In this WP, we will perform the clinical utility study on implementing WGS as first tier diagnostic test. A prospective parallel design will allow to compare the standard diagnostic pathway and the WGS first pathway at various outcome measures, including the diagnostic yield, the impact on medical decision making, cost-to-diagnosis. These outcomes will be subsequently used to perform a budget impact analysis and to model various scenarios to establish the position of WGS as first tier test.


  • 1.1: Determine the content and format of the laboratory and clinical data to be collected
  • 1.2: Optimize the WGS procedure
  • 1.3: Retrospective and prospective data collection of standard care and WGS first
  • 1.4: Determining of diagnostic yields and cost-to-diagnosis
  • 1.5: Comparison of cost-effectiveness of the WGS-first care and traditional diagnostic pathways in the parallel design (prospective)
  • 1.6: Assessment of the impact of WGS-bases diagnostics on the medical decision making process (for individual patients)
  • 1.7: Perform budget impact analysis (BIA) and decision analytic modeling for introducing WGS into the Dutch healthcare system

Description of Work

  • 1.1 Patient inclusion

    We will include 80 NDD and 65 NICU trios. Criteria for patient inclusion are that i) the phenotype of the patient is likely to have a genetic etiology (expert opinion of a clinical geneticist/ neonatologist) and ii) the patient is naïve for genetic testing (to allow for a prospective analysis). Specific exclusion criteria are i) the patient has had genetic testing before, and/or the phenotype is (likely) caused by external factors such as (birth) trauma and/or viral infections (Figure 2 and Figure 3). There will be no selection on gender, race or ethnicity.
  • 1.2 Whole genome sequencing

    Patients will undergo both the standard care pathway as well as the WGS-first care pathway. Together with WP2, standard formats for the WGS sequencing procedure will be generated including sequencing quality parameters, analysis pipeline, filtering strategies, etc. Also, interpretation and reporting guidelines as well as data sharing will be harmonized in collaboration with WP2 and WP3. To avoid unnecessary double testing costs, other genetic tests will be inferred from the WGS data wherever possible as it is expected that these data can be inferred from WGS data without loss of quality. Wherever WGS reveals the causative pathogenic mutations prior to standard testing, these results will be used in the care pathway of the patient.
    Cost, TAT, and effects on medical decision making for standard testing will be calculated/inferred as if the test was performed. Data on diagnostic outcomes will not be shared between the two diagnostic pathways prior to reaching a conclusion in each of them separately. Exceptions to this rule will be i) when WGS-first reaches a conclusive diagnosis in NICU patients (prior to the standard care being complete), as these may have impact on clinical decision making, ii) when WGS-first reaches a conclusive diagnosis in NDD patients (prior to completion of the standard care pathway or maximum follow-up time) with clinical implications.
  • 1.3 Health care resource use data collection

    1.3.1 Prospective data collection
    Economic analysis of the WGS-first care demands both prospective and retrospective data collection to gain full insight into cost-effectiveness of the WGS-first approach. The prospective data collection will collect the number of diagnosed patients and time to diagnosis will be collected as effect measures for both pathways. A case report form (CRF) will be developed, thereby having a structured and uniform data collection. The acquired data will consist of all the health care professional visits, hospitalizations, imaging, genetic tests, metabolic investigations, and biochemical investigations and will be collected in relevant units in the CRF. In addition, the societal perspective implies that not only direct (health care) costs will be monitored but also caregiver indirect costs such as work absenteeism (for parent or relative or other caregiver if relevant). Productivity losses will be estimated using the iPCQ questionnaire18 adapted to the parent as subject of measurement. The friction cost-method will be applied following the Dutch guidelines. Data collection in the prospective cohort will entail a maximum of two years thereby not including the full diagnostic odyssey of patients and families currently observed in standard care pathways. Moreover, impact of having a diagnosis on medical decision making processes will due to the relatively short time frame of this study not be incorporated in the prospective collection. Retrospective data collection is therefore needed to complement the prospective data collection regarding these outlined issues as well.
    A more detailed description of prospective data collection is provided in Appendix 1.

