Implementation of rapid genomic sequencing in safety-net neonatal intensive care units: protocol for the VIrtual GenOme CenteR (VIGOR) proof-of-concept study

Study setting and overview

The VIGOR study is run out of Boston Children’s Hospital (BCH), and the study team includes a diverse group of investigators with expertise in neonatology, genetics and genomics, genetic counselling, bioinformatics, health services research, implementation science and social disparities. Currently, there are five participating NICU sites (Baystate Medical Center, Boston Medical Center, The Children’s Regional Hospital at Cooper, UMass Chan Medical Center and The Women’s Hospital at Renaissance) and three additional sites are in the process of joining. Together, these sites comprise the VIGOR network (figure 1). To reach enrolment goals (n=250) and improve the diversity of our study sites, the VIGOR network is actively recruiting additional sites across the USA.

Figure 1
Figure 1

VIGOR study overview. The VIGOR study is run out of BCH and recruits subjects from five participating NICU sites. Enrolled infants receive clinically accredited rGS, the VIGOR team creates clinical interpretive reports, and local providers return results to families. Comprehensive implementation outcomes are examined using data collected from families, providers, and infant medical records. BCH, Boston Children’s Hospital; BMC, Boston Medical Center; CIR, clinical interpretive report; DHR, Doctor’s Hospital at Renaissance; NICUs, neonatal intensive care units; rGS, rapid genomic sequencing; VIGOR, VIrtual GenOme CenteR.

Neonatal providers at each site receive basic education about genomic medicine and identify eligible infants; the BCH-based VIGOR team enrols those eligible infants and their parents (Aim 1). Enrolled infants receive clinically accredited rGS, the VIGOR team creates CIRs based on the genomic sequencing results and local providers return the results to families. Additionally, the VIGOR team performs reanalysis of the sequencing data for infants with initially non-diagnostic results (Aim 2). Families are followed for 12 months after NICU discharge. Data is collected from infant medical records, families and providers and used to examine comprehensive implementation outcomes (Aim 3).

The VIGOR study is registered on clinicaltrials.gov (as of 25 January 2022) and is actively recruiting as of October 2022. The projected end date of recruitment is November 2025.

Aim 1 procedures

NICU site eligibility criteria

Safety-net NICUs are invited to participate in this study. Generally, safety-net centres predominately serve socially disadvantaged families.34 For this study, we defined safety-net centres as those serving >40% non-Hispanic black or Hispanic patients and thus also likely serving a high percentage of low-income families. These demographic groups have historically not been reached in genetic studies. Participating NICU sites are also required to meet the following criteria: (1) level 3 NICUs, as these NICUs often serve as the first site of evaluation for newborns with suspected genetic conditions, (2) NICUs with a minimum of 20 beds and 250 admissions per year, to enable goal enrolment and (3) NICUs not routinely able to offer rGS. A diverse group of sites meeting these criteria have been selected so far varying in unit size, area served (rural vs urban) and proportions of non-Hispanic black and Hispanic patients served.

Subject eligibility criteria

In this study, neonatal providers, infants and their families (parents and/or other caregivers) are study subjects. We provide details on eligibility, recruitment, consenting and retention below.

Neonatal providers: All neonatal providers at participating sites who are involved in the care of critically ill infants with probable genetic conditions are invited to participate in the study. This includes all neonatologists, and, at the discretion of the site principal investigator (PI), neonatal nurse practitioners and physician assistants.

Families: Infant inclusion and exclusion criteria are designed to be primarily phenotype driven (box 1) and are based on our previous phenotype-based study of rapid exome sequencing (ES) in the BCH NICU.17

Box 1

Infant inclusion and exclusion criteria

Inclusion criteria:

  • Infant is admitted to the neonatal intensive care unit (NICU) at a participating site

  • Infant is suspected to have a genetic condition, including, but not limited to:

    • Unexplained seizures

    • Unexplained hypotonia

    • Multiple congenital malformations

    • Metabolic disorders

    • Disorders of sex development

    • Interstitial lung disease

    • Immunodeficiency

  • At least one biological parent is available for consent and participation

  • Caregiver(s) intend to care for the infant in the USA for 12 months following NICU discharge

Exclusion criteria:

  • Infant has a prenatally known genetic diagnosis

  • Infant has clinical features pathognomonic for a recognisable chromosomal abnormality (eg, Trisomy 21)

  • Infant has an association of symptoms known to have a low genetic diagnostic yield (eg, vertebral defects, anal atresia, cardiac defects, trachea-oesophageal fistula, renal anomalies and limb abnormalities association)

  • Infant’s presentation is likely due to an environmental or other non-genetic cause (eg, fetal alcohol syndrome)

  • Infant has a known family history of genetic disease that is plausibly the cause of the infant’s phenotype

  • Infant is deceased, transferred or discharged from the NICU prior to enrolment

  • One or more parents object to the infant’s enrolment

Biological parents of the infant are eligible to participate in the rGS portion of the study. Primary caregivers are eligible to participate in follow-up surveys regardless of biological relationship to the infant, to accommodate diverse family structures (eg, same sex caregivers, use of donor egg or sperm, etc)

For this pilot study, we are limited to approaching English-speaking and Spanish-speaking families only. Additionally, due to the complexities of enrolling infants in state custody and the varying state laws which would have to be considered due to the locations of participating sites, we are not enrolling infants in state custody or who are intended to be placed in state custody.

