Heubi, James E.*; Guthery, Stephen†
A single-topic symposium was held before the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition (NASPGHAN) annual meeting in New Orleans on October 20, 2010. NASPGHAN leadership recognized that nutritional, hepatic, and gastrointestinal (GI) disorders are a substantial source of morbidity and mortality among infants, children, and adolescents. For many of these disorders, the pathophysiology, optimal prevention strategies, preferred diagnostic evaluation, and most effective treatments remain either unknown or poorly characterized. The National Institutes of Health (NIH) presently funds several multicenter collaborative networks in the field of GI disease. Results of discussions between the Executive Council and the Scientific Advisory/Research Committee of NASPGHAN elucidated that collaborative clinical networks are vital to improving the care of children with GI disease; there is no forum for clinicians and clinical investigators to discuss potential future collaborative opportunities; and there has been no systematic discussion about what biomarkers could be used as endpoints for clinical trials. The program, organized by the authors, was developed with the following goals:
1. To discuss collaborative networks, their implementation and function
2. To discuss potential funding sources for clinical trials
3. To discuss the science of determining robust endpoints and biomarkers of disease used for clinical trials
4. To discuss practical aspects of developing and completing clinical trials in pediatric gastroenterology, hepatology, and nutrition
The program was divided into 4 sessions with the following themes: how to create effective collaborative networks; clinical endpoints in pediatric liver and intestinal disease and nutrition; how to fund clinical trials; and breakout sessions designed to focus on areas relevant to developing and successful completion of clinical trials. The presentations are available at www.naspghan.org/wmspage.cfm?parm1=406.
Creating and sustaining collaborative networks was discussed by Edward Giannini, PhD, and Subra Kagathasan, MD. The creation of a collaborative research network begins with the identification of an unmet need of broad interest, and identification of a group of interested, appropriately qualified individuals who are willing to donate time and expertise. Early creation of bylaws, a Web site, and recognition of a home institution for the network's coordinating center (CC) are essential. The CC serves as the chief repository of certification credentials of network members, protocols, manuals of operations, regulatory documents, and a point of contact for potential sponsors and funding agencies. CCs frequently take on the responsibility of providing ongoing education of the network membership. An example is ensuring proficiency of its membership in assessing commonly used endpoints in clinical trials. Maintaining the infrastructure and staff of the CC may come from existing cores or center programs within the home institution, disease-specific foundations, fees collected for services provided, membership dues, and redistribution to the network of indirect costs paid to the institution. Maintaining infrastructure at a network clinical site, when not actively participating in an ongoing study, is extremely difficult. The size and sophistication of the network depend on the types and phases of studies it wishes to conduct, and the role it wishes to play (site management organization; academic research organization, clinical [or contract] research organization).
To date, there have been few investigator-initiated networks in our specialty (in contrast to industry-sponsored) or long-term natural history studies successfully conducted in childhood-onset inflammatory bowel disease (IBD). This is in sharp contrast to other pediatric subspecialties (Children Oncology Group and Cystic Fibrosis Foundation Care Network) in which successful collaborative networks that have been in place for decades executed landmark clinical trials that have changed the natural history of these disorders. Taking these networks as a “model,” the Crohn and Colitis Foundation of America (CCFA) Scientific Advisory Committee sponsored a collaborative meeting in 2006 and created a pediatric IBD network with a goal to be financially solvent, geographically diverse, and free of commercial bias. The primary aim of the network was to develop a large network to study clinical, translational, and natural history aspects of pediatric IBD. Out of this aim, the CCFA funded a multicenter research project and infrastructure to study risk stratification in children with new-onset Crohn disease (Risk Stratification Project). In 2008, the CCFA launched the Pediatric Resource Organization for Kids With Intestinal Inflammatory Disorders (PRO-KIIDS) as a comprehensive resource organization for children with IBD. The Risk Stratification Project was rolled into the PRO-KIIDS organization as part of the research network and expanded to develop other research projects in children with IBD. The PRO-KIIDS research network has now successfully grown into 46 centers throughout the United States and Canada. This group is successfully conducting a large inception cohort study involving children newly diagnosed as having Crohn disease while collecting their biospecimens (DNA, plasma, stool, and intestinal biopsies) to risk stratify their outcome and study translational aspects of disease progression. The success of the creation and sustainability of the network has been due to many factors including the following:
1. A “pilot phase” allowing the formation of a steering committee and obtaining institutional review board (IRB) approvals and material transfer agreements before the study began
2. Gradual expansion: Pilot phase included only 7 centers with a gradual increase in 4 phases to include a total of 46 centers
3. Sharing the workload: Multiple centers used their expertise to provide core services and research team management, including Emory University, providing administrative bio-banking cores, Cincinnati Children's Hospital Medical Center providing microarray and statistical core, The Hospital for Sick Children (SickKids) providing the data core, and Cedars Sinai providing the serology core
4. State-of-the-art remote data repository and management: Data management is performed by the study team with a remote data repository built in collaboration with Clinipace (a commercial database company)
5. Financial support from NIH, foundation or industry sources, with notable examples being CCFA supporting PRO-KIIDS
The continued success of the network is ensured by keeping constant communication channels open, transparency, and enthusiasm.
