Broad Population Genetic Screening Still Faces Implementation Challenges

November 1, 2019

Broad population-based genomic screening has the potential to improve patient care by detecting genetic causes of disease before they occur; however, the economics behind this approach have not fully been validated, according to a session on the clinical and economic utility of whole-genome sequencing at the AMCP Nexus 2019 meeting.

Broad population-based genomic screening has the potential to improve patient care by detecting genetic causes of disease before they occur; however, the economics behind this approach have not fully been validated, according to a session on the clinical and economic utility of whole-genome sequencing (WGS) at the AMCP Nexus 2019 meeting.

"We don't really have very many good models out there now that are looking at the cost effectiveness. We do need to quantify the costs but doing so is complicated," said Laney Jones, PharmD, MPH, assistant professor, Geisinger, during the presentation. "The personal utility of genomic knowledge is often missing from these current models."

The current methods for genetic testing include chromosomal microarray (CMA), which is designed to detect copy number variations and gene duplications and deletions. Broader testing techniques include whole-exome sequencing (WES) and WGS. In addition to deletions and insertions, WES, which focuses on 1% to 2% of the entire genome that contain 85% of the genes related to hereditary diseases, can also identify missense, nonsense, and splice site mutations. The broadest approach, WGS, covers DNA variations that can affect gene activity.

"There are 5200 genetic disorders for which the molecular basis is known," said Lon Castle, MD, chief of Lab and Specialty Drug Services, eviCore healthcare. "With WGS, you get the full genome. Just imagine the amount of genetic information you're getting there and how long it takes to sift through it and how long it takes to get it right. There's a lot of curation that has to go into those results. There's a lot of bioinformatics that goes into WGS and WES."

Geisinger Experience With Population Screening

The Geisinger hospital network launched a broad population level genetic screening program labeled MyCode in 2007. The hospital system, which is located in central Pennsylvania, has 2800 providers and 1.5 million active patients across 13 campuses. Tied to this, from a payment perspective, the Geisinger health plan has over 580,000 members and covered approximately 40% of those in the MyCode project.

Overall, 253,108 patients participated in the MyCode project, with samples provided for 172,819. Of these, 64,309 were analyzed with clinical results reported for 1068. The majority of these patients were of European ancestry (97%) with a median age of 59.21 years. The project looked specifically at the diseases indicated as tier 1 for genomic application by the CDC: hereditary breast and ovarian cancer (HBOC) syndrome, Lynch syndrome, and familial hypercholesterolemia. These conditions are expected in 1 in 400, 1 in 440, and 1 in 250 individuals, respectively.

Overall, BRCA1/2 mutations were found in 290 of these participants (27%), indicating HBOC. One hundred twenty-five (12%) had APOB or LDLR gene alterations, indicating familial hypercholesterolemia, leading to early onset coronary artery disease and stroke. There were an additional 101 patients (9%) who tested positive for Lynch syndrome, indicating the potential for early colon, uterine, and other cancers.

Of those detected, 42% did not have any evidence of a personal history of cancer and 50% did not have any evidence of a family history of cancer, suggesting that current screening methods would have missed these patients. Two-thirds of the detected individuals (68%) were eligible for risk management.

Not only did the detection of hereditary risk impact the individual who was tested but it also had implications for the rest of the family members who did not previously have a known family history of cancer. Findings from the Geisinger project aligned closely with previously reported results from JAMA, which showed that 55% of those detected by population screening did not meet the criteria for indication-based testing, according to National Comprehensive Cancer Network (NCCN) guidelines.1

"From this initial report and through our primary care docs and other providers in our system learning more about the MyCode project and its implications to clinical care, they've really pushed for the translation of this research project to be transitioned to clinical care," said Jones.

Following this push, population-based genomic screening was incorporated into 2 Geisinger clinics. There have been 679 patients as of August 2019 that had consented to the testing, with results received for 527. Of those tested, 527 were negative and 14 tested positive (2.6% yield).

"We're looking to anticipate care for disease prevention instead of reacting once someone has a disease," said Jones. "This strategy allows us to do earlier detection of disease and enables better management and improved outcomes. This provides more reliable identification of the risk of disease for patients and their facilities."

Drawbacks and Costs

Some of the drawbacks to population screening include scalability and the costs associated with testing, Jones noted. Reimbursement from insurance companies is still very guideline specific and genomic testing is not yet fully integrated into electronic health records. Chief among these challenges were costs, as most economic models are focused on drugs and not diagnostics.

