Adopting Genomic Technologies in Health Care: A Cost-Effectiveness Analysis of Microarrays in Learning Disability

Abstract:     
Rationale: Completing the draft sequence of the human genome enabled the development of many novel genomic technologies, such as microarray expression profiling. These technologies yield genetic information that may help with disease diagnosis, prognosis and treatment response. Many health care systems face pressure to adopt genomic technologies, some of which are very expensive. Microarray based comparative genomic hybridisation (aCGH) explores genomic gains or losses in chromosomes. Despite its proven diagnostic potential within research, most healthcare systems are unsure whether to place the technology into routine clinical practice, due to the actual array (glass slide) prices and unknown clinical utility.

Our study aimed to assess the potential economic impact of aCGH technology in routine clinical practice, applied to idiopathic (unknown cause) learning disability (ILD). Most individuals with moderate to severe LD need life long support and half of those with mild LD are significantly impaired throughout life. Despite associated clinical, social and psychological burdens, 60~80% of cases remain undiagnosed, arguably because of diagnostic limitations in the current testing method, karyotyping. Research evidence suggests that aCGH can detect 20-25% more diagnoses than karyotyping.

Objective: To perform a cost-effectiveness analysis of aCGH compared with karyotyping, using ILD as an applied example.

Methodology: Using a micro-costing approach and a National Health Service perspective, data on resource use and costs for the two diagnostic testing methods were collected from four laboratories in the United Kingdom. A range of laboratories were chosen to reflect differing configurations of resource use across laboratories and to increase study generalisability. Expert opinion and laboratory records established standard testing pathways and determined average sample throughput. Effectiveness was examined in terms of the number of cases likely to be diagnosed using the different approaches, by exploring a cohort of 100 ILD referrals to either aCGH or karyotyping, creating a cost per diagnosis. Extensive sensitivity analyses explored analysis uncertainty.

Results: The cost of a single aCGH and karyotyping test were 652 and 172 respectively, with the cost of the array slides themselves creating the greatest difference between the testing methods. aCGH detected 10% more diagnoses than karyotyping. The overall cost per diagnosis was 4,605 for aCGH and 7,322 for karyotyping.

Conclusions: This study has shown that the routine use of aCGH could be a cost-effective option in the diagnosis of ILD. Whilst karyotyping is the least expensive test on a stand-alone basis, fewer cases would be missed compared to using aCGH. These study results raise questions concerning whether increased diagnostic capability is worth the additional cost of this and other genomic technologies. We hope that this study provides useful information for health care funders faced with the decision about adopting microarray technology in their own settings. Our ongoing research is exploring the cost-effectiveness of further uses of aCGH, particularly in colorectal cancer. Finally, methodolological issues surrounding the difficulty of using QALYs in diagnostic testing are being considered.

Keywords: Genomics; Cost-effectiveness; Learning Disability

Working Paper Series