Protection of children as human research subjects, an issue of importance to all research in child health, was discussed by Susan Kornetsky, MPH. She presented the concept of the “sliding scale” for research consent in children, the concepts of permission from parent/guardians, and assent from a child or adolescent. Federal regulations define 4 (only 3 are listed) categories of research:
1. Minimal risk or greater than minimal risk with potential for direct benefit
2. Greater than minimal risk with no potential for direct benefit, which may yield generalizable knowledge about the subject's disorder or condition
3. Not otherwise approvable but provides an opportunity to understand, prevent, or alleviate a serious problem of children
Definition of minimal risk remains somewhat elusive. Does minimal risk include those “ordinarily encountered in daily life or the routine physical and psychologic examinations or tests” compared with a relative standard? Federal and Institute of Medicine advisory committees have endorsed the “absolute” interpretation of minimal risk to include risks encountered in normal, average healthy children living in safe conditions.
An excellent example of an area regarding minimal risk in GI research relates to the use of additional endoscopic biopsies for research. In this setting, investigators must provide rationale to allow IRBs to deem a study minimal risk or more than minimal risk. Approaches taken with IRBs can include data. A study by Yao et al (1), which included 253 research endoscopies, demonstrated the safety and tolerability of obtaining additional biopsies related to clinically indicated endoscopies. Luning et al (2) described a 16-year experience in the Netherlands that included 30,366 patients for whom the perforation rate was 0.12%. Alternatively, one could describe the largest number of biopsies that may be taken in a complicated endoscopic procedure and set that as a standard that cannot be exceeded for research purposes.
Assent, a requirement for older children and adolescents participating in research, is an affirmative agreement by the research participant. There is no specific age that is universally accepted for children. Maturity, psychological, emotional, and developmental states should be taken into consideration in obtaining assent.
Identifying resources at institutions for the support of clinical trials was discussed by James Heubi, MD. He discussed the resources available at the 60 Clinical Translational Science Award (CTSA) sites and consortium initiatives that may benefit clinical trial performance. The CTSA was created to provide an integrated environment to foster clinical and translational research (3). Every site has been charged with creating ways to help investigators using resources within their institutions and those provided by the CTSA grant. Most sites include a portal of access for supporting services such as biomedical informatics including REDCap for data management; biostatistics, epidemiology, ethics, and research design; clinical translational research centers (formerly the General Clinical Research Centers); community engagement with support for recruitment of study subjects such as ResearchMatch; regulatory knowledge and training; career development and education; and innovative technologies. There are existing collaborations between multiple CTSA institutions fostered by the CTSA Consortium. At a national level, efforts are being directed at examining the reasons for delays in IRB approval and contract negotiation. The results of this work should allow best practices to be identified, which can be spread across CTSA consortium and nonconsortium institutions. One of the efforts of the CTSA Consortium-Child Health Oversight Committee is to identify ways to create and implement new IRB models including the creation of a central IRB for the National Children's Study sponsored by the National Institute for Child Health and Human Development.
Pediatric research in private practice was discussed by Charles Thompson, MD. The drug discovery and development process is long, complicated, risky, expensive, and highly regulated. As advocates of pediatric health, it is important for practitioners and researchers to understand this process. He reviewed the discovery and development process, the evolution of Good Clinical, Laboratory, and Manufacturing Practices and the present regulations around pediatric research, both inside and outside the United States. He described the time required and hurdles to bring new discoveries to market and the phases of drug development (I–IV). Good Clinical Practices related to clinical trials of pharmaceuticals; Good Laboratory Practices that set the standards for how laboratory studies are performed, monitored, and reported; and Good Manufacturing Practices that set the standards for drug manufacturing, packaging, and storage play pivotal roles in the final drug-approval process. Recent legislation should encourage more pediatric studies. In 2002, the Best Pharmaceuticals for Children Act was passed. This legislation is voluntary and encourages industry to propose pediatric studies. The Food and Drug Administration (FDA) determines whether the drug is likely to benefit children and can issue a request for studies (pharmacokinetic/pharmacodynamic, safety, efficacy), and with completion of studies, they may award a 6-month patent extension. Similar legislation has been passed in the European Union—the Medical Products for Pediatric Use. The Pediatric Research Equity Act, passed in 2003, is mandatory for all products and requires pediatric data (safety, efficacy, dosing) included for all new chemical entities, indications, formulations, or changes in dosing or routes of administration.
Statistical and data management are essential to successful completion of trials. Stephen Guthery, MD, MSc, presented practical points regarding statistics relevant to clinical research. Often, statistical analyses are likened to lies. This opinion is often expressed in journal clubs and on rounds in academic arenas. In reality, however, it is up to editors, reviewers, and clinical practitioners to distinguish “lies” from statistics.
Desktop computing and the availability of large datasets have led to an explosion of quantitative knowledge. Lost in this explosion is that statistical principles remain the same. Guthery provides a relatively simple framework for physicians to conceptualize statistical hypothesis testing. In most biomedical research, statistical hypothesis testing is the assessment of differences between 2 groups. Statistical tests exclude chance as the explanation for an observed difference between 2 groups. Statistical hypothesis tests do not determine clinical relevance of an endpoint, do not assess clinical significance of an observed difference, do not measure bias, and do not assess generalizability.
