Methods | Evaluating whole genome sequencing for rare diseases in newborn screening: evidence synthesis from a series of systematic reviews

Evaluating whole genome sequencing for rare diseases in newborn screening: evidence synthesis from a series of systematic reviews · Methods

Methods

CC BY 4.0 · National Institute for Health and Care Research · approximately 36 min read

Some text in this chapter has been reproduced from Freeman K, Taylor D, Dinnes J, Clark CCA, Kander I, Scandrett K, et al. Challenges in evaluating whole genome sequencing for newborn screening: series of systematic reviews and roadmap for evidence generation for policy advisers. BMJ Medicine 2025;4:e001726. https://doi.org/10.1136/bmjmed-2025-001726. This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for non-commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc/4.0/. The text below includes minor additions and formatting changes to the original text.

The methods for conducting this review were predetermined and published on PROSPERO (registration number CRD42023475529).

Review of five conditions

The traditional approach to evaluating the evidence for a new screening programme is to conduct a thorough systematic review and cost-effectiveness analysis (CEA) of the test and condition to inform discussions around the extent to which UK NSC screening criteria are met. This traditional review approach was not considered to be feasible for each of the 200 rare diseases included in GEL’s Generation Study.6 Instead, a series of five systematic reviews covering 5 of the 200 rare diseases was undertaken to establish the evidence base per condition and to provide a reference case for comparison with alternative review approaches.

Objectives

Six key objectives aligned with evidencing UK NSC criteria were identified:

to identify the penetrance and expressivity of different gene variants associated with each condition in untreated infants/young people up to 18 years old
to identify the proportion of infants/young people up to 18 years with biochemical or biochemical and clinical features of each condition carrying the genetic variants known for the conditions
to evaluate the diagnostic accuracy (clinical validity) of gene sequencing for each condition
to evaluate the effectiveness of earlier intervention (treatment or surveillance) for each condition, or, if comparative data on early versus late intervention are unavailable, to separately evaluate the effectiveness of treatment in screen detected cases and following clinical presentation
to evaluate the effectiveness of WGS for newborn screening for each condition in terms of disease-related morbidity and mortality
to identify any harms of WGS for newborn screening for each condition and to identify any additional benefits beyond those afforded by earlier intervention.

The relevant UK NSC criteria per question are outlined in Table 1.

TABLE 1 Key questions for the evidence summary and relationship to UK NSC criteria

Key question^a	Related UK NSC criteria^b,16
Question 1: What is the penetrance and expressivity of different gene variants associated with each condition?	1. The condition should be an important health problem as judged by its frequency and/or severity. The epidemiology, incidence, prevalence and natural history of the condition should be understood, including development from latent to declared disease and/or there should be robust evidence about the association between the risk or disease marker and serious or treatable disease
	3. If the carriers of a variant are identified as a result of screening, the natural history of people with this status should be understood, including the psychological implications
Question 2: What is the prevalence of genetic variants in those with biochemical or biochemical and clinical features of each condition?	1. The condition should be an important health problem as judged by its frequency and/or severity. The epidemiology, incidence, prevalence and natural history of the condition should be understood, including development from latent to declared disease and/or there should be robust evidence about the association between the risk or disease marker and serious or treatable disease
Question 3: What is the diagnostic accuracy (clinical validity) of gene sequencing for each condition?	4. There should be a simple, safe, precise and validated screening test
	5. The distribution of test values in the target population should be known and a suitable cut-off level should be defined and agreed
	8. If the test is for a particular variant or set of genetic variants, the method for their selection and the means through which these will be kept under review in the programme should be clearly set out
Question 4: What is the effectiveness of earlier vs. later intervention (treatment or surveillance) for each condition?	9. There should be an effective intervention for patients identified through screening, with evidence that intervention at a pre-symptomatic phase leads to better outcomes for the screened individual when compared with usual care. Evidence relating to wider benefits of screening, e.g. those relating to family members, should be taken into account where available. However, where there is no prospect of benefit for the individual screened, then the screening programme should not be further considered
	10. There should be agreed evidence-based policies covering which individuals should be offered interventions and the appropriate intervention to be offered
Question 5: What is the effectiveness of WGS for newborn screening for each condition?	11. There should be evidence from high-quality RCTs that the screening programme is effective in reducing mortality or morbidity. Where screening is aimed solely at providing information to allow the person being screened to make an ‘informed choice’ (such as Down syndrome or cystic fibrosis carrier screening), there must be evidence from high-quality trials that the test accurately measures risk. The information that is provided about the test and its outcome must be of value and readily understood by the individual being screened
Question 6: What are the harms of WGS for newborn screening for each condition and any additional benefits beyond those from earlier treatment?	13. The benefit gained by individuals from the screening programme should outweigh any harms, e.g. from overdiagnosis, overtreatment, false positives, false reassurance, uncertain findings and complications

Selection of five conditions for review

A range of scenarios that would be likely to impact the UK NSC’s advice on the implementation of WGS for newborn screening were identified. These scenarios were primarily driven by the nature of the intervention(s) that could be introduced on detection of the condition, that is:

widely available and relatively low-cost treatment that carries a low risk of harm (e.g. adverse effects) for the patient and wider family, for example, vitamin therapy
intervention centred around long-term surveillance to allow earlier detection of the clinical manifestation of the condition with the associated anxiety and costs both in terms of resource use and time required to attend appointments, for example, regular outpatient appointments to assess biochemical or developmental changes
ongoing long-term high-cost treatment, potentially with more significant side effects and requiring hospital visits and long-term monitoring, for example, high cost or new drugs with unpleasant or unknown side effects
potentially curative interventions carrying high short-term risks and costs to NHS but long-term lower impact on both NHS, patients and their families, for example, stem cell transplantation
where existing screening and treatment pathways exist such that the impact of WGS would be incremental, for example, hearing screening using Automated Auditory Brainstem Response test and WGS.

