Improving the Infrastructure for Pediatric Medical Device Trials at Children’s Hospitals

Medical device innovation has produced critical new technologies to improve the way we prevent, diagnose, and treat diseases. Unfortunately, children experience this innovation differently than adults. Few devices have been designed and tested specifically for pediatric applications, resulting in children having fewer treatment options and being exposed to inconsistent benefit-risk profiles from the necessary off-label use of adult devices. Children also experience delays in access to technologies, and sometimes years to decades pass before a pediatric version of an adult device becomes available.¹ There are several scientific, clinical, regulatory, and financial barriers to making progress that have been previously described in detail.² Although federal agencies such as the National Institutes of Health (NIH), Food and Drug Administration (FDA), and Centers for Medicare and Medicaid Services (CMS) are responsible for discrete tasks such as funding device-related research, regulating devices, and guidance device reimbursement, there is no national-level network to support and organize systems-level solutions to these barriers.

One particular area that requires attention is the infrastructure supporting pediatric medical device (PMD) trials. Improving the various clinical research processes that support the development of novel PMD technologies is 1 key component of addressing this gap. Pediatric drug trials are well established and can be mandated through the Pediatric Research Equity Act, which gives the FDA authority to require studies for certain drugs and biologics. On the other hand, PMD clinical trials are not mandated, are iterative by design, and often require collaborations between academia and industry. These features and others contribute to a high level of complexity that has not been amenable to some of the programmatic solutions that have been implemented for pediatric clinical research, such as the Pediatric Trial Network,³ which heavily focuses on drug studies. To date, very little is known about best practices or programmatic features that set up an institution to successfully support PMD clinical trials.

The Pediatric Device Consortia (PDC) Program is an FDA initiative established by Congress in 2007 to fund nonprofit consortia to help stimulate projects that promote the development and availability of PMDs. Across 4 grant cycles, the FDA has awarded 23 consortia grants that have collectively supported over 1000 PMD projects across the United States. In 2021, Congress tasked the current PDCs and the FDA to evaluate the challenges associated with conducting PMD trials and potential solutions for improving associated infrastructure. Over the course of 2 years, the PDCs reviewed current clinical trial practices, interviewed key stakeholders, reviewed the available literature, and launched several clinical trials and clinical support initiatives. In this special article, we will describe the PDCs’ experience with and perspective on supporting PMD research, including barriers and facilitators, and make recommendations for a nationally applicable strategy to enhance research infrastructure.

To provide context to the information and recommendations presented here, details about the PDC host institutions during the 2018 to 2023 grant cycle are presented in Table 1. The data presented are from 2021, the midpoint of the 5-year grant cycle. Although each consortium has several affiliated sites, most PMD trials take place at the 7 institutions listed in Table 1. The rest of the information presented in this manuscript represents the activities and experiences at those 7 institutions and may not be reflective of the resources, infrastructure, or experiences at other PDC-affiliated sites.

The PMD development process typically starts with identifying an unmet clinical need, followed by creating and testing prototypes with key stakeholders. The design undergoes iterative improvements, and final testing focuses on biocompatibility, safety, and quality. Depending on the device, animal and human testing may be required, with clinical studies needing institutional review board (IRB) approval and often an investigational device exemption. Once evidence generation is complete, the company finalizes their FDA submission to obtain a marketing decision. The FDA uses a risk-based framework for regulating medical devices, and this risk classification determines what kind of clinical evidence is required, if any.