The goal of the vast majority of clinical research is to assess the differences between 2 groups of patients. Before performing the test or assessing the appropriateness of the statistical test, the investigator, the editor, the referee, and the reader should be asking themselves 4 extremely simple questions:
1. What is the dependent variable?
2. What kind of variable is the dependent variable?
3. What is the independent variable?
4. What type of variable is the independent variable?
The science of defining robust endpoints was presented by Daniel Turner, MD. There are endless indices and scales used to measure outcomes in clinical research, but many were not developed in accordance with accepted standardized criteria and were not subject to robust evaluation. The question is how can we critically appraise a given measurement tool? To address this question, we should first consider the conceptual framework for which the instrument is needed: what is the concept we want to measure (eg, disease activity, quality of life, disability); in what population; what is the purpose (discrimination, evaluation over time, or prediction)? Psychometrics is a field of measurement in which concepts such as “depression” and “quality of life” are explained by means of describing behaviors using many interrelated items. In contrast, clinimetrics addresses more measurable concepts (eg, disease activity, prediction); it has only a few weighted items, each of which taps a unique aspect of the measured phenomenon. Whether the index is clinimetric or psychometric in nature, its development and evaluation must undergo a structured rigorous process (Fig. 1). First, a list of all of the potential items must be generated and then reduced to a feasible size. After formatting, the items are graded and, in clinimetrics, undergo mathematical and/or judgmental weighting for their perceived importance in explaining the measured concept.
Second, the instrument must be validated. Criterion validity may be used when a criterion standard is available now or in the future. Otherwise, construct validity should be used, using as many minitheories as possible that, when combined, explain the concept under study. Face and content validity and feasibility are evaluated judgmentally and play a crucial role in determining the credibility of an instrument. Interrater reliability must be demonstrated because an unreliable index cannot be valid. Next, for evaluative instruments (eg, those used in clinical trials over time), responsiveness to change and test-retest reliability must be proved. Finally, cutoff values that correspond to the different disease states and response should be determined to allow presenting the differences in rates of disease states (eg, response, remission) and not only the mean change between groups. It is important to emphasize that validity, reliability, and responsiveness are not a property of the instruments themselves but how the instrument is being used (eg, a disease activity index that was validated in adults is not necessarily valid in children).
William Rodriguez, MD, PhD, from the FDA's Office of New Drugs, presented how endpoints are viewed from the perspective of the FDA. Overall, challenges in pediatric clinical trials include the fact that clinical endpoints for pediatric patients may be different from those for adults, harder to obtain or quantitate, or difficult to reproduce. In general, FDA prefers that clinical trials demonstrate the effectiveness of a drug by showing its effect on a clinical endpoint that reflects how a patient feels, functions, or survives. Sometimes it is necessary to use surrogate endpoints instead. A surrogate endpoint can be a biomarker intended to substitute for a clinical endpoint. A surrogate endpoint is expected to predict clinical benefit (or harm or lack of benefit) based on epidemiological, therapeutic, pathophysiological, or other scientific evidence.
Recognizing the importance of endpoints in the process of the study of drugs and labeling, a working group in the Office of New Drugs, Center for Drug Evaluation and Research, and FDA provides consultative review to determine the adequacy of study endpoints (development, validation, study protocol issues, data analysis issues). The present approach at the FDA can be summarized by the following points:
* FDA is requiring more meaningful and valid endpoint development. Clinician-reported endpoints generated in questionnaire format are receiving additional scrutiny based on our patient-reported outcomes (PROs).
* FDA is developing resources and training programs to adequately review endpoints.
The clinical studies section of labeling should describe those endpoints that are essential to establishing the efficacy of the drug and those that provide additional useful and valid information about the activities of the drug. Data on outcomes of treatment should be presented only if the data are convincing and the outcomes are clinically significant. When considering endpoint selection, the applicant should define the target population, specific claims, concepts that support the claim, and instruments that measure the claim. Thus, in selecting endpoint models, the approach must be specific to and defined by population, disease, and treatment, and the specific endpoints chosen.
A key driver in protocol design and analysis is the recognition that there are challenges in defining the population. A symptom-based classification system and the measurement of symptomatology chosen may overlap. Another issue is that there may be evolution of diagnostic criteria that define the condition/subgroups of each condition. Evidence of drug effectiveness is deemed substantial for claims in product labeling if supported by adequate and well-controlled clinical trials, using endpoint assessments that are well defined and reliable to measure the specific concept(s) stated or implied by the claims. This evidence should be supported by content validity. There must be evidence that the score produced by an instrument represents the concept that the instrument was intended to measure. This observation should be supported by evidence from qualitative studies (eg, patient input) that the instrument, including instructions, recall period, and response option, is meaningful, understandable, interpretable, and appropriate. FDA then evaluates the content validity to determine whether the instrument measures the concept it was intended to measure and whether the instrument measures the concept claimed. Other measurement properties may be considered, for example, construct validity that demonstrates relations with other measures, reliability that demonstrates stability of scores over time, intercorrelation of items that contribute to a score, or agreement between assessors, and ability to detect changes and demonstrate how score changes over time in response to an intervention. FDA instrument review is specific to the review of the instrument. Specifically, FDA can only evaluate an instrument in the context of its intended use (ie, specific clinical trial, desired labeling claim). The most critical consideration is whether content validity has been established with input from the patients in the target population. Therefore, in the absence of evidence of content validity, other measurement properties cannot be interpreted. The following are examples in which alternate endpoints for proton pump inhibitors for pediatric populations were used:
* Nexium (esomeprazole)—the primary efficacy endpoint for the study was the time to study discontinuation due to symptom worsening during the double-blind treatment phase. (The symptoms evaluated in the Interactive Voice Response Systems assessment were based upon the validated Orenstein Infant Gastroesophageal Reflux Questionnaire [I-GERQ].)