Genomics England Limited provided a shortlist of 27 monogenic conditions, which (1) were considered to meet their four principles (GEL score 1, judgement in July 2023)17 and (2) have a relatively high prevalence, within the context of rare disease. Following exclusion of conditions previously reviewed and not recommended for screening by the UK NSC, one condition with an intervention falling under each of the five scenarios above was selected at random. The final list consisted of the following five conditions:

Pyridoxine-dependent epilepsy (PDE): PDE is a neurological condition resulting from an enzyme deficiency, which causes the accumulation of metabolites which inactivate pyridoxine. This pyridoxine depletion causes intractable neonatal seizures that become recurrent and prolonged if left untreated. Treatment with high-dose pyridoxine (vitamin B6 supplementation) reduces the incidence and severity of seizures.
Heritable retinoblastoma (hRB): hRB is cancer of the eye caused by disrupted function of the tumour suppressor protein retinoblastoma (RB). hRB occurs where both copies of the faulty gene RB1 are present on conception (either inherited or occurring sporadically, referred to as ‘germline’ or ‘constitutional’ variants). Early intervention for hRB centres around regular ophthalmologic surveillance from birth to allow earlier identification and treatment.
X-linked hypophosphataemic rickets (XLHR): XLHR is an endocrine condition caused by loss of function in the phosphate-regulating neutral endopeptidase (PHEX) protein, which ultimately leads to hypophosphataemia (low phosphate levels) manifesting as rickets. Calcitriol and oral phosphate can be used as preventive treatments; the only available curative treatment being monoclonal antibody burosumab (currently not licensed for < 1-year-olds).
Familial haemophagocytic lymphohistiocytosis (fHLH): fHLH is an immune deficiency caused by malfunction of the perforin/granzyme cytotoxic pathway, leading to a proliferation of lymphocytes and overactive macrophages, which ultimately lead to infiltration and damage of organs, including the bone marrow, liver, spleen and brain. fHLH is usually activated by infection and presents as an acute illness. Active disease can initially be managed using chemoimmunotherapy with allogenic haematopoietic stem cell transplant (HSCT) providing curative treatment.
Medium-chain acyl-CoA dehydrogenase deficiency (MCADD): MCADD is a metabolic condition, currently included in the UK NBS screening programme. MCADD is caused by inactivity or deficiency of the MCAD protein, which affects proper functioning of the liver and can cause metabolic crises during periods of prolonged fasting or increased energy demands. MCADD is not curable, but it is managed preventively through dietary advice to avoid fasting and strict feeding regimens and through the provision of an emergency regimen (glucose polymer feed) to be used during illnesses.

See Report Supplementary Material 1 for full details of the five conditions.

The evidence was reviewed by the condition with all genetic variants considered rather than the pre-specifying gene variants to be reviewed. This approach is not quite the same as WGS, which aims to detect particular variants and therefore does not screen for conditions.

Reviews were undertaken using a rapid evidence assessment approach, producing an Evidence Summary as described in the UK NSC guidance on the evidence review process.7

Identification and selection of studies

Search strategy

Search strategies were developed for each condition by an Information Specialist. The searches were developed in a test database [MEDLINE (Ovid)] and were informed and refined through a series of scoping searches, checks of a proportion of results from these searches and iterative discussions between the Information Specialist (ND), project lead (KF) and members of the reviewing team (IK, JD and SC). The full process of search development is documented in Appendix 1.

The following databases were searched from inception to November 2023 (see Appendix 2 for exact dates and full search details): MEDLINE (via Ovid), EMBASE (via Ovid), Science Citation Index (SCI) (via Clarivate) and the Cochrane Library (via Wiley). No date, language or study-type filters were applied. Search results were managed using EndNote 20 [Clarivate Analytics (formerly Thomson Reuters), Philadelphia, PA, USA] and systematically deduplicated using the University of Leeds method.18

Inclusion criteria

Study eligibility criteria were defined for each review question, and these are summarised below. Full details are reported in Table 2 to allow the comparison of similarities and differences across review questions.