Notably, the study team encourages local providers to pursue any typical clinical care procedures that exist at their site for work up of infants with suspected genetic conditions, such as consultation with clinical genetics and/or other available genetic testing (ie, the VIGOR study is not meant to change existing clinical care).

Subject recruitment

Neonatal providers: Providers at each site are recruited through introduction by the site PI. Prior to approaching families regarding the study, neonatal providers are required to watch an orientation video explaining infant eligibility criteria and general study procedures and complete an attestation statement that orientation has been completed.

Families: Eligible infants are identified by neonatal providers at participating sites. The BCH VIGOR team is available for any questions regarding eligibility. Once an eligible infant is identified, a neonatal provider at the site approaches the family and introduces the study. If the family is interested, the NICU site team fills out a referral form and sends it securely to the VIGOR team. A VIGOR team member then reaches out to the family to set up a consent session via video conferencing or phone call.

Subject consent

Neonatal providers: Provider consent is implied through the completion of provider surveys (see survey details in Aim 3 procedures below).

Families: Interested families complete a preenrolment information session with a BCH VIGOR team GC or research assistant (RA) via video conferencing or phone call. During this session, the BCH VIGOR team member reviews study logistics, including possible risks and benefits of enrolling, provides an overview of the utility of rGS for critically ill infants and explains potential types of results. The consent form inquires about contacting the family about future research studies and reporting of American College of Medical Genetics and Genomics secondary findings.35 Consent is obtained electronically, and the study team uses certified Spanish interpreters for sessions with Spanish-speaking families.

Neonatal provider education

As part of the VIGOR study, neonatal providers at participating sites receive virtual education in genomic medicine. The BCH VIGOR team created a series of eight training modules available on the publicly accessible VIGOR web platform (https://www.virtualgenomecenter.org/education). VIGOR investigators with expertise in neonatal genomic medicine deliver the training modules with content geared towards neonatal providers with basic training in genetics and genomics. The sessions include: (1) VIGOR orientation, (2) ‘Genetics 101’, (3) recognising a patient with a potential genetic diagnosis, (4) current genetic diagnostic tools, (5) genomic sequencing results, reports and variant interpretation, (6) breadth of genomics in the neonatal period, (7) mock consent and disclosure sessions and (8) summary. The VIGOR team also provides one on one or group site training tailored to site needs for the duration of the study as needed.

Subject retention

Neonatal providers: Providers at participating sites can receive continuing medical education credits for watching the genomic medicine education modules and receive Amazon gift cards on completion of surveys and interviews.

Families: Families receive gift cards of graduating amounts on completion of follow-up surveys. The subset of families selected for qualitative interviews receive an additional gift card incentive.

Aim 2 procedures

Rapid genomic sequencing

A commercial clinically accredited vendor (GeneDx, Gaithersburg, Maryland, USA) performs the rGS (we used ES for the first year and then switched to genome sequencing as sequencing costs decreased over the study period).36 Local NICU staff collect a blood sample from the infant and buccal samples from enrolled biological parents using kits provided by the vendor. The site PI or RA orders the rGS test through the vendor portal as a clinically accredited test and samples are shipped from the study site to the vendor. To inform rGS analysis, the site provides relevant clinical information to the vendor. Additionally, the study site provides an infant medical history form to the VIGOR team and parents are sent a baseline survey regarding pregnancy and other medical history by the VIGOR team. After receiving all samples, the vendor provides a verbal preliminary rGS result within 7 days and a written rGS report within 14 days to the VIGOR team and the ordering provider at each site (figure 2). The rGS report from the vendor provides the result (eg, positive, negative, inconclusive) and limited information regarding potential next clinical steps.

Figure 2
Figure 2

rGS workflow. Clinically accredited rapid genomic sequencing is performed using samples collected by local neonatal intensive care unit staff from the infant and biological parents when available. The vendor provides a written report within 14 days and the BCH VIGOR team creates a comprehensive interpretive report to provide guidance to neonatal providers. Both reports are sent to the site and local providers disclose results to families. Negative or inconclusive results are reanalysed over the course of the study and positive results may be used to direct future care and resource management. ASO, antisense oligonucleotide; BCH, Boston Children’s Hospital; rGS, rapid genomic sequencing; VIGOR, VIrtual GenOme CenteR.