* Prevacid (lansoprazole)—percentage of feedings in which crying/fussiness/irritability episodes occurred during or within 1 hour after feeding or ≥50% reduction in average duration of episodes was used
* Protonix (pantoprazole)—weekly gastroesophageal reflux disease (GERD) symptom score—sum of 5 frequency scores for vomiting/regurgitation, choking/gagging, refusal to eat, difficulty in swallowing, and abdominal/belly pain
Presently, no proton pump inhibitor is labeled for the treatment of patients younger than 1 year with GERD. Pharmacokinetic and pharmacodynamic data in this age group show that the medication does, in fact, increase gastric pH; however, this has failed to translate into a clinical reduction in GERD symptoms. Given this information, it may be possible to infer that clinical trials have not shown efficacy due to a lack of correct endpoint selection, patient population, and/or trial design.
Surrogate endpoints capture any relation between the treatment and the true endpoint. Thus, therapeutic-induced changes in a surrogate endpoint are expected to reflect important changes in a clinically meaningful endpoint. Surrogate endpoints must be known or highly suspected to be in the causal pathway of the related patient-centered outcome. By definition, surrogate outcomes are sensitive to treatment effects (and easier to measure); therefore, they are considered to be more responsive than usual patient-centered outcomes. For example, a biomarker may occur at some threshold level much sooner than a corresponding related clinical symptom. This marker may be more sensitive and could permit earlier intervention, presumably with a medication, that may affect the clinical sign by virtue of its effect on the biomarker. At issue is the definition of clinical benefit and how to measure it. There may be an important disconnect between the biological effect (does it work?) and an actual clinical benefit (does it help?) (4).
Recognizing the potential role of biomarkers in facilitating drug development, the FDA, in collaboration with the NIH, academia, and industry through the Critical–Path Institute (C-Path), is part of an activity known as Development of Biomarkers in Early Drug Development and Decision Making Biomarker Consortium, which provides support for the development of meaningful biomarkers in drug development. Biomarkers ideally bridge the gap between mechanism-based preclinical development and early pharmacologic clinical evaluation. It is important to discuss their use in early pharmacologic clinical evaluation and later clinical trials measuring clinical efficacy early with the FDA. Biomarkers are a key to accelerating drug development and guiding the use of new therapeutic options. Biomarker development approaches include the evaluation of the pharmacologic activity using animal models, bridging animal and human pharmacology via proof-of-mechanism or other observations. The process of candidate selection includes establishing pharmacologically active doses in humans and evaluating safety in animal models and in early clinical development. Finally, the research group members must keep in mind that the FDA must be made aware in a clear fashion of their use of biomarkers as “bridges” to avoid resultant loss of information and clarity.
The challenges of endpoints for specific conditions were discussed by James Markowitz, MD, relating to 6-mercaptopurine (6-MP) in Crohn disease, and Samuel Nurko, MD, relating to functional GI disorders. Initial planning for the 6-MP pediatric Crohn disease trial, one of the first prospective, multicenter, collaborative clinical trials completed in the field of pediatric IBD, began in 1989 (5). The study design reflected treatment paradigms of the time, because children with Crohn disease treated during the 1970s and 1980s commonly received repeated courses of corticosteroids to control disease symptoms. Because the concept of chronic maintenance therapy was not universally accepted at the time, many children were weaned off corticosteroids to remain off treatment until a subsequent flare of disease required additional therapy. Immunomodulator therapy with 6-MP or azathioprine was generally considered only after multiple courses of corticosteroids or after surgery and disease recurrence.
At the time the trial was conceived, it was increasingly clear that repeated courses of corticosteroids were associated with significant toxicity in children and that the alternative, treatment with sulfasalazine or olsalazine, was not particularly efficacious. Although thiopurines were being used as rescue therapies after a child's illness had become complicated, the idea of what would come to be termed “top-down” therapy had yet to be formulated. The 6-MP trial was designed to test the hypothesis that the introduction of 6-MP at the time of diagnosis would improve patient outcome and decrease the long-term need for corticosteroids. Because clinicians noted that the sickest children at diagnosis were the ones who were most likely to receive repeated courses of corticosteroids for disease control, it was decided that the most appropriate subjects to study would be children presenting with moderate to severe disease activity. A multicenter, collaborative, placebo-controlled study was planned, with subjects blindly randomized to treatment with prednisone plus 6-MP or prednisone plus placebo.
