Characteristics and Results of Pediatric Medical Device Studies: 2017–2022

We performed a cross-sectional analysis of all interventional studies of medical devices registered on ClinicalTrials.gov that had a start date between January 1, 2017, and December 12, 2022, and included at least 1 US site. Studies were manually reviewed to select those that included FDA-regulated class II or III devices and that enrolled only children (aged 0–17 years) or both children and adults (aged 18 years and older). Class II devices are moderate-risk devices that represent about 43% of all devices and include products such as catheters, absorbable sutures, and surgical clips.⁷  Class III devices represent about 10% of devices and are defined as those that help sustain or support life, are implanted, or present substantial risk of illness or injury. These include devices such as pacemakers, prosthetics, and ventilators. Class II devices require regulatory clearance and class III devices approval before market entry. Laboratory-based diagnostic devices were excluded from the analysis. Although the FDA defines children as aged 0 to 21 years for the purpose of devices, we used clinical definitions and considered patients younger than age 18 years as pediatric.⁸  This study did not require institutional review board approval because it analyzed publicly available data and no patients or patient data were included in the study.

Data were extracted from ClinicalTrials.gov on study characteristics, including participant age eligibility, study centers (single or multicenter), randomization status, type of blinding (open label, single blind, double blind), study start year, and funding source (academic institution, industry, health care center, nonprofit organization, and government).⁹  Studies were classified as enrolling exclusively children or children and adults. Primary outcomes for studies were reviewed and classified as efficacy, safety, feasibility, or performance outcomes. Efficacy and safety outcomes were further categorized as constituting a clinical endpoint, surrogate measure, or clinical scale. Clinical endpoints were those that directly measure how a patient feels, functions, or survives; surrogate endpoints were markers or measurements serving as substitutes for clinical outcomes; and clinical scales were numeric measures and scores used to quantify symptoms related to the condition.^10,11  Studies with at least 1 clinical endpoint were classified as clinical.

Study status was determined from the ClinicalTrials.gov record. Studies with a status of terminated, suspended, or withdrawn were considered discontinued. For these studies, information provided by investigators in ClinicalTrials.gov on reasons for discontinuation was extracted. For studies listed as completed, we recorded the number of participants that had been enrolled.

For each device, we assigned a device type and attempted to identify the unique device studied. This was not possible if insufficient device identifiers were provided or if a study evaluated a generic device type. For each specific device that was identified, we used the Devices@FDA database¹²  to determine whether the device had been cleared or approved, and if the labeling included children (aged <18 years).

For each completed study, we determined whether results were reported in ClinicalTrials.gov or in a peer-reviewed publication as of December 27, 2022. Publications were identified through structured searches in PubMed, using the registry number, investigator names, disease studied, intervention, features of the study design, and study dates.⁹  A second investigator performed a search for all studies without a matching publication. The publication date was the date on which the publication first became available online. Journals in which studies were published were classified according to medical subject matter.

We conducted descriptive analyses using Microsoft Excel, version 16.62 (Microsoft Corporation). χ² tests and Mood’s median test were used to compare study characteristics and outcomes between studies enrolling only children and those enrolling both children and adults.

Our analysis included 482 studies that evaluated medical devices and were open to pediatric patients (Fig 1). Among these, 256 (53.1%) enrolled exclusively children and 226 (46.9%) were open to both children and adults (Table 1). There were 406 (84.2%) studies that examined class II devices and 76 (15.8%) that evaluated class III devices. For 355 studies, we were able to identify the specific device examined and determine that 58.1% (N = 151) were for an FDA-approved or cleared device, with 40.4% (N = 61) evaluating devices that included labeling for pediatric use.

The most common device types were diabetes-related devices (N = 57, 11.8%), monitors and measurement devices (N = 39, 8.1%), and digital therapeutics (N = 38, 7.9%). Most studies were single-center (N = 326, 67.6%) and used a nonrandomized (N = 255, 52.9%), open label (N = 350, 72.6%) design. There were 104 (21.6%) studies that enrolled neonates. Trials were primarily funded by academic institutions (N = 278, 57.7%) and industry (N = 142, 29.5%).