Low-risk class I devices that present minimal potential for harm to children are typically exempt from requiring any clinical data and must still go through an IRB review if the sponsor is collecting research subject data for publications, marketing, or FDA registration. Sponsors of most moderate-risk class II devices must identify a predicate and demonstrate that their device is substantially equivalent to that in the market. They do so by demonstrating equivalency in bench studies. In the context of PMD development, however, sponsors often add clinical study data to their research and development for fundraising and reimbursement strategy purposes. Moderate-risk devices for which there is no legally marketed predicate device, must undergo the de novo 510(K) regulatory pathway, often necessitating rigorous clinical studies. High-risk class III devices require full clinical studies to demonstrate safety and efficacy. The Humanitarian Device Exemption (HDE) pathway is similar to class III, except clinical evidence only needs to demonstrate that the device does not pose an unreasonable risk and has probable benefit. This pathway was created for devices used in conditions that affect less than 8000 people per year for which it might be nearly impossible to adequately power a clinical study to show efficacy. The class II de novo, class III, and HDE devices represent the most potential for innovation when it comes to generating novel methodologies and using data from both clinical trials and nonclinical trials sources.

Although the traditional 4-phase model of drug development is frequently taught throughout health care–related training programs, neither the FDA nor industry use the 4-phase model for the development and regulation of devices.^4–6 In general, the FDA requires different types of evidence depending on the intended use of the device and its overall risk classification. Clinical evidence is not always required, and when it is, those studies are typically defined as either feasibility studies (intended to evaluate the design, functionality, or overall concept of a device) or pivotal studies (intended to provide evidence of safety and efficacy).⁷ The FDA has different expectations of device trials compared with drug trials, but this subtlety is often overlooked; indeed, multiple reviews, textbooks, and public websites about clinical trials collapse drugs and devices into the same framework, perpetuating this misconception.^8,9 Characteristics of pediatric device trials that are unique and different from drugs include the following:

• Sample sizes in device studies are smaller than in drug studies. According to Quazi et al, in their study of all 2017 to 2022 PMD trials, they found that 51.6% enrolled fewer than 50 participants and 71.0% fewer than 100 participants.¹⁰ These numbers are even lower for high-risk PMD (2016–2021) trials, where the median enrollment was 14 patients and the range was 7 to 49 patients.¹¹ In stark contrast, the data from 2007 to 2014 drug trials indicate that only 56.1% enrolled fewer than 100.¹² 

• Device development is iterative. To add to this complexity, marketed pediatric devices may need to be often replaced due to the growing physiology of children. In the case of adults, a replacement may be seen as a failure that the device did not perform as expected.

There are multiple clinical evidence generation strategies for medical products. Although randomized controlled trials (RCTs) are the gold standard for drugs, for medical devices, the FDA (following an alignment with the manufacturer) often allows open-label, nonrandomized studies to support device marketing applications. This is well reflected in the 482 clinical PMD studies between 2017 and 2022, of which 52.9% were nonrandomized and 72.6% were open label.¹⁰ PMD clinical evidence is critical, but it is often generated through smaller and iterative evidence generation plans that incorporate nuanced and flexible study designs, instead of RCTs.¹¹ Use of real-world data has also been introduced to support PMD approvals.¹³ 

It is estimated that 60% to 75% of all medical products are used in pediatrics without prior clinical validation or a structured risk-benefit analysis that typically accompanies FDA’s clearance and approval processes.¹⁴ Challenges for PMD trials include smaller pediatric patient sample sizes and population heterogeneity, which limit trial recruitment.² Indeed, 32% of all study discontinuations were due to low patient accrual as indicated by the analysis of registered studies at clinicaltrials.gov between 2017 and 2022.¹⁰ As a vulnerable population, children are rarely considered for clinical trials during the initial phase of device development. Even high-risk devices with specific pediatric indications evaluate safety and effectiveness in children only 60% of the time.¹¹ Compounding this issue, FDA has rarely required postmarket studies to include children, despite having the authority to do so.¹⁵ Thus, PMDs greatly lag behind their adult counterparts¹ and adult devices continue being used in children off-label.^16,17 Table 2 provides a summary of known PMD clinical trial challenges as documented by our group and others. On balance, there has been significant progress made since the 2007 Pediatric Medical Device Safety and Improvement Act, and several new solutions have been proposed (and are actively being explored) to improve the PMD clinical trial infrastructure (Table 2).