To design a successful trial, clear endpoints had to be defined. The endpoints had to be relevant to the disease process and easy to interpret. They needed to be simple to measure accurately and be capable of differentiating between the 2 treatment strategies being studied within a reasonable period of time. Unfortunately, Crohn disease is an illness that lacks clear objective measures that accurately reflect disease activity. Concepts such as remission and relapse are fuzzy and lack clearcut, universally accepted diagnostic criteria. Disease activity varies along a continuum and can wax and wane over time. At the time the 6-MP study was designed, disease activity indices had not been validated in children and were nearly as subjective as a physician's assessment. The Pediatric Crohn Disease Activity Index had not yet been developed.
Ultimately, the most objective endpoint that could be defined was determined to be cumulative corticosteroid dose. This endpoint mandated that all of the investigators follow the same criteria in adjusting corticosteroid dosing over time. Because there was no criterion standard approach to the use of corticosteroids for the treatment of children with Crohn disease, a corticosteroid dosing schedule had to be agreed upon that reflected, as much as possible, general clinical practice. Such a schedule was developed by a group of 10 pediatric gastroenterologists from around the United States. Meeting together, each member of the group first described their individual approach to the dosing and duration of corticosteroid treatment that they used in clinical practice. Consensus was achieved through ongoing discussion, until a compromise dosing schedule was developed that mirrored most clinicians’ practice but did not specifically represent the treatment prescribed by any individual practitioner.
Once the dosing schedule was agreed upon, a treatment algorithm was needed to ensure that all of the investigators made the same determinations about when a corticosteroid dose adjustment was appropriate. Because the study protocol required decisions regarding treatment to be made both during clinic visits and during frequent telephone follow-ups, a simple disease activity index, the Harvey-Bradshaw score, was chosen to standardize clinical assessments. Using this score and the amount of change compared with the previous assessment, a treatment algorithm was designed that dictated dosing adjustments for the clinician. Using the latest technology of the time to minimize investigator error, the algorithm was computerized on
Equation (Uncited)Image Tools
-inch floppy disks and made available to all of the investigators for use in the research clinic.
Clear definitions of disease remission, response, and relapse were based on Harvey-Bradshaw scores maintained for predetermined periods of time. The result of these initiatives was an objective endpoint that reflected standardized corticosteroid therapy across all of the collaborative centers. Cumulative corticosteroid dose could be characterized accurately, and each subject's clinical status over time determined without significant variability among participating centers.
Functional GI diseases (FGIDs) are commonly seen in the pediatric population. They have no structural or organic basis and there is no biomarker, so they are diagnosed by using symptom-based criteria (6,7). Outcomes after treatment must rely on symptom improvement and patients’ perceptions, outcomes that have been called PROs.
PROs assess how the patient feels or functions with respect to his or her health condition and reflect unobservable outcomes known only to the patient, and it is thought that patient reports may be more reliable because they are not subject to third-party interpretation (7). Furthermore, improvement in clinical measures may not correspond to improvements in patients’ function/feeling, so PROs are needed to fully assess the effect of the treatment. There have been recent concerns that PROs that have been used as endpoints in clinical trials may not truly capture the effect of treatment on FGIDs, so the FDA has provided new guidelines for the development of satisfactory PROs that could be used as endpoints in clinical trials. The challenges of developing those instruments in the adult population have been recognized and efforts to design new valid endpoint are under way. The development of satisfactory PROs in children with FGID also is needed and is more challenging, given that there are some important aspects of pediatrics that need to be taken into account when developing these endpoints (7,8).
Particular aspects for the development of pediatric PROs include the following:
1. The cognitive abilities of patients need to be taken into account (9) and the items must be designed to an appropriate developmental and cognitive stage, considering receptive and expressive language, and social development, to ensure the child understands the questions being asked (10,11).
2. PROs need to identify domains that are developmentally appropriate and that reflect activities that are pertinent and relevant to children who are being studied, so it may be necessary to develop an instrument that will be different according to the age of the children being examined.
3. The differences in the ability to recall information according to age also need to be taken into account.
4. Given that young children are not reliable in reporting symptoms, instruments that are completed by the parents/proxy may be needed. At times proxy ratings may be complementary, but in younger children they may be the only available data (6,8,12). It must be remembered that there is always the possibility that the proxy perception may not actually reflect what the individual child is experiencing.
5. Pediatric PROs should also measure the emotional and economic effect of the child's illness on the family unit.
Opportunities for support for collaborative research from NIH were presented. Patricia Robuck, PhD, from the National Institute of Diabetes and Digestive and Kidney Diseases, discussed funding mechanisms for collaborative networks. Collaborative networks, generally funded as cooperative agreements, are the primary means for National Institute of Diabetes and Digestive and Kidney Diseases–funded clinical research to be conducted at multiple (≥3) institutions. Unlike the grant mechanism in which the NIH has little involvement with the funded research, the cooperative agreement funding mechanism allows program staff at the NIH to work with the funded investigators as a collaborator and provide input into the design, conduct, and analysis of projects funded using this mechanism. A considerable number of networks have been established that support research in diseases relevant to pediatric gastroenterology, hepatology, and nutrition, including the Childhood Liver Disease Research and Education Network (ChiLDREN), Cholestatic Liver Disease in Childhood (CLliC), Biliary Atresia Research Consortium (BARC), Pediatric Study of Hepatitis C (PEDS-C), and Pediatric Acute Liver Failure (PALF). Collaborative networks may be established to meet a research need identified by the NIH and publicized through a funding opportunity announcement as a program announcement, request for applications (RFA), or a request for proposal describing the nature and scope of the desired research along with specific instructions regarding the application requirements. Consult the NIH Office of Extramural Research Web site (http://grants.nih.gov/grants/guide/index.html) for a list of present funding opportunity announcements.
