these three VUS is very concerning because it is located in a gene where some acquired mutations are associated with a poor clinical outcome in the patient's tumor type. The patient's oncologist is worried because the VUS may indicate that the patient will not respond to targeted therapy and the treating oncologist is unsure whether the patient should have more aggressive chemotherapy or, alternatively, be given palliation and/or hospice care. Note that the databases used to identify SNPs contain fewer individuals of non‐European backgrounds. The concerning VUS in this patient's report is likely a common single nucleotide polymorphism in an ancestry group not sampled in the SNP database, and it has no pathologic or therapeutic impact. Unfortunately, his treating oncologist does not know about the database bias and believes that this VUS could be pathogenic.

In oncology care, NGS is often used to find mutations that were acquired and found only in the tumor tissue, which may function as a target for companion therapies. But NGS can, of course, also test nontumor tissue to identify genetic mutations that a patient carries in all of his or her cells, including genes predisposed to cancer. Sequencing both tumor tissue and matched normal tissue, such as normal blood, can help distinguish germline (inherited) mutations from somatic (acquired) mutations in cancer. Germline mutations are found in both tumor tissue and matched normal tissue, whereas acquired somatic mutations are found in the patient's tumor tissue only and are not seen in the matched normal tissue. Matched normal tissue also helps identify benign SNPs and other differences from the reference genome that are of no consequence but are simply found in patients with different ancestries or genetic makeup than those of the reference genome. Like inherited pathogenic mutations, benign SNPs are also seen in both the patient's tumor tissue and matched normal tissue. If an alteration is seen only in tumor tissue, it is unlikely to be a SNP. In this way, matched sequencing can be helpful in distinguishing SNPs from somatic mutations that function as biomarkers.

Unfortunately, sequencing matched normal tissue with tumor tissue requires twice the sequencing cost, and few institutions can justify that cost, especially given reimbursement rates. Instead, usually only the tumor tissue is sent for NGS, and variants identified by somatic NGS are checked against publicly available databases. Alterations seen in in more than 1% of the healthy database population are likely SNPs and can be reported as such (a benign variant) or eliminated from the report of pathological and potential pathological variants altogether. Given the differences in SNPs seen in different ancestries, a diverse database of SNPs is key to appropriately cataloging variants. If fewer than 1% of Europeans have a particular polymorphism, but many more than 1% of those of non‐European ancestry, and a SNP database is almost exclusively populated with healthy samples from people of European ancestry, that alteration will not be appropriately stratified as a SNP when found in a patient of non‐European descent.

In addition, other non‐SNP databases are often used in NGS variant interpretation. Some of these databases focus on germline alterations, and others focus on somatic alterations. Some of these databases are populated by published research, and others synthesize clinically sequenced patient samples or both. For example, COSMIC is a free publicly accessible database that catalogs somatic mutations reported in human cancer research literature, including research done retrospectively on patient samples sequenced as part of routine care. ¹³⁶ These databases can supply confidence when classifying a variant as pathogenic or likely pathogenic using up‐to‐date evidence. Like the SNP databases, these mutation databases also work best when they are diverse, but they often have biases coming from underenrollment in research or clinical sequencing by certain patient populations.

A 2021 study from St. Jude Children's Hospital found that approximately 15% of parents/guardians refused pediatric NGS for their child with cancer. Families of black children were significantly more likely to refuse testing than were families of white children, most commonly citing concerns about discrimination and a general feeling of being overwhelmed. ¹³⁷ This study is consistent with prior studies showing that black families are less likely to consent to genetic research for a range of factors including generally negative attitudes toward clinical trials, religious beliefs, lack of access to information about clinical trials, and structural barriers to health care access in general. ¹³⁸ , ¹³⁹ For standard‐of‐care genetic testing, not in the context of a clinical trial, disparities between black and white patients are also found in genetic testing, although some of this may be related to provider bias in referral, a common theme noted in prior literature on research involvement by race and ethnicity. ¹⁴⁰ , ¹⁴¹ Separately, DTC testing companies also sell data for use by drug companies and researchers and have a disproportionate representation of genomes from white participants, potentially furthering the bias. ¹⁴² , ¹⁴³ , ¹⁴⁴ , ¹⁴⁵ Whatever the reason, whether providers’ nonreferral, patients’ refusal, or enrollment bias, the result is the same: many SNP and mutation databases significantly underrepresent non‐European populations.