Clinical interpretive report (CIR)

Thus, the BCH VIGOR team creates a second report, a CIR, to provide guidance to the neonatal provider based on the rGS results. The VIGOR team uses available clinical information for the patient and genomic resources to create the CIR. The CIR interprets diagnostically relevant findings from the vendor report and links those findings to clinical management considerations, research opportunities and family resources. For a non-diagnostic or inconclusive finding, the CIR provides additional variant interpretation as relevant and describes plans for reanalysis. For example, for a variant of uncertain significance, the VIGOR team interprets the findings as likely, indeterminate or less likely of clinical relevance. The VIGOR team’s goal is to complete the CIR within three business days of receiving the vendor report. The CIR is sent to the study staff at the site via secure email and is then internally sent to the disclosing provider at the site. The VIGOR team is available to answer any questions about the information in the CIR. Notably, while the VIGOR team members are not clinical providers for enrolled infants, clinical providers at the sites may use the CIR in the context of an infant’s current clinical presentation.

Disclosure

If the preliminary rGS result indicates an urgent, life-threatening or immediately actionable condition, a BCH VIGOR team member immediately discloses results to the study site neonatal provider and family. Otherwise, a neonatal provider at the study site discloses the rGS results to the family, even if the infant has been discharged or transferred from the NICU prior to the results return. We planned for this method of disclosure for two reasons: first, disclosure from a trusted provider in a familiar environment may be the best approach for families; and second, experience with disclosure of genetic testing results may increase neonatal providers’ comfort with genomic medicine in the NICU. It is at the discretion of the neonatal provider caring for the infant and family to choose when to disclose the rGS results (after receiving the preliminary result, final vendor report and/or CIR), with whom to disclose results (clinical geneticist or GC at their site, if available or any other member(s) of the care team at their site (nurse, social worker, additional subspecialist, etc)) and how to disclose results (in person, by phone call, or by video conferencing). A VIGOR team member is available to answer questions and provide support to the neonatal provider through the disclosure process and is present during the primary disclosure session, but does not communicate directly with families about the results.

Reanalysis

The vendor sends the sequencing data to BCH where it is securely stored. For cases with initially negative or inconclusive rGS results, the BCH VIGOR team reanalyses the data within 3 months of obtaining it and annually thereafter for the duration of the study. Reanalysis can reveal new diagnostic findings as we learn new information about the patient’s clinical features, genomic analyses methods and gene–disease associations over time. The sequencing data is reanalysed by the VIGOR team using a custom in-house bioinformatics pipeline, Variant Explorer.37 If clinically relevant results are identified by reanalysis, the results are clinically confirmed by the vendor and updated reports are created by the vendor and VIGOR team. The VIGOR team communicates the updated results to the original ordering provider at the site and/or to the infant’s primary care physician.

Aim 3 procedures

Implementation framework

This study uses the Proctor Implementation Framework to guide comprehensive evaluation of implementation outcomes (figure 3).38 The Proctor Framework is based on the principle that both the evidence-based intervention and the strategies used to implement the intervention are needed to improve health outcomes. The framework incorporates three levels of outcomes: implementation, service and client. Using a mixed methods approach, we assess the implementation outcomes of appropriateness, feasibility and penetration, the service outcome of equity, and the client outcomes of function, symptomatology and satisfaction described in table 1.

Figure 3
Figure 3

Adapted Proctor Implementation Framework. Evaluation of implementation outcomes will be based upon the Proctor Implementation Framework. The framework posits that both the evidence-based intervention and the strategies used to implement the intervention are required to improve health outcomes. Three levels of outcomes are incorporated: implementation, service and client. rGS, rapid genomic sequencing; VIGOR, VIrtual GenOme CenteR.

Table 1

Outcomes and measurements

Data measures

This study collects data through qualitative and quantitative methods to assess outcomes at the neonatal provider, infant and family levels (table 1).

Data collection procedures

Qualitative component

Virtual focus groups: A study PI conducts focus groups of neonatal providers at participating sites via video conferencing before and after implementation of VIGOR at each site. These focus groups are approximately 60 min in length.

Qualitative interviews: A study PI or BCH VIGOR team member trained in qualitative analysis conducts qualitative interviews for neonatal providers and families via video conferencing. These interviews are approximately 30–60 min in length. Interviews with neonatal providers are conducted after a provider has completed 3–5 results disclosures. Interviews with families are conducted within 2 months of results disclosure. We are using purposive sampling to select a subset of approximately 60 families to ensure breadth of perspectives, including English-speaking and Spanish-speaking families, families across study sites and families with infants with positive, negative and inconclusive genomic sequencing results.

The study investigators create open-ended question guides to assess constructs of interest in focus groups and interviews and revise them as needed as part of the iterative analytic approach. Focus groups and interviews are conducted by study team members trained in qualitative interviewing and are transcribed verbatim.