Some collaborative networks, however, are established as the result of an investigator-initiated application not submitted in response to a specific RFA or request for proposal. Contact the appropriate NIH program staff for guidance about how to proceed; the processes are not standardized across the NIH and applicants can save a lot of time by talking with NIH program officials early in the application process.
Investigators should remember that many of the institutes and centers within the NIH fund research into areas of interest to NASPGHAN members. Subscribing to the Office of Extramural Research Listserv (http://grants.nih.gov/grants/oer.htm) is one of the best ways to keep up with present NIH trends, opportunities, and requirements.
Finally, aligning a well-designed research proposal with the goals identified in the Trans-NIH Action Plan for Liver Disease Research: A Report of the Liver Disease Subcommittee of the Digestive Diseases Interagency Coordinating Committee or the Opportunities and Challenges in Digestive Diseases Research: Recommendation of the National Commission on Digestive Diseases may be beneficial in obtaining funding (13,14). The Report on the Burden of Digestive Diseases in the United States is also an important resource when planning research objectives (15).
Orphan diseases and the pediatric investigator were discussed by Stephen Groft, Pharm D, NIH Office of Rare Diseases. A number of programs, including the intramural research and training and the extramural research programs, were discussed. A coordinated approach with a variety of agencies, advocacy groups, and academics is essential for developing orphan products. Development of orphan products, because they serve limited populations, presents unique challenges. Philanthropy commonly accompanies NIH funding, and collaboration with advocacy groups is essential. A number of challenges complicate the completion of studies: more sites are needed for recruitment than for nonorphan products; industry typically becomes interested only when relatively large populations can be enrolled; networks tend to work slowly; and multiple IRBs needed for approval of studies prolong the process.
Nader Youssef, MD, NPS Pharmaceuticals, presented the pharmaceutical company view of clinical trials in children. He presented the concept that there should be a partnership between pharmaceutical companies and investigators with the goal of patient safety and creating clean data. There should be objective feedback from scientific advisory boards regarding projects and an independent safety review performed using a data safety monitoring board. The benefits of partnerships between investigators and the pharmaceutical industry include identification of unmet needs in patient populations, novel therapeutic approaches, access to the latest data and thinking, development of real world protocols, synergism, and publications that can lead to peer recognition and financial rewards.
Andrew E. Mulberg, MD, Division Deputy Director, Gastroenterology Products, FDA, discussed how FDA, industry, and academia can work together. The opportunities for collaboration among academia, FDA, and industry are multiple and diversified. Examples include the development of standards that facilitate pediatric drug development, such as those that optimize central laboratory sampling and clinical trial design that are acceptable to worldwide regulatory agencies, including but not exclusive to EMA and FDA.
Coming together is a beginning; keeping together is progress; working together is success.
The critical role of central IRB propagation to expedite the process—the development of surrogate biomarkers and validation of patient reported outcomes, including questionnaires and psychometric tools for assessment of critical signs and symptoms—is relevant. The interaction of regulatory guidances, academic pediatric experts, and an evolving pediatric-focused workforce in industry understanding business case for safe and effective drugs for children portends successful pediatric drug development. Targeting the needs of children is an important task in all of the societies in which the effect of using proper therapies proven to be safe and effective has significant medical and economic consequences. FDA, academia, and the pharmaceutical industry must remain active and vigorously enthusiastic collaborative partners to yield safe and effective drugs for children. By cooperating more effectively, all will benefit, especially children, from our focus on streamlining and expediting our development of drugs for children with GI diseases.
During breakout sessions, topics were discussed relevant to the development of effective collaborative studies. Robert Squires, Jr, MD, discussed consortia organizations/practices. Pediatric gastroenterologists manage children with rare diseases. Clinical and pathophysiological features of even the more common intestinal and liver diseases differ from adults, which makes extrapolating management strategies for adults to children hazardous. Single-site studies require decades to log sufficient numbers of patients to achieve statistical power. Patient management that includes medical, surgical, nutritional, and transplant strategies vary depending upon the era in which they were enrolled. Thus, we are doubly challenged to enroll sufficient numbers of well-defined patients during a short period to provide evidence that can be generalized to a contemporary population. Only national, multicenter collaborations can provide that mechanism.
A number of pediatric consortia, supported by the NIH, philanthropy, and industry, are in place. Examples include the Children's Liver Disease Research and Education Network, the Pediatric Acute Liver Failure Study Group, the Pediatric IBD Collaboration Research Group, the Hepatitis B Network, and the Nonalcoholic Steatohepatitis Clinical Research Network. These consortia provide an opportunity for clinical investigators, both within and outside the networks, to develop hypotheses and propose ancillary studies that can be explored using database and biosample repositories.
