Challenges in conducting a systematic review of diagnostic test accuracy of genetic test: an example of the genetic diagnosis of familial hypercholesterolaemia

Article type
Sharma P1, Mowatt G1, Boachie C1, Stewart F1, Miedzybrodzka Z1, Simpson W1, Boyers D1, Kilonzo M1, McNamee P1
1University of Aberdeen, UK
Background: The availability and uptake of genetic tests is increasing. The UK National Institute for Health and Care Excellence commissioned a review on rapid genetic tests for the diagnosis of familial hypercholesterolaemia (FH).

Objectives: To explore methodological challenges of undertaking a review of diagnostic accuracy of genetic tests and to give recommendations for future reviews.

Methods: A systematic review of diagnostic accuracy of Elucigene FH20 and LIPOchip (rapid genetic tests) for the diagnosis of FH was performed. The problems encountered, decisions made and the recommendations suggested were tabulated and summarised descriptively. A review of systematic reviews on the diagnostic accuracy of genetic disorders was conducted to corroborate our issues. A comparative analysis of the methodological aspects, including reporting of sensitivity and specificity, a meta-analysis, susceptibility to bias and quality of reporting, was performed.

Results: Lack of valid data on diagnostic indices, heterogeneity across studies, the use of incomplete reference standards and poor quality of reporting imposed problems during the evaluation process. For example, due to the nature of genetic tests given in practice, no studies reported false positive data. An assumption of no false-positives, and therefore 100% specificity was made. Studies were also susceptible to differential and partial verification bias. Decisions to deal with these problems were made with the help of clinicians and the statistician involved in the process. Two systematic reviews were included from the review of systematic reviews. Data comparing methodological aspects of these reviews with our review will be reported. Both reviews stated the lack of carefully designed studies to make valid evaluation similar to our findings.

Conclusions: Reviewers should implement strategies (intensive scoping, clinician advice etc) early in the process to deal with the complexity of genetic tests. There is a need of specific guidelines for the reporting of diagnostic accuracy studies of genetic tests.