    1.3.2 Retrospective data collection
    To complement the prospective data collection, we aim to include in a retrospective approach at least 200 patients in total per disease area who have had standard diagnostics. We aim to clinically match these patients with those in the prospective cohort. For all patients we aim to collect all resources used from day of first hospital visit until their last visit in all participating hospitals. Similar CRFs can be used as developed for the prospective data collection. Effectiveness measures will again be percentage of diagnosed patients and time to diagnosis.
    After inclusion of patients, all available resource use data will be retrospectively collected from the hospital information systems and patient records. These collective investigations constitute the complete traditional diagnostic pathway and enables to identify all resources used after having or not having a diagnosis.
    Genetic tests performed at other genetics centres will also be collected in referral letters from general hospitals or other university medical centres. Data regarding sort of diagnosis, age at first visit and total length of the diagnostic trajectory will be collected as well.

    1.3.3 Costs
    Prospective and retrospective data enables assessment of differences between diagnostic costs for both pathways but as well for all other resources used over time. Moreover, retrospective data collection enables a complete assessment of the impact of having a diagnosis on the medical decision making process. To determine the cost-to-diagnosis, all individual units of care in both the prospective and retrospective cohort will eventually be linked to their unit costs. Per modality (WGS or standard genetic testing) standard costprices will therefore be determined using the Dutch guideline19 or else real/full cost prices via bottom up micro costing. Reimbursement prices issued by the Dutch Healthcare Authority (NZA) and national reference prices will also be used for this assessment as outlined in current Dutch pharmaco-economic guidance.
  • 1.4. Patient-related outcome measures (PROMs)

    In addition to the cost analyses, we aim to investigate the merits and burdens of both diagnostic pathways, as the parents experience them, and how the parents value the outcome of the diagnostic pathway.
    For the first PROM, parents of both NDD and NICU patients will be asked to score satisfaction both diagnostic care pathways on a visual analogue scale (VAS)21 when answering the question "Are you satisfied with the diagnostic pathway your child has received?". Parents will provide their answer on a horizontal line of 100mm, with two descriptors representing extremes of satisfaction which they will mark with a vertical line The measurement in mm is converted to the same number of points ranging from 0 (not satisfied) to 100 (extremely satisfied) points. Secondly, GCOS-24, a questionnaire measuring the (health related) aspects of genetic counseling, will be used.22 Given the relative short time path in this study and the fact that we are dealing with (extreme) young patients, it is expected that an evaluation using the EQ- 5D23 - a generic health-related quality of life instrument - will not be informative, not even if parents act as proxies. Both, the VAS satisfaction and the GCOS-24 will be measured at definitive diagnosis or end of study. The responses from each survey will be transformed into utility and capability scores, respectively (using published values).
    Lastly, parents of critically ill newborns face a wide range of medical decisions, ranging from relative easy decisions such as placing a temporary catheter to decisions involving most difficult end-of-life issues such as do-not-attempt-resuscitation or placing a limit on life-supporting technology. Of note, as the treatment options for NDD patients are expected to be limited, the perspective analysis of the joint decision-making process will only be performed in parents of NICU patients. To assess the parental decision-making preferences in the high-stress environment of the NICU, parents will be asked to complete a standardized and validated psychometric questionnaire, SDM-Q-9, measuring the shared decision making process, directly after consultation. In addition, the clinician will simultaneously complete a similar questionnaire, SDM-Q-Doc, tailored at measuring the physicians perspective of the shared decision making process.
  • 1.5. Analysis

    We will investigate whether the use of WGS-first is cost-effective compared to the standard genetic testing in NDD patients from a societal perspective, based on the underlying rationale that NGS is able to reduce time-to-diagnosis and saves on direct medical consumption (e.g. futile further, potentially invasive, diagnostics and treatments). In this rationale, it is assumed that on average, parents prefer an (earlier) diagnosis and consequently assign value/utility to this knowledge (e.g. for further reproductive choices and/or joining patient support groups). The economic evaluation is based on the general principles of a cost-effectiveness analysis and uses input from the prospective and retrospective data collection. Outcome measures for the economic evaluation on patient level are i) costs (both direct (in and outside health care) and indirect), ii) number of definitive genetic diagnosis, and iii) time to diagnosis.

