Combined population genomic screening for three high-risk conditions in Australia: a modelling study

Abstract

Background: No previous health-economic evaluation has assessed the impact and cost-effectiveness of offering combined adult population genomic screening for mutliple high-risk conditions in a national public healthcare system.

Methods: This modeling study assessed the impact of offering combined genomic screening for hereditary breast and ovarian cancer, Lynch syndrome and familial hypercholesterolaemia to all young adults in Australia, compared with the current practice of clinical criteria-based testing for each condition separately. The intervention of genomic screening, assumed as an up-front single cost in the first annual model cycle, would detect pathogenic variants in seven high-risk genes. The simulated population was 18-40 year-olds (8,324,242 individuals), modelling per-sample test costs ranging AU$100-$1200 (base-case AU$200) from the year 2023 onwards with testing uptake of 50%. Interventions for identified high-risk variant carriers follow current Australian guidelines, modelling imperfect uptake and adherence. Outcome measures were morbidity and mortality due to cancer (breast, ovarian, colorectal and endometrial) and coronary heart disease (CHD) over a lifetime horizon, from healthcare-system and societal perspectives. Outcomes included quality-adjusted life years (QALYs) and incremental cost-effectiveness ratio (ICER), discounted 5% annually (with 3% discounting in scenario analysis).

Findings: Over the population lifetime (to age 80 years), the model estimated that genomic screening per-100,000 individuals would lead to 747 QALYs gained by preventing 63 cancers, 31 CHD cases and 97 deaths. In the total model population, this would translate to 31,094 QALYs gained by preventing 2612 cancers, 542 non-fatal CHD events and 4047 total deaths. At AU$200 per-test, genomic screening would require an investment of AU$832 million for screening of 50% of the population. Our findings suggest that this intervention would be cost-effective from a healthcare-system perspective, yielding an ICER of AU$23,926 (∼£12,050/€14,110/US$15,345) per QALY gained over the status quo. In scenario analysis with 3% discounting, an ICER of AU$4758/QALY was obtained. Sensitivity analysis for the base case indicated that combined genomic screening would be cost-effective under 70% of simulations, cost-saving under 25% and not cost-effective under 5%. Threshold analysis showed that genomic screening would be cost-effective under the AU$50,000/QALY willingness-to-pay threshold at per-test costs up to AU$325 (∼£164/€192/US$208).

Interpretation: Our findings suggest that offering combined genomic screening for high-risk conditions to young adults would be cost-effective in the Australian public healthcare system, at currently realistic testing costs. Other matters, including psychosocial impacts, ethical and societal issues, and implementation challenges, also need consideration.

Funding: Australian Government, Department of Health, Medical Research Future Fund, Genomics Health Futures Mission (APP2009024). National Heart Foundation Future Leader Fellowship (102604).

Keywords: Cost-effectiveness analysis; Genomic testing; Genomics; Health-economic evaluation; Population screening; Prevention; Public health.

Fig. 1

Schematic: Decision tree of population genomic screening versus the status quo. The combined model used decision trees followed by cohort multistate transition models to compare the health and economic outcomes of two different testing strategies for the detection of pathogenic variant (PV) carriers for hereditary breast and ovarian cancer (BRCA1 and BRCA2 genes), Lynch syndrome (MLH1 and MHS2 genes) and familial hypercholesterolemia (LDLR, APOB and PCSK9 genes). Strategy 1 (the comparator) is the status quo in Australia of criteria-based genetic testing for each condition separately. Strategy 2 (the intervention) is combined population genomic screening for the three conditions combined in 18–40-year-olds.

Fig. 2

Markov models. Markov models were created for: a) Female BRCA1 and BRCA2 pathogenic variant carriers, capturing outcomes for breast and ovarian cancer; b) male MLH1 and MHS2 carriers, capturing outcomes for colorectal cancer; c) female MLH1 and MHS2 carriers, capturing outcomes for colorectal cancer and endometrial cancer; and d)LDLR, PCSK9 and APOB carriers, capturing outcomes for coronary heart disease.

Fig. 3

One-way sensitivity analysis. The incremental cost-effectiveness ratio (ICER) for the base-case of the combined model was AU$20,734/quality-adjusted life year (QALY) gained. Results of one-way sensitivity analysis for all input variations led to ICERs under the AU$50,000/QALY willingness-to-pay threshold. These ICERs ranged between AU$15,000–$25,000/QALY gained. Low-limit ICER estimates are shown in blue and upper-limit ICER estimates are shown in orange.

Fig. 4

Probabilistic sensitivity analysis. Results for the base case of the model (AU$200 per sample) are shown for: a) individual condition Markov models separately (hereditary breast and ovarian cancer in red, Lynch syndrome in green, familial hypercholesterolemia in blue), and b) the combined model (all three conditions together in black). Results are presented as incremental costs (x-axis) per carrier versus incremental effect (y-axis) per carrier. The diagonal black line indicates the AU$50,000 per quality-adjusted life year (QALY) wiliness-to-pay threshold or cost-effectiveness plane.