TABLE 2 Eligibility criteria for each review question

PICO category condition	Eligibility criteria according to RQs
Population	RQs 2, 4		RQs 1, 3, 5, 6
PDE	Clinical: intractable seizures that respond to pyridoxine treatmentBiochemical: multiple biomarkers (mainly, α-AASA, pipecolic acid) in the urine, blood or cerebral spinal fluidGene: ALDH7A1		1. Newborn babies with no symptoms or known FH of the five selected conditions (i.e. newborn screening cohorts)
hRB	Clinical: positive eye examination (indirect ophthalmoscopic examination with scleral indentation) usually following dim or absent red reflex testingBiochemical: N/AGene: RB1		2. Sibling or family studies or other approximations
XLHR	Clinical: bone deformity, dental abscesses, stunted growth and bone and joint painBiochemical: low serum phosphate, high urine phosphateGene: PHEX
fHLH	Clinical: fever, enlarged spleen and liver, lymphadenopathy, an array of neurological symptomsBiochemical: high ferritin, abnormal cell counts, disturbed liver function markersGene: PRF1, UNC13D, STX11, STXBP2 (fHLH)
MCADD	Clinical: hypoglycaemic episodes characterised by seizure and metabolic decompensation (vomiting, coma and death)Biochemical: C8 acylcarnitine levels ≥ 0.5 µmol/l and C8:C10 ratio ≥ 1.0Gene: ACADM
Intervention	RQs 1, 2, 3	RQ 4 intervention	RQ 4 timing of intervention	RQs 5, 6
PDE	Presence of pathogenic variants in the relevant gene(s) was detected by direct sequencing using any technology	Supplements of pyridoxine or vitamin B6; LRT	Early treatment or surveillance of ‘gene-positive’ newborns or an approximation thereof	Screening with WGS
hRB		Regular ophthalmoscopic examination usually accompanied by red reflex testing
XLHR		Supplements of oral phosphate, active vitamin D, monoclonal antibody burosumab
fHLH	n	Chemoimmunotherapy and allogeneic HSCT
MCADD		dietary advice
Comparator	RQ 1, 2, 3	RQ 4 intervention	RQ 4 timing of intervention	RQ 5, 6
All conditions	No comparator	As for ‘Intervention’ above	Late treatment or surveillance of symptomatically detected newborns, children or young people up to the age of 18 or an approximation	no WGS screeningcurrent practice, for example NBS screeningno comparator
Outcomes	RQ 1, 2, 3	RQ 4	RQ 5	RQ 6
All conditions	RQ 1, 2, 3: Presence of the condition (clinical or clinical and biochemically defined) in those with the genetic variant of interest	Disease-specific morbidity and mortality	Health-related health outcomes that could be measured across conditions (e.g. quality of life, time to diagnosis and intervention and mortality)	Harms or other benefits from WGS in newborns

Population

For the evaluation of penetrance and expressivity (question 1), diagnostic accuracy of gene sequencing (question 3) and the effectiveness and harms and benefits of WGS (questions 5 and 6), studies of newborn babies with no symptoms or known family history (FH) of the five selected conditions (i.e. newborn screening cohorts) were eligible for inclusion. Where evidence in newborns was limited, sibling or family studies or other approximations were considered.

For the evaluation of the prevalence of different genetic variants (question 2) and the evaluation of the effectiveness of earlier intervention (question 4), studies conducted in newborns, children and young people up to age 18 years with clinical, biochemical or genetic indicators of the five conditions were eligible. The clinical and biochemical definitions for each condition, where applicable, are provided in Table 2. Intervention studies (question 4) ideally included participants who were asymptomatic for the condition of interest (‘early’ intervention) and those treated following clinical presentation (‘later’ intervention).

Exposure/intervention

For review questions 1–3, the exposure was defined as the presence of pathogenic variants in the relevant gene(s) for each of the five conditions detected by direct sequencing using any technology [WGS, WES, next-generation sequencing (NGS), Sanger sequencing, etc.]. The identification of a genetic variant as pathogenic or ‘likely pathogenic’ is a complex process; however, established classification systems, such as the one developed by the American Board of Medical Genetics,9 are often used. We relied on study authors’ classification of pathogenicity.

For review question 4 (effectiveness of earlier intervention), eligible interventions were defined separately for each condition (Table 2). ‘Early’ intervention was ideally defined as the treatment or surveillance of newborns identified as positive for pathogenic variants of the relevant gene(s) (any test) for each condition or an approximation thereof, for example, treatment or surveillance of second or third siblings of the original proband who carried the same pathogenic variants. ‘Late’ intervention was considered to occur in symptomatically detected (without screening) newborns, children or young people up to the age of 18 years for each of the five conditions or an approximation (anything else available). Where evidence was limited, any author-defined ‘early’ versus ‘late’ intervention was accepted.

For review questions 5 and 6 (effectiveness and harms from WGS), the intervention was screening using WGS.

Comparator

No comparators were considered for review questions 1–3.

For review question 4, the comparators were the same as the interventions per condition, introduced at a later time point in the disease, that is, following symptomatic detection of the conditions (‘late’ as defined by the study authors).

For review questions 5 and 6, the comparator strategy was no screening using WGS, comparison with current practice (which in the UK is NBS screening for nine different rare diseases using methods such as MS/MS) or no comparator.

Outcomes

For review questions 1 (penetrance and expressivity) and 3 (diagnostic accuracy of sequencing), the ‘outcome’ is the target condition and reference standard used to measure the target condition. For each question, the target conditions were clinically, or clinically and biochemically, defined conditions. For review question 2 (proportion of different genetic variants in those with the conditions of interest), the population and target conditions are the same (i.e. those with the conditions of interest). Each combination of biochemical and clinical features should be considered as the ‘reference standard’ for detecting the presence of disease and calculation of diagnostic accuracy (review question 3).

For review question 4 (effectiveness of earlier intervention), eligible outcomes were measures of disease-specific morbidity and mortality.

For review question 5 (effectiveness of WGS), any health-related health outcomes that could be measured across conditions (e.g. quality of life, time to diagnosis and intervention and mortality) were eligible.

For review question 6 (harms or other benefits from WGS in newborns), harms included effects associated with false-positive results, overdiagnosis (including identification of variants of uncertain significance), ethical issues, parental or proband anxiety, referral to surveillance pathway, missing management pathways, data storage and adverse events from treating asymptomatic newborns. Potential benefits included greater certainty (doctors and patients), reduced anxiety, fewer investigations, appropriate surveillance or management plan (therapeutic yield), earlier diagnosis and earlier treatment.