Partly because of an underrepresentation in databases used to classify a variant as benign (such as a SNP) or pathogenic (a disease‐causing mutation), when minorities do receive genetic testing, they are more likely to have a variant of uncertain significance. ¹⁴¹ Patients with VUS experience more cancer‐related distress and uncertainty, and physicians of patients with VUS also experience uncertainty. ¹⁴⁶ , ¹⁴⁷ An article in the New England Journal of Medicine reported that patients of African ancestry were not only more likely to be told they had a VUS than were their European‐heritage counterparts, but they were also more likely to be told incorrectly that they had a mutation that could lead to sudden cardiac death. ¹⁴⁸

We should pause and mention here that although race is tied to ancestry and ethnicity, race itself is a social, not a genetic, category. Interracial genetic changes are only a small fraction (<1%) of the total genetic differences among different human genomes and therefore cannot be used to imply the genetic inferiority or superiority of a particular group of humans. ¹⁴⁹ This small, otherwise inconsequential, fraction of benign changes at most merely leads to issues during the genetic interpretation of results from large‐panel sequencing when the interpretation requires comparing dissimilar groups of patients and databases or when identifying new disease‐causing mutations in research. ¹⁵⁰

Lack of Patient and Public Trust

Patient Scenario. An African American female presented to an urgent care center with several hours of severe and localized chest pain with shortness of breath. She was asked to wait in her car due to the COVID‐19 pandemic, pending screening for COVID‐19 in order to be let into the care center. After one hour of waiting, without seeing a provider or having her vital signs evaluated, she was directed to a different urgent care center because of staffing issues. At the second urgent care center, she was greeted with overt frustration and contempt for being referred from the first clinic. She was again asked to wait in her car for COVID‐19 testing before being let into the clinic again, without any of her vital signs measured. One hour of waiting in her car at the new clinic, and more than two hours from initial presentation to the medical community, she still had not been seen by a provider for chest pain. Her husband, a medical doctor, knew the serious potential causes of chest pain and that his wife should have been seen long ago according to the current standard of care. Accordingly, he called the clinic from the car several times, demanding immediate attention. ¹⁵¹ Eventually a nurse came to the car and used the COVID‐19 nasal swab in a retaliatory manner to intentionally harm the patient. Thirty minutes passed before the patient was declared COVID‐19 free, and a medical doctor came to evaluate the patient, still in her car. The medical doctor minimized the patient's pain and determined, without any imaging or vital signs, that the symptoms were due to postnasal drip.

At the husband's insistence, the medical doctor flippantly agreed to order a chest X‐ray. The patient was allowed to enter the facility, and a pulse oximeter was administered, but the doctor disregarded the results stating, “They don't work on colored fingers,” and the X‐ray results showing signs of a potential pulmonary embolism or pneumonia was ignored. The patient was diagnosed with “congestion” and sent home. Over two days after the initial presentation, the patient presented to a third urgent care center, where she finally got the current standard of care. A CT scan, echocardiogram, and full blood work were ordered, and these tests found a pulmonary embolism, a potentially fatal blood clot, which had been left untreated for so long that it had caused the death of lung tissue. The patient spent the next four days in the hospital. This story is summarized from a true patient experience initially published in Forbes in July 2021. ¹⁵²

When interviewed, the patient acknowledged that her mistreatment and the clinic's malpractice were almost certainly related to her race. ¹⁵² The health care team's minimization of pain and disease severity, as well as the lack of empathy and even outright malice, are never overtly racist. The nurse never says, “I am ignoring and hurting you intentionally because you are Black” but each action and inaction conveys the same message in a disguised and subtle way.

