“Population genomic screening, which was envisioned almost 20 years ago with the publication of the draft human genome sequence, is recognized by the Centers for Disease Control and Prevention (CDC) as having notable potential to improve public health. In 2014, the CDC’s Office of Public Health Genomics designated Tier 1 genomic applications as those with evidence-based guidelines and recommendations to prevent morbidity and mortality associated with genetic risk. The initial 3 autosomal dominant conditions assigned to Tier 1—Lynch syndrome (LS) (high risk for colorectal cancer [CRC]), hereditary breast and ovarian cancer syndrome (HBOC), and familial hypercholesterolemia (FH)—cause a high lifetime incidence of cancer or cardiovascular disease and have proven risk-reducing interventions. To date, the clinical and economic value of screening for Tier 1 conditions in an unselected U.S. population has been uncertain. [..]
Methods
[..] This decision analytic modeling analysis used computer simulations of the clinical history of hypothetical cohorts of 100 000 persons that were racially and ethnically representative of the U.S. population, in which approximately 1.5% have LS, HBOC, or FH variants according to studies of biobank participants and a recent update of the Healthy Nevada Project screening study cohort. We used annual model cycles, a lifetime time horizon, and a U.S. payer perspective (focused on direct medical care costs [in 2020 U.S. dollars] only) and discounted all future costs and health outcomes by 3% per year to reflect their present value. [..]
Identified probands could undertake evidence-based guideline-recommended interventions to prevent or mitigate disease through prophylactic surgical intervention, including risk-reducing mastectomy (RRM) and salpingo-oophorectomy (RRSO) (for HBOC probands); intensified screening and surveillance leading to earlier-stage cancer detection, including magnetic resonance imaging (for HBOC probands) and increased colonoscopy frequency (for LS probands); or preventive treatment, including lipid-lowering therapy (for FH probands). [..]
We assumed that the cost of a hypothetical population screening test was $250, based in part on similar currently available genomic tests. For identified probands, we assumed an additional cost of $250 for targeted confirmation testing and genetic counseling. The costs for mammography, magnetic resonance imaging, and prophylactic surgeries were derived from published sources. Costs for HBOC-associated cancer were assessed for the first year of treatment, for continuation of treatment in subsequent years, and as palliative therapy for the last year of life. We estimated the cost of colonoscopy from an economic analysis of 20 primary LS screening strategies. Stage-specific CRC costs were applied for the first year of treatment, for continued treatment in subsequent years, and as palliative therapy for the last year of life from a cost-effectiveness analysis of DNA stool testing to screen for CRC. Costs associated with MI, stroke, the post-MI health state, the poststroke health state, and statin therapy were derived from published sources. Post-MI and poststroke costs continued if patients survived in either health state. [..]
Results
Screening 100 000 unselected 30-year-olds resulted in 101 (95% UI, 77 to 127) fewer overall cancer cases and 15 (95% UI, 4 to 28) fewer cardiovascular events. Specifically, screening prevented 45 (95% UI, 31 to 61) incident breast cancer cases, 8 (95% UI, 3 to 12) ovarian cancer cases, 48 (95% UI, 32 to 68) CRC cases, 9 (95% UI, 3 to 18) MIs, and 6 (95% UI, 2 to 10) strokes. These reductions in disease incidence (as well as severity) led to a cumulative increase of 495 QALYs [quality-adjusted life years] (95% UI, 401 to 757) at an incremental cost of $33.9 million (95% UI, $27.0 million to $41.1 million). The ICER was $68 600 per QALY gained (95% UI, $41 800 to $88 900).
The results were most sensitive to uncertainty in the screening assay cost, the relative risk for MI for FH probands, prior MI in FH probands, 5-year survival among women with breast cancer, and CRC stage at diagnosis. Results of the probabilistic sensitivity analysis showed that population screening in 30-year-olds had a 16%, 99.4%, and 100% probability of being cost-effective versus no genomic screening at thresholds of $50 000, $100 000, and $150 000 per QALY, respectively. [..]
Discussion
We found that population screening of young adults for 3 monogenic conditions was likely to be cost-effective compared with the usual practice of testing only those individuals who have a high-risk family history. Testing between the ages of 20 and 40 years was under the commonly used U.S. willingness-to-pay threshold of $100 000 per QALY in the majority of probabilistic model simulations. Screening was cost-effective across various scenarios and sensitivity analyses. The results also support pairing cascade testing for first-degree family members with population screening to efficiently identify additional persons who could benefit from intensified screening or surveillance.
The modeled genomic conditions (LS, HBOC, and FH) have features that are essential to establishing cost-effectiveness. All are driven by genetic risks with higher prevalence across ancestries compared with other genetic disorders. The conditions also induce early-onset disease that substantially shortens life expectancy and decreases quality of life. In addition, the risk for cancer and cardiovascular disease associated with the conditions can be reduced with evidence-based, guideline-recommended care. We also showed that testing for multiple genetic risks simultaneously is critical for the cost-effectiveness of Tier 1 screening, as our previous work demonstrated that individual screening for FH or LS was not cost-effective in U.S. populations and HBOC screening was borderline cost-effective.
[..] several criteria must be met to achieve cost-effective implementation. First, the screening panel should include only genes and variants with strong evidence of pathogenicity. Adding conditions to a screening panel could improve the value of genomic screening, but the cost, potential harms, and efficacy of recommended preventive care and clinical outcomes would need independent evaluation. Second, the testing should be inexpensive; we found that the cost threshold to screen persons who were aged 30 years was $413, but this decreased to $290 for 40-year-olds and $166 for 50-year-olds. Third, high-risk persons should be appropriately counseled about their results, should receive evidence-based recommendations, and should have access to longitudinal care management of their genomic risks. To avoid the risks of false reassurance, individuals with a negative screening result (>98% of all who are tested) should receive effective communication that standard cancer and cardiovascular screening tests are still recommended. Evidence-based approaches to implementation of genomic screening and value-based reimbursement policies will enhance the likelihood that the genomic screening will retain cost-effectiveness in real-world implementations.
Previous cost-effectiveness studies of LS, HBOC, and FH testing have considered these conditions independently, most often in affected persons or their families, and used testing costs up to an order of magnitude higher, reflecting testing costs at the time. [..] Two recent U.S.-based cohort studies evaluated the health services effect of identifying genetic risks that would not otherwise be identified by family history screening. Among 64 392 persons screened for the Tier 1 conditions as part of the MyCode program, 87% of the 351 identified probands did not have a prior genetic diagnosis of their Tier 1 condition. In the Healthy Nevada Project, 90% of participants screened for Tier 1 conditions and identified as carrying a pathogenic variant were not previously identified, 21.9% already had developed an associated disease, and only a quarter reported a relevant family history. These findings indicate that population genomic screening can add clinical value even if the utilization of family history is substantially improved above the 9% pooled rate modeled for this study.
Our study has several limitations. The prevalence and penetrance of monogenic risks are primarily informed by data from European ancestry cohorts [..]. We modeled an average rate of uptake for risk-reducing interventions that likely varies across different real-world populations and health care environments. Although sensitivity and scenario analyses account for some of the variation, our analysis does not yet explicitly acknowledge the disparities in insurance coverage and economic barriers to receiving preventive care services in the United States. Finally, the screening program modeled in this analysis does not return variants of uncertain significance, which allows a lower cost for return of results and reduces complexity and screened persons’ uncertainty but would not identify those at risk due to reclassified variants. Some population screening programs incorporate periodic reanalysis of data, which would reduce this risk.”
Full article, GF Guzauskas, S Garbett, Z Zhou et al., Annals of Internal Medicine, 2023.5.9