Study designs

All the conditions selected for this review are considered to be rare; therefore, we expected a relative lack of evidence for each of the six review questions. We did, however, rank study designs in the order of priority for each review question. Systematic reviews were eligible for all six review questions.

For review questions 1 and 3, we considered the following types of primary study:

observational studies of newborn screening population without treatment and follow up to disease
observational studies of any screening population with treatment and follow up to disease with matched comparator (or with no comparator)
observational studies of screening of which only gene variant positives are included
case-control studies (review question 3 only)
sibling studies
case series (i.e. more than one case or family).

For review question 2, any observational studies reporting sequencing results for individuals with the conditions of interest were eligible. A minimum of four cases with the condition was required.

For review questions 4–6, randomised controlled trials (RCTs), non-randomised studies, before-and-after studies and other cohort studies were eligible.

Exclusion criteria

Papers that fulfil the following criteria were excluded:

studies of people older than 18 years at diagnosis
studies of non-hereditary forms of the five conditions
qualitative studies
studies that provided insufficient information for the assessment of methodological quality/risk of bias
studies reporting outcomes not relevant to our review questions
studies where > 10% of the sample did not meet our inclusion criteria and are not reported separately
study reports not available in the English language
single case studies (studies of one case or one family; however, we reported the number of case studies per condition)
letters, reviews, editorials, communications, conference abstracts and other grey literature, publications that contained no numerical outcomes data.

Selection of studies

Titles and abstracts of records identified by the searches were screened by one reviewer. A random 20% sample of records were screened independently in duplicate by a second reviewer, and any records with any uncertainty over inclusion (coded ‘Maybe’ by either reviewer) were discussed and a consensus decision was reached. The full publications of all records selected were obtained and assessed by one reviewer, with a random 20% sample assessed independently by a second reviewer. Disagreements were resolved by consensus, or through discussion with a third reviewer.

Data collection and analysis

Data extraction

All data extractions were extracted into a piloted electronic data collection form. Data were extracted by one reviewer and checked by a second reviewer. Any disagreements were resolved by consensus or discussion with a third reviewer.

Assessment of methodological quality

At the outset, we planned to carry out quality appraisal using dedicated tools for each study type, for example, using the Risk of Bias in Systematic Reviews (ROBIS) tool for systematic reviews,19 the Quality in Prognostic Studies tool20 or tools developed by the Joanna Briggs Institute.21
Ultimately, the majority of reports included were of relatively small series of patients with the conditions of interest. The decision was, therefore, taken to use a single tool designed for the appraisal of case series and case reports, tailored to the review question.22 Use of a single tool allowed an overall picture of study quality across study types and conditions.

Quality assessment was conducted by one reviewer with all assessments checked by a second reviewer.

Methods for synthesis

We planned to employ an order of priority approach to synthesis, providing a narrative synthesis of studies that employ the highest priority design available with an accompanying tabulation of key details from studies using lower priority designs (e.g. study design, countries in which the studies took place, sample sizes and key outcomes) (see Protocol for full details, PROSPERO: CRD42023475529).

Due to the paucity of studies using higher priority study designs and the lack of available data for review questions 1 (penetrance and expressivity), 3 (diagnostic accuracy of sequencing), 5 and 6 (effectiveness and harms or other benefits from WGS in newborns), an alternative approach to synthesis was adopted.

Studies providing data for review question 2 (prevalence of different genetic variants in those with the condition of interest) were classified into subgroups according to the definition of disease using clinical and/or biochemical measures (broadest to narrowest), and the largest most representative study from each subgroup was synthesised. For any subgroups where a single study was insufficient to cover the breadth of tests (e.g. numbers of genes included) and population, more than one study was selected for synthesis. Data concerning study design, country in which the study took place, definition of the condition, number of participants, genes tested and genetic tests used, gene frequency and types of variant identified were presented. Available data on expressivity in those with the condition were also presented. A subset of data items was reported for the remaining studies that were considered relevant for the review question 2 (see Appendix 3, Table 16).

Studie providing data for review question 4 (evidence on earlier vs. later treatment) were classified into subgroups according to the definition of ‘early’ versus ‘late’ treatment that was used. Studies using definitions most closely related to the review question were synthesised. For completeness, case reports (typically, family or sibling studies with four or fewer patients) reporting a comparison of early versus late treatment were reported in Appendix 5, Table 31. Data concerning study design, country in which the study took place, definition of the condition, number of participants, definition of early and late treatments, outcome measures, time point of measurement and results were presented. A subset of data items was reported for the remaining studies considered relevant for the review question 4 (see Appendix 3, Table 19).

Exploring Clinical Genome Resource as an evidence source

The existing online research ClinGen was explored as an evidence source for the clinical actionability of rare paediatric genetic diseases (i.e. the risk of variants is known to be high, and intervening will prevent/mitigate disease), and findings were compared to the traditional reviews of the five conditions PDE, hRB, XLHR, fHLH and MCADD. The main aim was to assess the resource as a potential evidence source for the UK NSC, considering that traditional reviews for over 200 conditions are unlikely to be feasible.

Research question (RQ): What is the utility of ClinGen as an evidence source for the clinical actionability for the genes associated with the 5 selected conditions and scaled up to 200 conditions?