To arrive at our final recommendations, we reviewed all of the information we gathered from the literature, key stakeholder interviews, and our own institutional practices, and identified core concepts that were relevant to PMD clinical trials (Figure 1). We grouped these concepts into 5 major domains: internal or local regulatory considerations, FDA regulatory considerations, financial considerations, logistical considerations, and data considerations. Once domains were identified, new concepts were added to the list. The list of concepts and domains were iteratively refined as new insights were added. Finally, to make our recommendations more practical, the domains were then translated into 3 main areas of focus for recommendations: pediatric device clinical trials units (CTUs), pediatric device trial initiation bottlenecks, and data considerations for pediatric device trials (Figure 1). We focused our recommendations on the core infrastructure that children’s hospitals, research institutes, and academic medical centers should provide to accelerate PMD trials. That said, it is also important that organizations cultivate a workforce of clinicians, investigators, and research professionals that are knowledgeable and interested in PMD development.

The IRBs, scientific committees, and CTUs at most academic medical centers (including pediatric institutions) are typically more familiar with evaluating and conducting drug trials than device trials in terms of ethical concerns, safety, and efficacy. However, as discussed previously, drug trials and device trials have significant differences and the knowledge of one cannot be easily applied to the other. This knowledge gap is even more concerning when PMD sponsors face delays or rejection of their research protocol because it was reviewed through the lens of drug trials. Other sponsors experience delays related to contracting, recruiting, hiring, budget negotiations, IRB amendments, and other logistical concerns. We encourage institutions to establish a dedicated unit to support device trials of all risk profiles. Pediatric device CTUs do not necessarily need separate physical space or staff to run; they can be administrative and process elements that connect existing clinical trial infrastructure in a way that is specific for PMDs. Institutions can consider starting with 1 or 2 staff with device-specific expertise who can liaise and coordinate across the various parts of the organization (eg, contracting, budget, cohort discovery, protocol development, IRB). Institutions can consider exploring partnerships with Clinical and Translational Science Awards (CTSA) Programs, engineering schools, regulatory science departments, or incubators and accelerators to develop device-specific expertise. Over time, institutions may consider investing additional resources to grow these programs and bring more sponsored device trials to the organization. Pediatric device CTUs can ultimately improve the current state by providing technical, scientific, and logistic support to physicians, sponsors, research subject participants, IRB, and scientific reviewers.

Children’s hospitals dedicate significant resources to the oversight of clinical trials to minimize patient risk and institutional risk that are inherent to any novel clinical trial. This oversight, combined with less experience with PMDs, has the unfortunate side effect of creating several bottlenecks that can hinder the initiation of pediatric device clinical trials and have led to approval wait times between 3 and 12 months. Trial initiation can be greatly streamlined by having the regulatory, financial, and contracting workstreams occur concurrently (and ideally within a dedicated pediatric device CTU) to support sponsors and clinicians (Figure 2).

The following best practice recommendations may assist in overcoming the challenges in initiating PMD clinical trials:

• IRB approval: For multisite pediatric device clinical trials, the use of reliance agreements/“single IRB” similar to the current requirements for NIH-supported trials would be beneficial because the smaller pediatric patient populations often require multiple study sites for sufficient patient enrollment numbers. In addition, the availability of pediatric device experts on IRBs and the training of reviewers on pediatric device trials is needed with an emphasis on the differences between drug and device clinical trials.

• Budget considerations: Pediatric device–specific clinical trial budget templates that reflect the type of costs and activities that are unique to pediatric device clinical trials are needed. Specifically, unlike drug studies that are conducted to determine safety and tolerability and often require pharmacokinetic and pharmacodynamic studies, device studies are often conducted to test the safety and performance of the device and look for more patient-focused outcomes, often incorporating human factor studies. Therefore, when delineating research vs standard-of-care elements of the budget, device CTUs can add unique expertise.

• Standardized clinical trial agreements/rental agreements: Preapproved language by all hospital stakeholders in the agreement templates (legal, supply chain/purchasing, etc) would reduce the processing times for these documents and expedite trial initiation. In addition, standardization of pediatric device clinical trial agreements among children’s hospitals would avoid duplicative reviews similar to the benefits of reliance agreements for IRB approvals with NIH-supported trials.