Studies enrolling only children were more likely to study class II than class III devices when compared with studies enrolling both children and adults (90.2% vs 77.5%, respectively; P < .001). Exclusively pediatric studies were also more likely to evaluate devices that had already been approved for pediatric use (54.4% vs 25.0%, respectively; P < .001). The distribution of funding sources differed between the 2 study groups (P < .001), with a greater proportion of studies enrolling only children funded by academic institutions (65.6% vs 48.7%, respectively) and fewer funded by industry (21.1% vs 38.9%, respectively). The age distribution of participants differed as well (P < .001), with exclusively pediatric studies more likely to be open to neonates and infants compared with studies enrolling both children and adults. In terms of study design, studies enrolling only children were more likely to be randomized (52.7% vs 40.7%, respectively; P = .008).

There were 291 (60.4%) studies that included a primary outcome of efficacy only, 35 (7.3%) that assessed safety only, and 40 (8.3%) that studied both efficacy and safety (Table 2). Among trials that included safety or efficacy outcomes, 260 (71.0%) used a clinical outcome. Most trials were ongoing (N = 307, 63.7%), whereas 124 (25.7%) were completed and 34 (7.1%) discontinued. For trials that were discontinued, the most frequent reasons for discontinuation were difficulty with patient accrual (N = 11, 32.4%), company decisions (N = 7, 20.6%), and problems conducting the study (N = 5, 14.7%). Among the 124 trials that were completed, more than one-half (N = 64, 51.6%) enrolled fewer than 50 participants and 71.0% (N = 88) enrolled fewer than 100. The median enrollment was 47 participants (interquartile range [IQR], 20–106).

Studies that enrolled only children were similar to those enrolling both children and adults with respect to outcomes studied and current status. The median number of participants enrolled in exclusively pediatric studies was 40 (IQR, 20–75) compared with 52 (IQR, 27–126) in studies enrolling both children and adults (P = .22).

After a median follow-up of 3.0 years, results for completed studies were available in peer-reviewed publications for 27 (21.8%) studies and were posted in ClinicalTrials.gov for 66 (53.2%) studies. Studies were published in a variety of subspecialty journals, including most frequently in pediatric journals (n = 4, 14.8%), diabetes-related journals (n = 4, 14.8%), and neuroscience-related journals (n = 4, 14.8%). The percentage of studies published in peer-reviewed journals and reporting results in ClinicalTrials.gov was similar between studies enrolling only children and those enrolling both children and adults (P = .19 and P = .64, respectively).

This analysis of pediatric medical device studies registered in ClinicalTrials.gov indicates that most devices evaluated in children are already FDA approved or cleared, though less than half of these include pediatric-specific labeling. Pediatric medical device studies are commonly funded by academic centers, use nonrandomized, open label designs, focus primarily on efficacy outcomes, and enroll small numbers of participants. The majority of studies have results posted in ClinicalTrials.gov, but only 22% have findings reported in a peer-reviewed publication after a median of 3.0 years since trial completion. Increased understanding of clinical research conducted to evaluate devices in children, including funding sources, study design, and availability of trial findings, will serve to further inform pediatric needs in device development and clinical evidence generation.

Pediatric device innovation and development lags substantially behind the advances that have been achieved for adult diseases and conditions, with industry focused primarily on the development of devices targeting adult populations.^13,14  In our analysis, less than one-third of device studies enrolling any children were sponsored by industry and only 21% of studies enrolling exclusively pediatric patients were funded by industry. The disparity between pediatric and adult investment is related to several challenges associated with developing devices specifically for children, which can be broadly classified into 4 categories. First, there are clinical considerations, including distinct pediatric physiology and unique disease and treatment requirements that preclude simple extension of adult products to pediatric populations.¹⁵  Second, there are technical challenges that must be addressed, including smaller patient size, children’s growth over time, and the need to demonstrate device longevity, which may require distinct products or device iterations.⁴  Third, pediatric-specific regulatory considerations include additional consent and assent requirements, stricter conditions for research conducted in vulnerable populations, and variable pediatric device expertise at both the FDA and local institutional review boards.^5,13  Last, financial barriers impact sponsor decisions to pursue pediatric-specific device development, including smaller market sizes, higher development costs, and lower reimbursement rates.^4,13  Although the first 3 categories can be addressed through scientific, engineering, and educational efforts, financial barriers are often cited as the most difficult to overcome. Addressing these will require significant changes to the mechanisms in place to regulate and pay for medical devices. Without adequate incentives and requirements (such as those that have been put in place for drug development¹⁶ ), the current market dynamics do not support sufficient incentives for medical device manufacturers to invest in developing products for children.