Background to Clinical Genome Resource

The or ClinGen is a National Institutes of Health [US]-funded, open access and centralised resource that defines the clinical relevance and actionability of genomic variants. ClinGen’s action group developed practical methods to identify genetic disorders with the greatest clinical utility when detected in previously undiagnosed adults23 and adapted the methods to the paediatric context.24 Working groups use a standardised protocol to produce summary reports and semi-quantitative metric scores for childhood-onset rare genetic disorders. The methods to produce summary reports are adapted from work by Goddard et al.25 to guide decisions about returning incidental findings. The method provides a transparent, systematic, evidence-based process for identifying and quality-rating of evidence. The evidence is reviewed by an expert panel that applies a semi-quantitative metric based on Berg et al.26 to score the overall clinical actionability of gene variants. Each topic is scored independently by multiple members, and the scores are then discussed using consensus for assigning a single actionability score.

The ClinGen focuses on four main curation activities: gene–disease validity (pathogenic variants in the gene clearly cause disease), dosage sensitivity (loss or gain of the gene results in disease), variant pathogenicity (categorisation of variants in the gene into benign, uncertain and pathogenic) and clinical actionability. As of March 2024, the database reported 6357 curated variants across 100 genes (not restricted to paediatric-onset diseases) and included 144 paediatric actionability reports.24 Conditions/genes have to meet a set of minimum requirements to be reviewed. These include:

Guidelines on an intervention relevant to an undiagnosed paediatric population exist (focusing on disease prevention, lowering the clinical burden and improving clinical outcomes and not including ‘personal utility’, reproductive decision-making and ‘ending the diagnostic odyssey’).
At least one variant in the gene should have moderate-to-high penetrance (40% penetrance, relative risk = 2), or no information on penetrance is available.
The health condition is significant.

Actionability reports determine the clinical actionability of secondary findings in paediatric patients undergoing clinically indicated diagnostic testing. Although the paediatric protocol states that elements relevant for population-based screening decisions may be captured, there is a lack of consideration of systems-based practice and availability of population-scale follow-up. Actionability ratings from ClinGen are therefore insufficient for recommending screening in asymptomatic cohorts.24

Paediatric summary reports summarise the evidence on gene–condition pairs under four dimensions: severity of disease, likelihood of disease (similar to penetrance), effectiveness of intervention and nature of intervention. The dimensions are scored from 0 to 3 (3 being best for actionability) based on the evidence, and the evidence is rated for the likelihood and effectiveness dimensions (poor evidence to substantial evidence). The scores are summarised across the four dimensions to provide an overall score for each outcome–intervention pair. The scores and the evidence are taken into consideration for the final assertion on actionability. The evidence review allows for a pragmatic approach allowing non-systematic and expert based references, because it is recognised that this is the most commonly available evidence for rare genetic disorders.

Clinical Genome Resource information for the five conditions reviewed

We searched the ClinGen database on 19 February 2024 for each of the genes included in our review of five conditions and identified reports from the paediatric actionability working group. For each of the five conditions, we extracted the scores for condition–intervention pairs where available, the overall assertion, the evidence provided underlying the scores and the references cited. We tabulated the scores and evidence level comparatively against the evaluation from GEL and against our assessment against the UK NSC criteria. We used the upper quintile of the ClinGen score as the cut-off for actionability as recommended by the Locus-Variant Binning Committee, who developed the transparent semiquantitative metric for evaluating the clinical actionability for pathogenic variants.26

We performed a narrative synthesis of our assessment.

Review of genomic studies of newborn screening cohorts reporting penetrance for pathogenic variants

Knowledge of penetrance is important for the decision about which pathogenic variants of genes associated with childhood diseases should be reported for action following detection by sequencing. This is particularly important in the screening context to maximise the number of babies who are likely to benefit from detection at screening and those who are likely overdiagnosed (detection rate of pathogenic variants is not sufficient). The main aim of this focused review was to identify studies reporting penetrance as an outcome following sequencing in the newborn screening setting for any paediatric condition – to explore the feasibility of identifying highly penetrant pathogenic variants that could be considered for a screening programme (i.e. minimising harm from reporting variants of unknown or uncertain significance).

RQ: What is the penetrance or actionability of pathogenic/likely pathogenic variants of rare genetic child-onset diseases identified in newborn screening populations using WGS?

Identification and selection of studies

Searches were developed iteratively in a single database (MEDLINE via Ovid). The final search combines the concept of newborn screening with either WGS, WES, penetrance, actionability, sequencing or allele frequency.

The following databases were searched from inception to January 2024 (see Appendix 2 for exact dates and full search details): MEDLINE (via Ovid), EMBASE (via Ovid), SCI (via Clarivate) and the Cochrane Library (via Wiley). No date, language or study-type filters were applied. Records were exported to EndNote and systematically de-duplicated using a process based on the University of Leeds method.18

Eligibility criteria

This approach relies on the availability of information on penetrance/expressivity and generalisability of findings to the screening context. We, therefore, only included studies of newborn screening populations that report as a minimum the penetrance (or an approximation) of gene variants linked to rare genetic diseases with childhood onset.

Studies that satisfied the criteria listed in Table 3 were included.