Trust is the foundation on which the physician‐patient relationship is built. According to the Kaiser Family Foundation (KFF), Black Americans, compared with White Americans, place less trust in doctors, hospitals, and the health care system as a whole. ¹⁵³ This is not surprising, given than one‐third of Black Americans reported personally experiencing discrimination while receiving health care. ¹⁵⁴ Their mistrust extends to medical research. One study requesting blood for pharmacogenomic research found that 12% of African American patients declined to submit a blood sample, citing a mistrust of genetic research. ¹⁵⁵ A lack of trust is seen not only among Black Americans. Patients’ general trust of their physicians has declined over time, partly because of better access to information about science and medicine, which is somewhat paradoxical. ¹⁵⁶ Genetic testing reflects the mistrust of the health care system as a whole, with some specific concerns about DNA testing. ¹⁵⁷

DNA is the blueprint for all the cells in our body. It has a remarkable ability to predict what diseases we may get and even tendencies toward behavior and addiction. ¹⁵⁸ , ¹⁵⁹ Newer genetic testing is able to identify many epigenetic changes in addition to differences in DNA itself. Epigenetic changes are chemical modifications, essentially tags, on the DNA from our interaction with the environment, which help turn up and down gene expression. Thus, with some types of genetic testing, we can “see” the interplay of “nature” and “nurture.” ¹⁶⁰ With such an intimate view of who we are, and who we may become, it is unsurprising that many people are concerned about genetic privacy and discrimination. ¹⁶¹ Submitting DNA to laboratories for research or clinical purposes requires a degree of trust beyond that for other types of research and clinical testing.

One particularly important legal case that has altered the course of genomic research and implementation of precision medicine techniques is that of the Havasupai Tribe v. Arizona Board of Regents. ¹⁶² In this case, researchers from Arizona State University offered the Havasupai people, a small, isolated community living in the Grand Canyon, a proposition to help the tribe better understand how their genetic makeup predisposes or protects them from developing diabetes. Unfortunately, the tribe was not told that their samples would be used in future research projects, ranging from studying the genetic risk of developing schizophrenia to studying the potential migration of these people from Eurasia. The Havasupai people not only found these studies to be offensive, but they also went against their religious beliefs and caused a serious sense of betrayal between the Havasupai people and the biomedical research community. This betrayal is so severe that this community will no longer participate in human medical research. In fact, this mistrust has become pervasive in many Native American communities, with the Navajo Nation having called a moratorium on genetic research involving its people since 2002 and a large proportion of its people reporting that they are unsure if genetic research on Navajo people should resume. ¹⁶³ Although this is only one case, it is important because it illustrates how one injustice can affect an entire community and/or patient population.

As mentioned previously, the lack of inclusion in genomic databases is an ongoing source of bias against certain communities. ¹⁶⁴ , ¹⁶⁵ Another inclusionary issue is the chiasm between researchers and their study participants. A lack of community input during experimental design has been pervasive, with only recent attempts to include patient advocates in the design. ¹⁶⁴ , ¹⁶⁵ , ¹⁶⁶ , ¹⁶⁷ , ¹⁶⁸ Errors resulting from biases involving experimental design or from missing data from lack of participation can ultimately result in false conclusions and even more mistrust of the medical and scientific communities. Artificial intelligence (AI) algorithms that learn from databases lacking in diversity can also perpetuate biases in diagnostic decision trees (Figure 2). ¹⁶⁹

Figure 2.

The Cycles of Mistrust and Bias in Molecular Medicine [Colour figure can be viewed at wileyonlinelibrary.com]

Looking Forward

Clearly there is much work to do. Billing and reimbursement need to be changed. Because of the broad underpayment for molecular laboratory testing by third‐party payers, health care institutions have shifted charges to underinsured or uninsured patients. In turn, this shift has created disincentives for hospitals to invest in laboratory technology and has also contributed to disparate outcomes for patients of differing economic classes and living in different regions. ¹²⁵ , ¹³⁹ It is uncertain whether changes in billing and reimbursement require overhauling the entire third‐party payer system (i.e., universal health care), more extensive national coverage determination policies, or revisions of just the molecular CPT codes and associated fees with a minor refinement of existing policies.

