Summary of Literature on Pediatric Real-world Evidence and Effectiveness

Abstract

This cross-sectional study investigates the use of real-world data and real-world evidence in informing medicine efficacy in pediatric literature.

The US Food and Drug Administration (FDA) has long used clinical real-world data (RWD) and clinical real-world evidence (RWE) to supplement information from traditional randomized clinical trials on the safety of medications in children and adults, but RWD and RWE are not commonly used to demonstrate effectiveness, especially in pediatrics. With the widespread use of electronic health care data, there may be new opportunities for RWE to assess the effectiveness of medical products that are on the market. The 21st Century Cures Act of 2016 provides specific milestones for the FDA to achieve in evaluating potential uses of clinical RWE to support regulatory decision-making. As part of that effort, the FDA published its Framework for FDA’s Real-World Evidence Program in 2018 (https://www.fda.gov/media/120060/download). The initial question within the domain of clinical pediatric RWE is: what are the descriptive characteristics of existing clinical pediatric RWE effectiveness studies? This review of clinical pediatric RWE studies assessed studies from ClinicalTrials.gov and other published drug studies that focused on therapeutic effectiveness from 5 subspecialties (ie, psychiatry, cardiology, rheumatology, pulmonology, and oncology), which, based in part on analogous adult studies, could have promising clinical RWE application and could differ from one another in their approach to using clinical RWE to evaluate effectiveness.

Methods

This cross-sectional study used Embase and PubMed to search articles in the 5 subspecialties that were published between January 1, 2017, and December 31, 2019 (Figure). We screened all articles by title and abstract. Publications were selected if more than 50% of the participants were children and the study had at least 2 arms. Since the data used in this study were publicly available and the study was deemed not human subjects research, this article did not require institutional review board approval or informed consent.

Figure. Study Design.

Search terms included real-world data and RWD, real-world evidence and RWE, pragmatic clinical trials, longitudinal data, in utero exposure, treatment outcome, claims data, registries, mobile device, cell phone, electronic health records and EHR, Medicaid data, case studies, observational studies, observational data, and evidence-based medicine.

Results

Two reviewers agreed that 6 noninterventional publications met the criteria (Figure).^1,2,3,4,5,6 A review of completed pragmatic trials for pediatrics in the ClinicalTrials.gov database from 2017 to 2019 did not identify any clinical trials that reported primarily relying on clinical RWD in the assessment of drug effectiveness. In examining common practices to control for bias in observational studies, we found that half of the 6 publications controlled for confounding by statistical analysis (Table), and none of the publications used propensity score matching.

Table. Summary of 6 Studies Identified in Literature Review.

Source	Subspecialty	Population	Data source	Study design	Efficacy end points
Cabanillas Stanchi et al,1 2019	Oncology	Comparison of 41 patients taking osaprepitant and 42 patients taking granisetron for vomiting	EHR	Prospective cohort; no adjusting for confounding	Total number of vomiting events, number of chemotherapy courses with vomiting, proportion of patients with vomiting or who received dimenhydrinate, number of dimenhydrinate doses
Cruickshank et al,2 2019	Oncology	Comparison of patients using lidocaine and prilocaine (148 LPs) vs patients using lidocaine alone (142 LPs) for lumbar puncture analgesia	EHR	Retrospective cohort; no adjusting for confounding	Time to LP completion, time to recovery, number of maintenance doses of propofol required
Fouda et al,3 2018	Oncology	Cancer survivors (n = 68) vs healthy controls (n = 30) 25-OH-VD levels before and after supplementation	EHR	Retrospective cohort; no adjusting for confounding	Level of 25-OH-VD at 18 mo of supplementation with 50 000 IU 25-OH-VD/d or 1000 IU 25-OH-OH/d
Brown et al,4 2017	Psychiatry	PANS and PANDAS treated with corticosteroids (85 flares treated with 102 total courses of oral corticosteroids; 318 flares untreated)	EHR	Retrospective cohort; data were analyzed by using multilevel random-effects models	Length of flare in days
Leonard et al,5 2018	Pulmonary	120 Patients with SCD (65 ICS; 55 non-ICS)	EHR	Retrospective cohort; regression analyses were performed to control for differences in baseline status of known confounders	Rates of transfusion, oxygen requirement, BiPAP initiation, PICU transfer, intubation, readmission, hospital cost, and length of stay
Kimura et al,6 2017	Rheumatology	Untreated systemic juvenile idiopathic arthritis in CARRA; 8 patients started a nonbiologic CTP (2 GC, 6 MTX) and 22 patients on a biologic CTP (12 IL1i, 10 IL6i) at disease onset	Registry	Prospective cohort pilot study; demographic and disease features were similar between CTP groups	Untreated systemic JIA patients enrolled in CARRA began on 1 of 4 CTPs chosen by the treating physician and patient or family (glucocorticoid alone; methotrexate ± GC; IL1i ± GC; IL6i ± GC); the primary outcome of clinical inactive disease without current GC use was assessed at 9 mo

Abbreviations: BiPAP, bilevel positive airway pressure; CARRA, Childhood Arthritis and Research Alliance; CTP, consensus treatment plan; GC, glutocorticoid; ICS, inhaled corticosteroid; IL1i, interleukin 1 inhibitor; IL6i, interleukin 6 inhibitor; JIA, juvenile idiopathic arthritis; LP, lumbar puncture; MTX, methotrexate; 25-OH-VD, 25-hydroxy-vitamin D; PANDAS, pediatric autoimmune neuropsychiatric disorders associated with streptococcal infections; PANS, pediatric acute-onset neuropsychiatric syndrome; PICU, pediatric intensive care unit; SCD, sickle cell disease.

Discussion

Limitations of this study include the selection of a subset of only 5 pediatric subspecialties. Had we included other leading subspecialties, such as gastroenterology, we might have had different results. Additionally, this review was not a formal systematic review, meaning that we may not have captured the full scope of the literature on the effectiveness of RWD and RWE in pediatrics, even within the 5 subspecialties included. We did gather enough data to suggest that only a small proportion of pediatric RWE effectiveness studies may have been designed to optimally inform the effectiveness of drugs and biologics.

Our review of clinical RWE effectiveness literature found a paucity of studies with 2 or more arms in a set of pediatric subspecialties. Even fewer studies used methods that could potentially control for bias, limiting the ability to draw causal inference. This may be because they were designed as hypothesis-generating studies. At least 1 study relied on a registry that collected data elements that could potentially be used for more methodologically rigorous studies. The value of registries has been recognized by the FDA in its evaluation of drugs for cystic fibrosis. The establishment of systems in which data are routinely collected in a structured method that can be used to generate RWE may provide additional opportunities for pediatric RWE to inform pediatric effectiveness.