TABLE 3 Study eligibility criteria

Population	Newborn babies without symptoms or known FH of rare genetic diseases
Target condition	Any rare genetic condition with childhood onset of symptoms
Exposure/intervention	Screening using WGS or WESStep down: direct sequencing using panel tests or single gene testing in newborns without symptoms or known FH
Comparator	Order of priority (comparator is only needed for approximation of penetrance using allele frequency):
	1. No comparator necessary
	2. Randomisation WGS vs. standard care
	3. Contemporaneous cohort of matched/random newborns without genetic screening
	4. General population of newborns without genetic screening
	5. General population of healthy adults
Outcomes	For comparator 1:
	a. Follow-up without treatment to clinical presentation
	Step down:
	b. Follow-up to/comparison to biochemical tests (e.g. conventional NBS test), indicating presence of disease processes
	c. Categorisation into levels of penetrance based on existing literature
	d. Categorisation into pathogenic/likely pathogenic based on existing classification systems
	For comparators 2–5:Variant frequency matching the inheritance pattern of the relevant condition in screened and comparator to approximate penetrance
Study designs	Order of priority:
	Observational study of WGS of asymptomatic newborns without reporting to parents (no treatment) and follow-up to clinical features (or reporting results, but no treatment until symptomatic or reporting results and then later stopping treatment to determine whether necessary)
	Observational study of WGS with follow-up testing (e.g. biochemical test)Observational study of WGS follow-up to symptom onset despite treatment Randomised trial of WGS (with results reported and treatment) vs. standard care to determine variant frequency and approximate gene penetranceComparative studies using comparators detailed above to determine variant frequency and approximate gene penetrance
Language	English language

Papers that fulfilled the following criteria were excluded:

Studies of populations other than newborns, studies on populations at risk or with symptoms, studies where WGS is the second-tier test, qualitative studies, studies only reporting variant frequency without an estimation of penetrance, studies only reporting carrier frequency, studies that provide insufficient information for assessment of methodological quality/risk of bias, studies reporting outcomes not listed in our inclusion criteria, studies where > 10% of the sample did not meet our inclusion criteria and were not reported separately, articles not available in the English language, single case studies, letters, reviews, editorials, communications, conference abstracts and other grey literature, publications that contain no numerical outcomes data.

Review strategy

Titles and abstracts of records identified by the searches were screened by one reviewer. A second reviewer independently assessed a random 20% sample of the titles/abstracts plus records labelled as unclear by the first reviewer. Disagreements were resolved by consensus. Full-text articles were independently assessed against the inclusion/exclusion criteria by two reviewers. Disagreements were resolved by consensus. Records rejected at full-text stage are listed in Report Supplementary Material 2 with reasons for exclusion.

Data extraction strategy

Data were extracted into a piloted electronic data collection form by one reviewer and were checked by a second reviewer.

Assessment of methodological quality

Methodological quality of each study was assessed based on study design, as no appropriate tool was identified to assess bias in single-arm cohort studies.

Methods for analysis/synthesis

We tabulated and narratively synthesised information on the proportion of sequencing positive cases, agreement with conventional screening results and the proportion with confirmed disease on clinical follow-up (penetrance) for each condition separately and for all conditions combined per study. We considered positive genetic screening outcomes as unconfirmed if within the studies’ type and length of follow-up, the condition could not be confirmed with confirmatory testing and/or clinical follow-up. We provided our learning from this review for the context of WGS in newborn screening.

Review of cost-effectiveness evaluations of whole genome sequencing and whole exome sequencing

Based on a quick search, we could not find any cost-effectiveness evaluations of WGS or WES specifically for newborn screening. We included studies evaluating the use of WES because there is likely to be an overlap in the methodological challenges faced when considering their cost-effectiveness. To help inform future cost-effectiveness evaluations of using WGS for screening, the objectives of this review were to provide: (1) an overview of the contexts in which the cost-effectiveness of WGS or WES has been evaluated to date, (2) a high-level summary of the methodological approaches adopted in these evaluations, (3) an in-depth appraisal of how health-related quality of life (HRQoL) has been measured and incorporated into cost–utility studies and (4) a summary of the cost components included in micro-costing studies. The motivation for providing an overview of the contexts in which WGS or WES has been evaluated to date is that, if focusing on WGS or WES as a means to detect a disease earlier, the approach adopted to capture the benefits and harms of earlier diagnosis and potentially treatment may be applicable when considering a screening context.

We kept our search strategy and inclusion criteria broad to ensure that any WGS or WES studies conducted in a screening context would be identified.

Objective: To produce a methodological overview of published economic evaluations and costing studies of WGS or WES in any context or population.

Search strategy

Searches for cost or economic evaluations were conducted in the following databases in February 2024:

MEDLINE (Ovid)
EMBASE (Ovid)
CEA registry
Web of Science (WoS)
American Economic Association electronic bibliography (EconLit)

Pre-print sources were not searched.

Our search terms were broadly based on those used in a previously published systematic review of WGS/WES cost-effectiveness studies.27 Searches combined concepts relating to sequencing analyses, including, but not limited to, terms for WGS and WES and (1) costing studies, (2) budget impact studies, (3) economic evaluations and (4) economic models. A previous scoping review, Nurchis et al.28, was identified, which focused specifically on Health Technology Assessments (HTAs) of WGS. Targeted searches for HTAs were not included in the Schwarze et al.27 review, so we added targeted searches of MEDLINE, EMBASE and International HTA from 2022 (i.e. post the searches conducted by Nurchis et al.)28 to ensure identification of any additional HTA reports that our original search may have missed. Searches were restricted to English language and humans. The search development methods can be found in Appendix 1 and the full search details are reported in Appendix 2.