This situation is not all negative, though. An example of a partial policy success is the Affordable Care Act's Medicaid expansion program: participating states have improved access to BRCA1 and BRCA2 inherited genetic testing overall, with no cost sharing for qualified patients. ¹⁷⁰ , ¹⁷¹

But because systematic change is difficult and tedious, before any of these policy changes go into effect, physician leaders can help patients today by participating in expert working‐groups like the ABIM's Choosing Wisely campaign or other specialty‐specific working groups. ¹⁷² These working groups should be explicit in their EBM recommendations for which diseases and circumstances and which kinds of testing require molecular testing. Expert physicians should take a cost‐conscious approach when considering the evidence and revise their recommendations when appropriate and in a timely manner. Leaving such determinations to individual insurance providers and Medicare, without explicit support from experts, allows ambiguity and ultimately the ability for third‐party payers to reject claims.

Expert physicians should also get involved in teaching and outreach to their peers, including primary care providers, helping them understand the indications and analysis of molecular tests. As we stated, expert working groups should be explicit in their EBM testing recommendations. Concerns about pigeonholing under‐resourced health care facilities with requirements for NGS in the diagnosis and management of diseases are reasonable but should not stop the recommendation if it is truly required. Finally, strong interprofessional relationships between patient‐facing providers and laboratory medicine physicians can help direct clinical decision making, avoid wasted laboratory spending, and strengthen diagnostic algorithms.

Regaining the trust of minority communities will likely be the hardest hurdle. Researchers and health care providers inflicted many real harms on communities in the past, and these harms will not be quickly forgotten. Given the direction of rhetoric in the national sphere, healthy skepticism is still warranted. Institutional review boards for human research should be vigilant and continue to protect the rights and welfare of all subjects. Ethics training for researchers and clinicians should be required when working with genetic data. Research funding should encourage the development of treatments and tests that help patients of all ancestries, not just those of European heritage because they are the easiest to enroll. Providers should stay vigilant of their own biases to make sure they are offering the same resources and opportunities to all their patients. Only by demonstrating commitment to the well‐being of all people can the medical and research community hope to slowly rebuild trust.

Conclusions

Molecular testing has moved firmly from the research realm into the contemporary practice of evidence‐based medicine. It is showing clinical utility in primary care and not just specialty care or in esoteric diseases and syndromes. General practitioners, who are not specialists in genetics, are ordering this testing more often, and some patients are requesting it more often. ¹⁷³ , ¹⁷⁴ , ¹⁷⁵ , ¹⁷⁶ Given this growth of utility, now is the right time to ask: is this testing being provided equitably? The answer to this question now is a resounding no. A patient's socioeconomic class, race/ethnicity/ancestry, insurance provider, geography, and physician all contribute to his or her access to appropriate care. To reach a more equitable system of molecular pathology distribution in the United States, many groups of people, including providers, payers, laboratorians, pathologists and laboratory directors, hospital administrators, researchers, bioethicists, and policymakers must take action (Table 3).

Table 3.

Actionable Items for Stakeholders in Molecular Pathology

Stakeholder	Actionable Items
Providers	Provide high‐quality patient care using evidence‐based medicine (EBM) to incorporate molecular pathology when indicated. Become aware of their own prejudices and biases and seek out training for how to improve care to all populations. Experts should use their expertise to create official, unambiguous EBM guidelines.
Payers	Stay abreast of current evidence‐based practices and attempt to understand the indications for new tests and procedures. Provide insurance coverage in an ethical and transparent way while doing no harm to patients.
Laboratorians, pathologists, and laboratory directors	Provide high‐quality patient care using evidence‐based medicine. Engage in quality improvement programs in the laboratory to continue to drive down costs. Assist in educate patient‐facing colleagues about testing indications and test selection. Improve direct communication through pathology reports and laboratory results that are returned to both the patient and the patient's care team. Understand changing laws and policies and get involved with pathology organizations like CAP to help guide policy and improve patient care. Safeguard genetic information.
Hospital administrators	Solicit information to understand the patient population that their hospital serves. Communicate with patient facing clinicians, as well as pathologists and hospital laboratories. Hire a diverse workforce at all levels. Work with third‐party payers to make pricing transparent. Reduce barriers to health‐care provider CME.
Researchers	Become aware of their own prejudices and biases while also including participant communities in their research design. Follow institutional and federal regulatory standards, especially those pertaining to ethics and integrity. Consider unintentional biases in genome‐sequencing enrollment and work to offer research opportunities to all populations equally. Safeguard genetic information.
Bioethicists	Consider exploring the complex ethical issues involved in pathology and laboratory medicine. Comment on issues such as those proposed in this article in order to start a conversation about just and equitable molecular pathology care.
Policymakers	Engage with experts in the health‐care industry to better understand how policy proposals will impact patients. Broaden access to health care and improve its affordability. Ensure that lobbyists from commercial health‐care companies and private insurance do not have undue influence on policy proposals.