Another finding in our study was that among devices that were approved or cleared, most did not include pediatric-specific labeling. This may be because the device had not been previously assessed in children and clinical data on safety and evidence were lacking, or it may be related to current practices in incorporating pediatric data in device labels. Some devices are designed and tested in both children and adults but are then labeled only for adult use. Others have a generic indication that is silent on patient age and does not specify approval for pediatric use. Unlike for drugs and biologics,¹⁷  the FDA does not have age-labeling requirements or standards for devices, resulting in inconsistent documentation of approval for pediatric age groups. There is little public information on this aspect of the regulatory process, making it difficult to fully define the factors dictating whether and how pediatric approval is sought and how pediatric information is included in device labels. Further investigation into processes to increase the quantity and consistency of pediatric information in device labels, including data on safety and efficacy even when a device is not approved for pediatric use, could improve standards for pediatric labeling and support evidence-based decision-making.

When devices lack pediatric approval and labeling, they may be used in an off-label fashion by health care providers.¹⁸  Such off-label use is critically important when no adequate alternatives exist but has also perpetuated the disparities between pediatric and adult device development. Clinicians have developed approaches to effectively adapt or modify adult devices for off-label use in children, whereas certain pediatric specialties have also developed guidelines and reports on the safety and efficacy for these off-label uses.⁶  These practices have created a disincentive for device manufacturers, in which clinicians and researchers generate the clinical data and guidance that might have been required by the FDA from manufacturers had they sought approval for use in pediatric patients. Because the FDA does not have the authority to require manufacturers to perform pediatric studies (even if a product is relevant to pediatric conditions and there is a medical need), the system largely relies on investigator-initiated studies to fill in the gap and generate clinical data on off-label applications during the postmarket phase.

Our study highlighted several aspects in the conduct of pediatric device studies that warrant further attention. First, less than 20% included a safety outcome, focusing instead primarily on efficacy outcomes alone. Additional safety outcomes could be incorporated into clinical studies along with efficacy endpoints. Pediatric studies also tended to be small, with a median of less than 50 patients enrolled, and about two-thirds of studies consisted of single-center studies. Related to these features, the most frequent reason for trial discontinuation was poor patient accrual. Greater investment in and reliance on pediatric research networks and consortia could support improved enrollment and successful completion of pediatric studies.¹⁹  Finally, less than one-quarter of completed studies were published in peer-reviewed publications, which is somewhat lower than has been previously documented in studies examining nonpublication across a range of patient cohorts and conditions.^9,20–22  Although most trials in our analysis had results deposited in ClinicalTrials.gov, peer-reviewed publications serve a complementary function and must continue to be prioritized because they remain the most accessible information source for clinicians and the scientific community, including as the basis for clinical guideline development, meta-analyses, and development of future studies and research agendas.

Although we examined clinical studies registered in ClinicalTrials.gov, the registry does not capture observational studies and we were not able to comment on the use of these studies to study medical devices in children. Observational studies using real-world data (data generated in the course of routine care delivery outside of research or trials) have the potential to augment pediatric clinical evidence. For devices that are already approved, such evidence can be used to support labeling expansions to include children, conduct confirmatory studies, and perform safety monitoring.^13,23,24  However, the FDA is unlikely to accept real-world evidence for first approvals except in rare cases. Furthermore, to support scalable and robust research methods based on real-world data, health systems will need to improve how they record and track devices used in clinical practice.²⁵  The unique device identification system, launched by the FDA in 2012,²⁶  is a promising mechanism to track devices in clinical settings, though health systems have been slow to integrate this identifier as a structured data field within electronic health information systems, limiting use of real-world data for evaluation of medical devices.²⁷ 

There were several limitations to this study. First, we relied on publicly available data in Clinicaltrials.gov, which lacks structured data fields for medical device-specific concepts and may be incomplete for certain variables. In particular, there was insufficient information to uniquely identify all of the devices evaluated in studies, which precluded additional examination of the devices under study. In addition, because of the small number of published studies, we were unable to perform additional analyses examining factors associated with study publication and time to publication.

Our findings provide baseline information on current research activities for pediatric medical devices and serve to inform programs and initiatives seeking to increase pediatric-specific device development. In addition to considerations on ensuring the most rigorous trial design, greater focus is needed on timely and complete dissemination of trial results generated in pediatric device studies.

FDA

US Food and Drug Administration

IQR

interquartile range