Studies included in the Schwarze et al. systematic review27 and the Nurchis et al. HTA scoping review28 were assessed against our inclusion criteria. References of relevant systematic reviews were checked for any additional primary studies that were not identified by our search.

Inclusion and exclusion criteria

The inclusion and exclusion criteria for the review (Table 4) broadly replicate those used in the Schwarze et al. review, apart from the outcome inclusion criteria.27 This review had a broader scope than our review in that it also included eight studies which only focused on health outcomes (even though the searches targeted economic evaluations). Because our inclusion criteria were restricted to studies that included costs as an outcome, we excluded these eight studies. We also excluded studies that focused on using WGS or WES to diagnose or monitor communicable diseases (e.g. bacterial infections). The reason for this was that the role of sequencing in this context was often very different compared to using it to diagnose a rare disease in an individual. For example, the motivation for sequencing was often to monitor for outbreaks and/or detect treatment resistance – potentially independent of the health of the individual concerned. The timing of the test is crucial in this context, and samples from multiple individuals were often batched together. We also excluded conference abstracts because word count restrictions meant that the studies could not be reported in sufficient methodological detail to facilitate meaningful evidence synthesis.

TABLE 4 Inclusion and exclusion criteria

	Inclusion criteria	Exclusion criteria
Population	Studies in human health care	Studies which focused on populations with suspected communicable diseases
Intervention	WGS and WES (any platform)	Targeted sequencing
Comparator	Any comparator or no comparator
Outcomes	Costs	Included health outcomes only
Study design/ publication type	HTAs, costing studies, budget impact analyses, economic evaluations (partial and full), evidence-based guidelines, clinical trials	Conference abstracts

Data were not extracted from included literature reviews, but the studies included in the reviews were checked against our database search results to ensure that we added any eligible studies that had been missed.

Screening

Initial screening of titles and abstracts, followed by full-text screening was carried out using Rayyan.29 The titles and abstracts of records identified by the searches were independently screened by two reviewers (BS and AO). Where any disagreements occurred, the record was taken through to full-text screening. Full texts were assessed against inclusion/exclusion criteria by one reviewer (BS or AO), while 20% was independently checked by the other reviewer. Disagreements at this stage were resolved through discussion. There were some studies where it was unclear whether the test being evaluated was WGS or WES. These were shared with our genetics advisors for confirmation. The reasons for excluding records at the full-text stage were documented.

Data extraction

An electronic data collection form was developed in Microsoft Excel^® (Microsoft Corporation, Redmond, WA, USA), piloted and refined. The form included the same study characteristics extracted in the Schwarze et al. systematic review (for continuity) but with additional components to capture additional methodological issues relating to costing of WGS/WES pathways and comparator pathways.27 Studies which produced a comparative analysis of costs and health outcomes were categorised as full economic evaluations. This included cost–utility studies, where the outcome is defined as the cost per quality-adjusted life-year (QALY), or cost-effectiveness studies, where the outcome is defined as the cost per change in a particular outcome. Partial economic evaluations, such as cost–consequence analyses, were defined as studies where costs and outcomes are reported, but they are reported in isolation. Costing studies are those which did not evaluate patient outcomes and just reported costs.

Data were extracted by one reviewer (BS or AO), with a random 20% checked by the other reviewer. Disagreements were resolved by consensus or discussion with a third reviewer.

Critical appraisal

Given that the main focus of this cost-effectiveness review was to develop an understanding of the methodology used, rather than the results of the studies identified, we did not appraise the methodological quality of included studies.

Evidence synthesis

For the studies meeting our eligibility criteria, we first provide a tabular/graphical overview of the key study characteristics, including number of publications by year, continent/country, the type of economic analysis conducted, which sequencing approaches were evaluated (WES, WGS or both) and the target population (grouped broadly by age).

To summarise the contexts in which WGS or WES has been evaluated, we provide a narrative overview of the included studies, structured around the PICO:

Population: who were the target population, that is age, disease area.
Intervention: specifics about the timing of the test or aspects relating to how the test was conducted and integrated into the diagnostic pathway.
Comparator(s): what were the comparator(s) in the studies, if any, and what types of data were used to inform the comparator pathways and associated resource use.
Outcomes: what was the main outcome of the study, for example, cost per patient, cost per diagnosis, cost per QALY.

We then provide a summary of some specific aspects of the methodological approach adopted, which are of particular interest when considering the economic evaluation of WGS. These include the costing perspective adopted, the time horizon of the analysis and if and what discounting was applied and whether any sensitivity or scenario analyses were conducted to capture and demonstrate uncertainties in the underlying evidence base. The time horizon adopted is important as WGS often has long-term implications, particularly for conditions with lifelong impacts or where early detection can alter long-term outcomes. Given the likely long-term nature of the impacts of WGS (and screening more broadly), the use and choice of a discount rate is also likely to significantly influence the value of future costs and benefits. The costing perspective is important as there are disparities and debate around who is responsible for the immediate and long-term cost consequences associated with WGS. Using WGS at scale may also lead to societal costs/cost savings that require consideration. There is also likely to be a high level of uncertainty in the evidence base underpinning economic evaluations of WGS, and we wanted to assess whether sensitivity or scenario analyses were routinely conducted to capture these uncertainties. There are other methodological factors important to consider when critiquing an economic evaluation, such as characterising heterogeneity across population subgroups or the analytic approach adopted, but we felt the appropriate approach to these are more context-specific and therefore challenging to critique in a review with such a broad clinical focus.