Opportunities exist to establish new consortia related to pancreatic disease, intestinal failure, immune-mediated liver disease, functional bowel disorders, and others. A strong functioning network requires a number of components: a clinical condition that can be identified and characterized; a set of relevant hypotheses that, when tested, will provide new knowledge of the pathophysiology, clinical nuances, and management of the condition being studied; investigators who are willing to work hard, interact (collaborate) well, appreciate the challenges of clinical research and have local resources to support clinical investigation; opportunities to support translational and mechanistic studies; an experienced data CC; and a steady and reliable funding source. It is important to remember that, given the present budget challenges, what one wants to do and what one can do may not align. Determining and ranking the priorities for the consortium will be critical to its success.
The administrative structure of the consortium requires significant planning. Successful consortia have both seasoned and junior investigators with geographic, ethnic, and gender diversity. The size of the network will depend upon the disease frequency and variability, the types of clinical questions, whether a clinical trial is anticipated, and the budget. An administrative committee structure to support the unique needs of the consortium should be established coupled with a set of goals and objectives for each committee and term limits (if desired) for committee leadership and members. The committee structure should include not only principal investigators but also clinical coordinators, advocacy groups, and laypersons depending upon the nature of the consortium. Policies related to the proposal and approval of ancillary studies, publication and presentation of data, data quality, and use of biosamples should be established and approved by all of the members of the consortium.
Not all of the efforts will be successful and the challenges are many. The initial vision may not be the final product; however, new knowledge that will alter our understanding of a disease process or its treatment is difficult to come by. It requires hard work, collaboration, patience, tolerance, and a focus on the priorities of the project, with egos and self-promotion left at the door.
Endpoints in clinical trials were discussed by Anthony Otley, MD, and Kathleen Schwarz, MD. In 1948, the World Health Organization defined health as being not only the absence of disease and infirmity but also the presence of physical, mental, and social well-being. During the last several decades, a dramatic increase in the use of health-related quality of life (HRQOL) outcome measures has been evident in the adult and pediatric clinical trials literature. With HRQOL, one is attempting to ascertain the effect of the disease, concentrating on the health-related aspects of quality of life.
A self-administered questionnaire, which is easy to understand and complete and that covers all of the important aspects of the patient's HRQOL, is an ideal means of assessing HRQOL. There are 2 basic types of HRQOL measures: generic and disease specific. A generic measure is designed to measure all of the aspects of health and its related quality of life, and can include items and domains that are broadly applicable to various diseases and populations. Disease-specific questionnaires are more sensitive to disease-related changes in patients’ health status than generic questionnaires.
Any measurement tool should be tested before use to ensure that it fulfils the fundamental psychometric characteristics of a good measure. The psychometric characteristics to be assessed include sensibility, reliability, validity, and responsiveness to change (Table 1). Sensibility is a measurement characteristic with many aspects, and a questionnaire can include the assessment of feasibility for both the person administering and completing the questionnaire (ie, time to complete and mark, readability) and a critical review of the appropriateness of items included or omitted. Reliability looks at whether a measure has reproducibility. Validity is concerned with whether a questionnaire actually measures what it is intended to measure. Ideally, one would like to measure the validity of a HRQOL measure, comparing it with a criterion standard. The final characteristic, responsiveness to change, relates to the ability of the questionnaire to detect change over time, characteristics that are important for use in clinical settings (16–18).
The key endpoints that have been used in clinical trials in children with hepatitis B virus (HBV) have included suppression of viral replication, reduced liver inflammation, normalized alanine aminotransferase (ALT) level, and improved liver histology. With respect to suppression of viral replication, this endpoint has been evaluated in a number of ways, including reduction of viral load until serum HBV DNA is undetectable by polymerase chain reaction, and/or there is durable HBeAg seroconversion, and HBsAg seroconversion is achieved; however, the “devil is in the details” regarding the interpretation and reporting of these endpoints can be problematic due to a host of confounding factors, including HBV genotype, the mode of acquisition, and the phase of HBV (Table 2). Normalizing ALT is a goal often reported in adult studies, but in pediatric trials in which study participants may span from infancy to late teen years, the normal range for ALT will also vary for age, 40 IU/L up to 5 years and 26 IU/L between 12 and 18 years of age.
Although the endpoints in hepatitis C virus (HCV) trials share many similarities with those in HBV studies, given the differences between the viruses, there are likewise some key differences in the endpoints. Unlike with HBV, anti-HCV is not a desired endpoint because it is not a neutralizing antibody. Common overarching goals are to achieve suppression of viral replication and reduced liver inflammation. Determination of suppression of HCV replication is usually defined by a durable viral response 6 months off therapy or sustained viral response and was used for PEDS-C as an endpoint. Sustained viral response indicates an undetectable HCV RNA in serum or plasma 6 months postcompletion of therapy. Additional terms for viral response include “null response” for a <2 log drop in HCV RNA at week 12 of therapy, or “partial response” when there is a ≥2 log drop in HCV RNA at week 12, but the patients do not become HCV RNA negative at week 24. Relapse is said to have occurred when HCV RNA is negative at the end of therapy but positive at 6 months postcompletion of therapy.