    1.5.1 Cost-to-diagnosis
    Both prospective as retrospective input will be used to assess the difference between diagnostic costs for both pathways (costs from first visit until having or not having a diagnosis). Assessment of this difference will be used as input for the cost-effectiveness analysis in 1.5.4.

    1.5.2 Impact on medical decision making
    To measure the impact of a diagnosis on medical decision making and costs, we will ask health care professionals to report follow-up exams and/or changes made to the standard care. For this analysis, questionnaires will be used that will serve as a guideline to interview neonatologists/clinical geneticists involved in clinical care of the patients enrolled in this study. The main question that will be addressed, for each patient individually, is "has knowledge on the genetic diagnosis changed/influenced clinical care provided to this patient?" That is, it may be expected that for some patients, the genetic diagnosis has led to referral to e.g. cardiologist for surveillance of cardiac performance, whereas without the knowledge of the genetic etiology, this referral may not have taken place. Especially for NICU patients, certain difficult to diagnose acute conditions may lead to withdrawal of further treatment or re-direction of care, whereas other diagnosis may lead to medical interventions. Changes in therapeutic strategies will be monitored and related to the health care costs for each individual patient. The output will be added to the CRFs. In addition to quantitative assessments, we aim to perform structured interviews with medical geneticists and other specialists to outline their viewpoints of approaches in changing patient disease management after diagnosis. Outlined changes will be compared to the collected retrospective dataset to envision whether changes have appeared and whether more guidance in patient management is needed. In addition, for analytical decision modeling, changes in health care resource are used for calculations.

    1.5.3 Time-to-diagnosis
    For time to diagnosis a time-to-event analysis will be performed, with definitive diagnosis being the event. NDD patients for whom no definitive diagnosis was made, contribute complete follow up time (two years) to the time of diagnosis as a censored observation. The mean time to definitive diagnosis saved over the duration of interest is the difference in area under the time-to-definitive diagnosis curve (Kaplan-Meier curve) of both the standard and WGS-first pathways.

    1.5.4 Decision analytic modeling
    Eventually outcomes regarding costs-to-diagnosis (1.5.1) and impact on medical decision making (1.5.2) will be combined in a cost-effectiveness analysis. Outcomes will be depicted as Incremental Cost Effectiveness Ratio’s (ICERs) which express the extra investment (if any) that is needed to achieve an extra diagnosis or a specified shortening of time to diagnosis.
    The decision analytic modeling is to find the most cost-effective place for WGS in the diagnostic process on the medium and longer term. These models will allow us to generate scenarios and guidelines when who to offer WGS in a diagnostic setting. At least four scenarios will be considered, as presented in “Toelichting kosteneffectiviteitsanalyse en budgetimpact” respectively, being i) current care, ii) immediate 100% WGSfirst, iii) gradual implementation WGS-first and iv) partial implementation of WGS-first.

    1.5.5 Uncertainty analysis
    Uncertainty surrounding difference in costs of diagnostic pathways and ICERs will be determined using the bootstrap method or the Fieller method. The impact of uncertainty surrounding deterministic parameters (for example prices) on the differences in total costs and the ICER will be explored using one-way sensitivity analyses on the range of extremes.
    Additionally, decision analytic modeling will be used to determine the costs that may be associated with knowledge of the diagnosis. That is, it may for instance be expected that based on knowledge of the genetic diagnosis, patients may be consulted by a cardiologist that otherwise would not have happened.