• Centralized clinical research support: Dedicated pediatric device CTUs can serve as a centralized research support organization at children’s hospitals that essentially creates a “one-stop shop” with device trial-specific services for innovators and clinicians. In addition, the unit’s pool of research coordinators with device trial experience and part-time availability for a particular project can help to accelerate trial initiation and patient enrollment.

The data needs of PMD research can be varied, ranging from simple cohort discovery and patient demographics to complex device data extraction and modeling. Institutions who intend to support PMD research should have well-established data and informatics services, with all the appropriate security and compliance measures required for the protected health information and regulatory submissions. The common goal is the determination of safety and performance of the investigational device. Early-stage PMD innovators may not have the capacity or expertise needed to understand, plan, implement, and analyze a data management plan and would benefit from guidance and consultation services from the academic hospitals at which they plan to conduct a PMD study. The following recommendations might prevent or alleviate PMD issues:

• Cohort discovery: In determining whether a site meets the recruitment goals, cohort discovery is critical to determine whether there are enough patients who meet the inclusion criteria. The process of cohort discovery uses data systems to identify historical patient accrual rates to infer a hospital’s ability to successfully recruit eligible study participants. Common cohort discovery tools include TriNetX, i2b2, enterprise data warehouses, national research collaboratives such as PCORnet and PEDSnet, and population health management platforms such as HealtheIntent. These tools facilitate study feasibility assessment at a given site, and their inclusion into a decision-making process is poised to reduce high numbers of study discontinuations.

• Data coordination and management: Institutions hosting medical device trials require robust and repeatable processes and systems to support data collection, management, and coordination. These include clinical trial management systems, electronic data capture tools such as REDCap, and enterprise data warehouses or data lakes. These last 2 examples enable broad access to curated retrospective and prospective data products. Occasionally, sponsors provide their own data collection form, which most institutions can accommodate, but this requires evaluation and clearance by information security. Delivering data to sponsors can be a cause of delays, especially when it involves historical or clinical data collected at the institution. This process often requires several different agreements (eg, data use agreements, collaboration agreements), and there is sometimes a lack of defined workflows to clarity on which stakeholders and governing bodies need to be involved, resulting in the issue being tackled ad hoc rather than systematically.

• Information security and regulatory requirements: Evaluation of the safety of devices and device data, tracking devices, and collecting data for regulatory submissions are critical to supporting medical device research, and robust information security review protocols for data and connected devices are necessary. CFR 21, part 11, compliance, the standard that governs data intended for regulatory submissions, can be handled in different ways, either through a sponsor-provided platform or by maintaining an institutional solution. Tracking devices are typically handled by the research teams through shipping and accountability records, clinical teams tracking device labels and serial numbers, and biomedical engineering tagging and logging devices by serial number and expiration dates where applicable. The unique device identification system implementation has not been broadly adopted by health care systems but could likely provide an excellent solution for close tracking of devices during clinical trials.^18,19 

Successful PMD clinical trials rely on robust data management and coordination, secure information systems, and adherence to regulatory requirements. Institutions conducting such trials must have well-established processes and tools to accelerate progress and ensure the safety and efficacy of investigational devices. Providing guidance and support to early-stage innovators who may lack expertise in these areas can further improve trial outcomes and efficiency.

PMD development is complicated by several clinical, technical, regulatory, and financial barriers.¹ These issues extend to the pediatric academic medical centers that play a key role in device research and development. Most of the sponsored research infrastructure, training, contracting, and budgeting at these centers has been built around drug trials. Unfortunately, the significant differences in regulation,^4,5 development,^14,15 and financing^20–22 between drugs and devices mean that there is limited value in translating approaches from one to the other. As a result, even nationally recognized research institutions with a strong record of successfully completing drug trials struggle to initiate and conduct device trials.