William Rodriquez, MD, PhD, Office of New Drugs, FDA, and Linda Ulrich, Office of Orphan Products, FDA, discussed orphan products and the FDA. Orphan products are those used to treat diseases or conditions that affect <200,000 people in the United States. Drugs, biologics, medical devices, and even some medical foods used for rare disease or conditions may be orphan products lacking sponsors to develop and market them. Orphan products usually lack sponsors for financial reasons. The market for a product to treat a rare disease is likely to be small. With costly research and the prospect of limited profit, companies have few incentives to bring the product to market. The Office of Orphan Products Development (OOPD) at the FDA administers a variety of programs to encourage the development of products for rare diseases. These programs include the Orphan Drug Designation Program, which qualifies a product for special financial incentives; the Humanitarian Use Device Program, which motivates businesses to develop medical devices for rare diseases and conditions; and the Orphan Products Grant Program, which provides funding for clinical trials. Orphan drug designation is the process by which a sponsor for a drug or biologic product can take advantage of 3 special incentives: 7 years of marketing exclusivity to the first sponsor to obtain FDA marketing approval for the designated disease or condition; a tax credit of half of the qualified clinical research costs for a designated orphan product; and a waiver of prescription drug user fees. Also, the OOPD administers grants to defray the costs of clinical research needed to study orphan products as authorized by the Orphan Drug Act. Phase I studies may receive up to $200,000/year for up to 3 years; Phase II and III studies may receive up to $400,000/year for up to 4 years. The RFA is published in the Federal Register. The announcement is also located on the OOPD Web site (www.fda.gov/orphan), on www.grants.gov, and in the NIH Guide.
Parental considerations/support groups regarding children participating in clinical trials were discussed by Dennis Black, MD, and Cindy Hahn, Alagille Syndrome Alliance. The role and mutual advantages of involving parental support groups in pediatric clinical trials through the specific examples of the Alagille Syndrome Alliance, and the Studies of Primary Sclerosing Cholangitis and Withdrawal/Reinstitution of Ursodeoxycholic Acid in Pediatric Primary Sclerosing Cholangitis (studies aligned with the PSC Partners Seeking a Cure support group), were presented. They also applied lessons learned from these examples to other support group and research settings. Both the Alagille Syndrome Alliance and PSC Partners Seeking a Cure offer investigators helpful support and feedback on study design and recruitment, especially practical and safety aspects; a connection to knowledgeable, enthusiastic, and engaged patients and parents; assistance with participant recruitment and access to a broader participant pool; dispersal of information about upcoming, active, and completed studies; and potential funding support. In exchange, investigators offer support groups continued progress on clinical trials to validate presently used therapies and discover new ones; expertise through serving on the groups’ advisory committee; assistance in organizing and speaking at support group conferences; review of patient educational materials; assistance with research grant review; and strategic planning consultation. By being involved in clinical trials and collaborating with investigators, both the Alagille Syndrome Alliance and PSC Partners Seeking a Cure play a part in moving research forward and improving QOL for children with these diseases. Collaboration is a win-win endeavor that, when approached with mutual respect, trust, and support, has no downside.
Multicenter trials for private practionners were discussed by Charles Thompson, MD, and Stanley Cohen, MD. For practicing clinicians, specifically those in private practice, participating in research is driven primarily by a desire to contribute to the body of clinical research and participate in an intellectual process beyond daily clinical care because it is generally not driven by the need for promotion or career advancement. In addition, the personal costs for participating in clinical research (ie, time and dollars) are typically not borne by any supporting organization or institution (eg, medical school or hospital), significantly changing the costs of involvement in clinical and translational research. Instead, the costs of conducting clinical research are borne by the individual and/or the practice. Still, many clinicians actively pursue clinical and/or translational research to answer important management- or disease-related questions, to provide opportunities for their patients to benefit from national collaborative projects and pharmaceutical trials, to increase patient numbers in consortia projects that require large sample size (with private practices often providing an optimal environment for such trials), to participate in establishing the evidence base to improve patient care, and to enable the clinician to enhance his or her own growth and personal satisfaction. As always, the degree of participation in a specific research project is dependent on whether the clinician is the primary investigator initiating the project or a subinvestigator, and in that regard there is little difference between the private and academic clinician.
What does change is how the clinician sequesters time from his or her already encumbered schedule and the resources and infrastructure needed to underwrite and otherwise support those efforts in an office-based, private practice setting. Having well-trained and efficient research coordinators who can manage the numerous details implicit in every research project is essential. This assumes a real financial commitment from the investigator. New coordinator hires may require mentoring and may not be cost-effective initially. Special attention to the potential liabilities, specifically the indemnity clause in research contracts, should be thoroughly reviewed. Outside counsel is often necessary to review contracts to make sure that the language protects a clinician's practice and does not increase risk and liability.
There are a number of reasons why private practitioners participate in clinical or translational research. Some are different from those that drive research for the academically employed physician. Yet, the inclusion of the private practice clinician in research will augment our disease- and patient-centered knowledge and to ensure that a comprehensive view is represented in quality improvement studies and evidence-based medicine.