Patient‐facing providers need to practice according to evidence‐based medicine, incorporating molecular pathology when indicated, looking for ways to contain costs, and staying up to date through continuing education. Providers also need to become aware of their own prejudices and biases and seek out training to improve care to all populations. Clinical trials need to be offered to all patients at the same rates, regardless of the patients’ ethnicity, socioeconomic status, and the like. Finally, expert health care providers should lend their expertise to create clear guidelines and recommendations following EBM, so that there is more clarity on what testing is necessary.

Payers need to update policies to cover current evidence‐based practices, including NCCN guidelines, provide more financial transparency to their clients, better assess medical necessity before denying claims, and remove preauthorization policies that harm patients.

Laboratories are responsible for providing high‐quality patient care using evidence‐based medicine while also looking for ways to drive down costs. Pathologists have a duty to educate patient‐facing colleagues about new testing and testing technology with a cost‐conscious approach. Because clinical laboratories are a historically underappreciated component of patient care, hospital laboratories should follow current laws and policies but also advocate policy changes that would improve patient care. Finally, molecular genetic laboratories should safeguard genetic information obtained during the course of clinical care.

Hospital administrations should hire a diverse workforce at all levels and in every area of their hospital. Hospitals should work to understand the patient populations they serve as well as the needs of different groups within the hospital, including laboratories. Hospital leadership also must work with payers to provide greater price transparency to patients and to their own departments. Hospitals should consider investing in new medical technology needed for patient care, even when third‐party payers are lagging. Finally, hospitals should provide incentives, adequate time, and financial support for providers’ continuing education, including bias training.

Researchers should be aware of their own prejudices and biases and should include participant communities in designing research. Researchers must ensure that all populations are enrolled in genetic studies at similar rates wherever possible so as to eliminate as much database bias as possible. Last, researchers must safeguard genetic information and ensure that genetic information is used only according to the participants’ consent.

Bioethicists would be well served to try to understand the fields of pathology and laboratory medicine in order to start a conversation about just and equitable molecular pathology care.

Finally, policymakers must understand the issues and the unintended consequences of new legislation by consulting experts in health care, as well as those in laboratory medicine. Policymakers must show leadership, passing legislation to broaden access and affordability to health care. They must act in the interest of their constituents and ensure that lobbying by commercial health care companies and the insurance industry does not have undue influence on policy proposals. The third‐party payer system, as illustrated here, is broken; what kinds of repair are required (or if it can be repaired) will be a policy decision, not a medical one.

With all stakeholders working together, molecular pathology can be distributed in an equitable and just manner in the United States, but this change will not take place overnight. While work is done to bring molecular pathology to a broader patient population, it is important to remember that the standard of care is a moving target. Molecular pathology will continue to play a role in the increasing move toward precision in medicine, and, we hope, with the changes we have suggested, will contribute a small part to health care equity.

Funding/Support: This research received no specific grant from any funding agency in the public, commercial, or not‐for‐profit sectors.

Acknowledgments: The authors of this paper would like to thank the Loyola University Chicago Health Science Library for their administrative support during the literature review process. The authors would also like to thank the Milbank Quarterly peer reviewers for their formative feedback.

Conflict of Interest Disclosures: The authors of this article have no conflicts of interest to declare.