The PDCs are in a unique position as they work directly with PMD innovators, their home institutions, and the FDA to address these shortcomings. This has created an opportunity to both catalog barriers and propose and test innovative solutions to address them. We have synthesized our insights into 3 areas of focus (pediatric device CTUs, pediatric device trial initiation bottlenecks, and data considerations for pediatric device trials) and summarized our recommendations for other institutions who are interested in diving deeper into PMD clinical trials in Table 3. Legislative and policy actions will still be necessary to address many of the issues we highlighted in Table 2, but progress can be made by optimizing existing processes.¹⁵ 

It is worth noting that these recommendations are based on our own experiences and our conversations and interactions with sponsors, regulators, and researchers. There is very little peer-reviewed literature addressing this topic. As part of the same initiative that resulted in this manuscript, the PDCs are collaborating to aggregate what is known about the field of PMD research infrastructure through a scoping review and a Delphi process to generate more generalizable recommendations.

Although improving the pediatric device trial infrastructure is part of the solution, more needs to be done to strengthen the PMD pipeline and ecosystem. Financial models need to be developed that incentivize pediatric innovation.^2,20–22 The FDA may consider exploring pediatric-specific pathways and ensuring there is adequate pediatric device expertise across the agency, including at the Center for Devices and Radiological Health. Congress should explore expanding the funding and capacity of the FDA to specifically address pediatric devices.¹⁵ Our training and education of clinicians, researchers, and regulators must incorporate more medical device–specific information so that there is a workforce that can support this work.²³ And finally, better coordination across stakeholders at a national level will be particularly critical for high-risk devices that have high costs, small patient populations, and long development timelines. The System of Hospitals for Innovation in Pediatrics–Medical Devices (SHIP-MD)²⁴ that was developed by the FDA and many other stakeholders has been adopted by the Foundation for the National Institutes of Health for creation of a Pediatric Medical Device Public-Private Partnership (PMD PPP).²⁵ Coordinating national multistakeholder efforts via the SHIP-MD/PMD PPP will create a national framework to support evidence generation, evaluation, regulatory tasks, and fundraising for PMDs.

We should be clear that this manuscript represents the experiences of only a handful of institutions that have had frequent contact with and received funding from the FDA. It may be difficult to implement some of these recommendations, as they will require investment in specially trained staff and services at each institution, as well as redesigning several key elements of sponsored research vehicles. Finally, longitudinal studies will be needed to measure the impact of these recommendations on outcomes; device development occurs over long timelines (7–10 years by some estimates) and the outcome measures require careful tracking and surveillance. Focusing on process milestones (such as achieving design lock, preclinical studies, presubmission meetings, etc) may be a more practical approach in the short term.

Significant infrastructure changes are needed at children’s hospitals to account for the differences between pediatric drug clinical trials and pediatric device clinical trials. The PDCs possess a unique perspective to implement solutions to the barriers discussed. Our recommendations for other organizations interested in engaging in more PMD research are grouped into key functional areas. Further research is needed to assess the implementation strategies and evaluate any gaps. Cross-sector industry/small business-academia collaboration will be critical to addressing the gaps in PMD development and availability. National programs and frameworks like SHIP-MD and the PDC are critical to coordinating those collaborations effectively.

The Pediatric Device Consortia (PDC) Research Collaborative includes Robert J. Levy, Usha R. Thekkedath, Janene H. Fuerch, Kunj R. Sheth, James K. Wall, Michael R. Harrison, Elizabeth A. Gress, R. Brandon Hunter, Kolaleh Eskandanian, Julia Finkel, Francesca Joseph, Ioannis Koutroulis, Bobbe Thomas, Payal Shah, Joshua Dienstman, Vasum Peiris, Chester J. Koh, J. Lloyd Johnson, Karen S. Russell, Gumei Liu, Grzegorz Zapotoczny, Salima Jamal, Catherine Park, Nadine Afari, Yaniv Bar-Cohen, and Juan Espinoza. The authors are grateful for all the PDC team members, collaborators, and pediatric medical device innovators who make this work possible every day.