Quantitative component

Chart abstraction: Chart abstraction is performed at the time of infant enrolment, at the time of infant transfer/discharge from the NICU and 12 months after results disclosure. A study team member at each site completes chart abstraction forms which the BCH VIGOR team enters into a centralised REDCap database.

Surveys: All neonatal provider and caregiver surveys are administered by email or text message by REDCap. We send caregivers surveys at enrolment, within 2 weeks of results disclosure and 3, 6 and 12 months after results disclosure. After delivery of the initial survey at each time point, we send reminder surveys weekly for a maximum of 3 weeks. We send neonatal providers surveys before and after completion of virtual genomic medicine training and after each results disclosure.

Outcome definitions

Implementation outcomes

The implementation outcomes include: (1) penetration, which is the primary outcome and is defined as the percentage of eligible infants that are enrolled, have rGS sent (on the infant with or without biological parents), have vendor rGS report delivered, have CIR delivered and have a family disclosure session; (2) appropriateness, which is defined as neonatal providers’ perception of the value of rGS and use of VIGOR in optimising NICU clinical care and (3) feasibility, which is defined as neonatal providers’ perception of the ability of VIGOR to be implemented as designed in their NICUs.

Service outcome

Equity is defined as equal penetration by infant race/ethnicity, infant insurance status (public/private/uninsured) and language preferred by family for study communication (English/Spanish).

Client outcomes

Client outcomes include function, symptomatology and satisfaction. We examine function at the neonatal provider level, as comfort with caring for infants with likely genetic disorders, including identification, testing, disclosing results and management. We also examine function at the family level as two common mental health outcomes: stress measured according to the Perceived Stress Scale39 and depression measured according to the Edinburgh Postnatal Depression Scale.40 We measure symptomatology at the infant level over the first year post-NICU discharge. We will assess postdischarge healthcare utilisation with measures adapted from the National Survey of Children with Special Health Care Needs41 and other clinical outcomes, including genetic diagnosis, management changes (eg, referrals, lab tests, imaging, treatment changes, procedures, surgeries), palliative care decisions and/or death, and family planning decisions. At the neonatal provider level, satisfaction including perceived comfort with the process of results disclosure, the level of VIGOR team support and perceived barriers and facilitators to implementation of VIGOR is measured. At the caregiver level, we measure satisfaction with the decision to enrol in the study and receipt of genomic sequencing results by neonatal providers.

Data analysis and power considerations

Qualitative component

The qualitative analysis uses a grounded theory iterative approach.42 Study investigators with broad expertise in genomic medicine, neonatal care, social disparities, experience working in safety-net NICU settings and qualitative research, independently review transcripts and meet to develop and revise a uniform codebook. Investigators then code transcripts and convene to assure uniform coding and resolve any discrepancies through group discussion. Themes are developed and revised from coded manuscripts iteratively until no new themes emerge (thematic saturation). Investigator triangulation is used to assess reliability, whereby investigators read transcripts independently and subsequently reach consensus through group discussion, and by member checking, whereby findings are communicated to participants to confirm results.43

Quantitative component

As the goal of this ‘proof of concept’ study is to assess implementation of a novel delivery system for genomic care, we do not have a control group, and therefore the quantitative analysis will be primarily descriptive. We will estimate penetration, the primary implementation outcome, by calculating the percentage of enrolled subjects completing the genomic sequencing and disclosure process as described above. To assess equity, we will use contingency-table statistics to compare penetration among subgroups (infant race/ethnicity, infant insurance status, language preferred by family for study communication). Measures to be collected at multiple time points include clinical outcome, family mental health and family and provider satisfaction (table 1), with most family surveys administered to both parents. We will analyse the longitudinal outcomes with repeated-measures linear regression for continuous outcomes and logistic regression for discrete outcomes. Each model will include random effects to account for serial correlation, within-family correlation and within-site clustering as well as fixed effects by time, subgroup and parent. We will construct contrasts from the fitted model parameters to describe the time course of each outcome and to test hypotheses comparing the mean response and the time course among subgroups.

In the planned sample of 250 enrolled infants and their families (N), we anticipate 80%–90% penetration (P). The estimate of achieved penetration rate will therefore carry precision of 1.9%–2.5%. For assessing equity, the detectable difference in penetration between subgroups (eg, by insurance status) depends on the subgroup sizes; however, we anticipate that with subgroup sizes between 0.2 and 0.4, the difference in penetration detectable with 80% power and 5% type I error will lie between 11% and 18%. The above estimates give us confidence that the planned sample size will provide ample precision for its descriptive aims and sufficient power to demonstrate relatively subtle subgroup differences and changes with statistical significance.

This post was originally published on https://bmjopen.bmj.com