Currently, there is no evidence supporting the large-scale implementation of WGS of newborns with concomitant simultaneous detection of many conditions. The cost and the balance of benefit and harm are unknown, and implementation would prevent the research required to measure the benefits and harms (see Research recommendations section). Our review of genomic studies of newborn screening cohorts demonstrates unequivocally that introduction of WGS without substantial further research would cause harm and uncertainty.
We found that the traditional approach to evidence synthesis can currently not be applied to WGS of newborns, the quantity of work is very high and the data provided are of insufficient quality and therefore insufficient value. Evidence from clinically detected cases is often not generalisable to screening, where the spectrum of disease differs, and less clinically significant disease types are more common. While there are many unknowns about the benefits and harms of screening, we advocate addressing the accurate measurement of penetrance first and filtering review efforts concerning other aspects, such as benefit of earlier treatment, only in those conditions with variants with sufficient penetrance. This proposal is a pragmatic stepwise approach based on the paucity of data and the need to engage with the broader scientific community to deliver the types of studies which will deliver the required penetrance data (see Research recommendations section).
Another possible approach to evidence synthesis is to use existing gene and variant curation by genomics projects. Decisions on which conditions, genes and variants to report are generally based on the severity of disease, disease onset, penetrance and expressivity of pathogenic variants, treatability and access to treatment. Gene curation processes vary greatly but are all based on a condition-by-condition approach. Our assessment of the GEL approach and ClinGen approach showed that they do not meet the requirements of the UK NSC and are not suitable for adoption. An assessment of approaches more widely may be needed.
Therefore, what is needed is commissioning of carefully designed research to generate new evidence with a focus on penetrance data in a screening setting for genes previously selected by gene curation processes. Future review efforts can focus on the narrower questions of penetrance and expressivity of pathogenic variants in large screening cohorts first, with additional questions about earlier treatment benefits only in those with sufficient penetrance and expressivity. Future large-scale WGS testing and screening programmes should be established in ways that support robust evaluation, research and assessment of clinical effectiveness to determine immediate and longer-term costs, benefits and harms.
To date, economic evaluations of WGS have focused on symptomatic populations and have focused predominantly on the costs and outcomes associated with the diagnostic process itself. Studies that have attempted to estimate cost per QALYs associated with WGS/WES testing, of which there are few, were either focused on very specific heath conditions where it is easier to define distinct health states, or have applied very crude quality of life and survival estimates to extremely broad health states. Creating a model which accounts for all the conditions is highly likely to be unfeasible. It may be possible to group some of the conditions by their biological pathways or clinical manifestations, providing a mechanism for identifying more homogenous health states to which you could apply more precise HRQoL, survival and cost estimates. There will inevitably be a high degree of uncertainty associated with any model developed for this evaluation question, driven by an inability to capture the whole decision problem rather than parameter uncertainty or model structure limitations. Methods to interrogate this type of uncertainty and understand the risk of making the wrong decision (i.e. recommending/not recommending WGS for newborn screening) in this context would be valuable to support policy decision-making.