The direct costs associated with population screening are for the assays, analysis, and return of the results. Potential downstream costs exist for provider and patient education, confirmation testing, and interventions and surveillance. Other ancillary costs might also be accrued by population screening, including for false reassurances, interventions or surveillance in response to false positives, and the effects on insurance or employment.

Findings from a randomized pilot study looking at short-term costs between family history alone or WGS were published in 2018.2 In a cardiology setting, WGS cost $5098 more than family history alone. In the primary care setting, there was a $5073 difference. The downstream costs did not differ significantly between groups.

"We don't always know the prevalence of genomic conditions in unselected populations, especially in fields like familial hypercholesterolemia. These markers are always changing and how we define these diseases is changing," said Jones. "We also don't always know the penetrance of the genomic condition or the effectiveness in presymptomatic interventions."

Adding to the complexity of reimbursement is the lack of a single governing entity that is responsible for payment decisions in the genetic testing space. For medications, FDA approval is typically sufficient for payers to make a coverage decision.

"In the payer world, CMS decides if they'll pay or not pay for a genetic test, but that's only for the Medicare population and that has no bearing on what commercial entities do," said Castle. "CMS is looking at a biased population, as individuals over 65, some tests work well in that population and some do not. That is why payers do not blindly follow CMS like some follow the FDA. It's not as clear of a picture."

Outside of one overarching group to weigh-in on payment decisions, the ACCE model has been developed by the CDC to assess new molecular diagnostic, genomic tests, and technologies in an evidence-based fashion, according to Castle. For the model, the framework assesses analytic validity, clinical validity, clinical utility, and ethical, legal and social implications. Coverage decisions are largely based on the first 2 aspects, he noted.

Other sources of potential payment decisions for genetic testing comes from guidelines or pathway organizations, technology assessment committees, regulatory and other governmental agencies, professional societies, and individually contracted providers and industry experts. The key for each of these, however, is to understand how their recommendations are reached. For instance, Castle noted that NCCN is not evidence-based but is consensus-based, which is an important distinction.

Improving WGS Diagnostic Yield

At this time, evidence on the effectiveness of WGS has not yet been strong enough to support broad payer acceptance. Moreover, challenges still remain for interpretation of the data. The NTRK gene is an example of the complexity of interpretation, Castle noted. The treatments larotrectinib (Vitrakvi) and entrectinib (Rozlytrek) are approved for patients with an NTRK gene fusion, which is significantly different than an NTRK gene mutation.

At this time, WES remains significantly cheaper and faster than WGS, further complicating the payment decision. Additionally, in a large meta-analysis the diagnostic yield was not significantly different between WGS and WES.3 In the analysis, which looked at 20,068 children across 37 studies, WGS had a diagnostic yield of 41% whereas WES had a yield of 36%. Both provided a significantly higher yield than CMA (10% yield).

"There's a lot of this we still need to figure out how to handle as a healthcare community. There are all sorts of informed consent issues that we need to grapple with," said Castle. "There are a lot of things to think about with WGS, other than just the technology and validity of the tests."

As the price of WGS continues to decline, work is being done to refine the diagnostic yield. Narrowing the focus of studies to more clearly defined genetic causes could help elevate the yield, Castle said. In an example of this, he cited findings from a study looking at 103 children with phenotypes suggesting a genetic cause for their disease.4 In this analysis, the diagnostic yield with WGS was 41%. The authors of the paper noted that 18 new diagnoses were made in the study based on variants detect by WGS that were not detectable by WES.

"This is how we can get this testing done, if you limit it to where it is going to be the most helpful for the largest segment of the population. That's who you want to test. Other people perhaps based on appeals but that's not where we should start out of the gate, we aren't going to start with everybody," said Castle.


  1. Mandelker D, Zhang L, Kemel Y, et al. Mutation Detection in Patients With Advanced Cancer by Universal Sequencing of Cancer-Related Genes in Tumor and Normal DNA vs Guideline-Based Germline Testing. JAMA. 2017;318(9):825-835.
  2. Christensen KD, Vassy JL, Phillips KA, et al. Short-term costs of integrating whole-genome sequencing into primary care and cardiology settings: a pilot randomized trial. Genet Med. 2018;20(12):1544-1553.
  3. Clark MM, Stark Z, Farnaes L, et al. Meta-analysis of the diagnostic and clinical utility of genome and exome sequencing and chromosomal microarray in children with suspected genetic diseases. NPJ Genom Med. 2018;3:16.
  4. Lionel AC, Costain G, Monfared N, et al. Improved diagnostic yield compared with targeted gene sequencing panels suggests a role for whole-genome sequencing as a first-tier genetic test. Genet Med. 2018;20(4):435-443.