Since the reference standard outcome for a UK NSC economic evaluation is cost per QALY, we present a more detailed methodological review of the included cost–utility studies. Specifically, we focus on how HRQoL has been measured and incorporated into these evaluations.

Additionally, we provide a narrative summary of the costing studies that utilised a bottom-up approach to estimating the costs of WGS. Examining the components included in these costings and the methodological approaches employed offers valuable insights for future evaluations of newborn screening programs incorporating WGS.

Consideration of the public voice in the evaluation of whole genome sequencing

The broad aims of the patient and patient and public involvement and engagement (PPIE) were to:

build an understanding of the perspectives and experiences of members of the rare disease community around the key challenges and opportunities of WGS in newborn screening
explore views towards, and understandings of, screening programmes and the role of the UK NSC
discuss methodological challenges identified during the review process
note views on limitations in the evidence base
contribute meaningfully to the development of future PPIE in this area.

Recruitment targeted people with differing lived experiences of rare genetic conditions (including parents and adults living with rare conditions), third-sector representatives and prospective parents from the public. Recruitment took place between October and December 2023. Due to the restricted time frame, recruitment was initially through a targeted approach via known contacts at several third-sector organisations and through GEL’s PPIE groups. Through these contacts, we recruited the charity representatives (including an adult living with a genetic condition). Four of the parent representatives were recruited through a post on a social media page for families living with rare genetic conditions, and one approached the research team after reading a press release about the study. The public representative had previously been involved in PPIE work related to screening for a rare genetic condition, so they were able to meaningfully contribute without needing extensive guidance/input on genomics, screening and health care.

The group consisted initially of eight members: five women and three men:

a member of the public, who was also an expectant, and later a new, parent
an adult living with a rare genetic condition, who was also a Diversity, Equity, Inclusion and Belonging consultant and representative for a charity organisation supporting people living with a genetic condition
representative from a charity advocating for and supporting people living with genetic conditions
five parents of children living with a rare genetic condition (ages of their children ranged from 5 years to young adults).

One of the parents withdrew from the group very early in the process due to their caring commitments.

Five (2-hour) virtual meetings took place between 15 January and 21 May 2024 (for discussion topics, see Table 5). Meetings 1 and 4 included the involvement of a member of the review team (KF) to present the review and its progress, with the remaining three meetings being independent of the review process. Meetings were recorded to facilitate note-taking. To maximise inclusion, participants who were unable to join any of the meetings were offered the opportunity to catch up by watching recordings of missed meetings (with the permission of the participants present at the meeting) and sending feedback via e-mail.

TABLE 5 Schedule of PPIE meeting discussion topics

Meeting	Topic
1	Introduction to the group and projectTeam and PPIE group introductions. Presentation on the purpose and process of the review. Answering questions from the PPIE group on this
2	Harms and benefitsPre-reading and presentation on: background/context to newborn screening, concept of harms and benefits, the role of the UK NSC, how WGS differs from current screening and that different stakeholder groups may have different views. Open discussion on ‘What do you think are the key harms and benefits of screening that policy-makers should be taking into account’
3	Genetic uncertaintiesPre-reading and presentation: current understanding of genetic uncertainties, including defining terms such as penetrance, expressivity, variants of uncertain significance. Open discussion on the impact of this information on balancing harms and benefits and how to approach these complexities when explaining genetic screening to parents
4	Systematic review findingsPre-reading: executive summary of draft review and open discussion on the methodological challenges and evidence gaps identified during the review process and the conclusions reached
5	Conclusions and role of PPIE in future reviewsOpen discussion on topics, including: the most important harms and benefits; how these can be captured and traded off; how these harms and benefits can be transformed into evidence to support the decision-making of the UK NSC; how can PPIE be best incorporated into future reviews and who needs to be involved in these conversations

Meetings were deliberative, to create the knowledge space required for discussion, drawing on relevant evidence and the experience of the participants. Ahead of each meeting, participants were informed of the topic for discussion and were given up to 1 hour of preparatory work in the form of reading from a variety of sources and/or formulating ideas/questions.

Following each meeting, participants were asked to complete a short evaluation questionnaire. This had two main purposes: (1) to evaluate the format and content of the meeting and (2) to allow participants time to reflect on the content of the meeting and provide any short summary points on this. In the final meeting, which brought together the group’s thinking on all of the topics covered, participants were asked to contribute to an online whiteboard (before, during and after the meeting) with their responses to prompts: What do you see as the main or most important? (1) Benefits and (2) harms to WGS for screening in newborns, (3) how should potential harms be prevented/dealt with, (4) how should benefits and harms be balanced against each other, (6) what kinds of evidence should the UK NSC prioritise, (7) what should happen where the evidence is not available, and finally, (8) how should PPIE contribute to this process (including when and who)?

Attribution. Karoline Freeman; Jacqueline Dinnes; Bethany Shinkins; Corinna Clark; Inès Kander; Katie Scandrett et al. (2025). Evaluating whole genome sequencing for rare diseases in newborn screening: evidence synthesis from a series of systematic reviews. National Institute for Health and Care Research. Copyright © 2025 Freeman et al. Licensed under CC BY 4.0. NCBI Bookshelf NBK619683. CC BY 4.0. Original source ↗. Text reproduced from the Bookshelf XML without changing wording; figures, images and supplements are not reproduced.