    1.5.6 PROMs
    The VAS satisfaction and GCOS-24 responses from each survey will be transformed into utility and capability scores, respectively (using published values). The relationship between the differences between SDM-Q-9/SDM-Q-Doc scores will be assessed in each care pathway by MANCOVA, student's t-test and Pearson correlations.

  • 1.6. Budget impact analysis (BIA)

    The aim of this BIA is to perform an assessment of the financial consequences of implementing a (rapid) WGS-first approach, as a substitution for standard genetic testing, in the Dutch health care system in the short-to-medium term from the budget holders perspective (for example, Health Care/VWS, Third party payers).
    The BIA will apply to two model disorders:
    NICU – The prevalence and incidence of NICU patients will be investigated/updated by use of epidemiological data in the Dutch nationwide and local context. In the time period 2011-2013, on average, 4,100 newborns were admitted to the NICU each year. With at least 6-8% of NICU patients having a (rare) genetic disorder30, each year, ~287 patients are to be expected. We performed a preliminary analysis of the global average cost estimate for the subset of patients with an expected genetic disease, taking into account hospital costs (in-hospital days) as well as genetic counseling and testing (prices according to CVZ19). On average, the costs involved in the standard care pathway are €40,500 per patient per annum (Bijlage 6-1 ‘Doelmatigheidswinst NICU’). The total budget impact for all NICU patients with an expected genetic disorder thus totals to €11,623,787 per annum. The cost-estimate for the WGS-first care consumption is expected to be € 32,974 euro/patient/annum, with savings introduced due to shorter stays at the NICU (2 days on average, but maintaining the total in-hospital days), and one generic rapid WGSfirst genetic test. Assuming that WGS-first substitutes 100% of requested genetics tests, the yearly consumption in the new situation would be €9,463,538, thereby introducing a budgetary saving of €2,160,249 each year (18.6% of current estimated costs). Of note, calculation of this budgetary savings is relatively conservative given that is has not taken into account savings due to futile testing and/or shorting of the total time spent in-hospital, as these savings, if any, will differ for each patient. NDD – We performed a preliminary analysis of the global average costs estimate for NDD, taking into account that annually, 2,500 new NDD patients are seen in clinic. Costs in standard care in the Dutch health care system are based on reported costs-to-diagnosis20,31 supplemented by estimates of productivity loss, which amount to be €15,386 per patient per annum (Bijlage 6-2 ‘Doelmatigheidswinst NDD’). The total budget impact for all NDD patients thus totals €38,465,000 per annum. The cost-estimate for the WGS-first care consumption is expected to be €8,885 euro/patient/annum. Assuming that WGS-first substitutes 100% of requested genetics tests, the yearly consumption in the new situation would be €22,212,500, thereby introducing a budgetary saving of €16,252,500 each year (57% of current estimated costs). When performing the BIA in this project, we will provide more precise (and updated) estimates for both NDD and NICU patients. The BIA adheres to the new guidelines (Zorginstituut, 2015) and applies the perspectives: societal, health insurance/third party payer and health care (Budgettair Kader Zorg (BKZ)). All of these perspective related care consumption is collected in WP1. We will investigate the prevalence and incidence of NDD by using epidemiological data in the Dutch and local context to determine the average cost estimate for NDD patients. The BIA Prices will be linked to perspectives using the data collected in the CEA, including societal-CEA based prices, BKZ-average rates according to NZa, for health insurance perspective also NZa average rates and, for example, for a local health care provider perspective specific passenger rates (‘passanten tarieven’).


  • D1.1 Harmonized WGS sequencing procedure (M3)
  • D1.2 Sequencing of cases complete (M27)
  • D1.3 Retrospective data collection (M12)
  • D1.4 Report on obtained diagnostic yield, achievable TAT (M27)
  • D1.5 Analysis of medical impact (M27)
  • D1.6 Cost-to-diagnosis (M30)
  • D1.7 PROMs evaluating the WGS-first approach (M36)
  • D1.8 Budget impact analysis (M36)
  • D1.9 Decision analytical models (M36)