Article Text

Download PDFPDF

StaR Child Health: developing evidence-based guidance for the design, conduct and reporting of paediatric trials
  1. William van't Hoff1,
  2. Martin Offringa2
  3. for the Star Child Health group (http://www.starchildhealth.org)
  1. 1Somers Clinical Research Facility, Great Ormond Street Hospital NHS Foundation Trust, London, UK
  2. 2Child Health Evaluative Sciences (CHES), Hospital for Sick Children, University of Toronto, Toronto, Ontario, Canada
  1. Correspondence to Dr William van't Hoff, Somers Clinical Research Facility, Great Ormond Street Hospital NHS Foundation Trust, Great Ormond Street, London WC1N 3JH, UK; William.vanthoff{at}gosh.nhs.uk

Abstract

There has been a huge upsurge in clinical research in children in the last decade, stimulated in England by dedicated research infrastructure and support through the National Institute for Health Research. This infrastructure offering research design, expert review, trial management, research nurse, data support and dedicated facilities enables paediatricians to conduct more and better research. The challenge is how to design and conduct trials that will make a real difference to children's health. Standards for Research (StaR) in Child Health was founded in 2009 to address the paucity and shortcomings of paediatric clinical trials. This global initiative involves methodologists, clinicians, patient advocacy groups and policy makers dedicated to developing practical, evidence-based standards for enhancing the reliability and relevance of paediatric clinical research. In this overview, we highlight the contribution of StaR to this agenda, describe the international context, and suggest how StaR's future plans could be integrated with new and existing support for research.

  • Evidence Based Medicine
  • Outcomes research
View Full Text

Statistics from Altmetric.com

Introduction

In every working day, children's health professionals encounter practices that are weak in evidence. Only fairly recently has this large gap in the evidence, that should underpin clinical choices for safe and effective treatments, come to the larger public's attention. It has motivated the emergence of dedicated children's research networks worldwide (e.g. National Institute for Health Research (NIHR) Medicines for Children Research Network (MCRN) in England). Until the last decade, there were relatively few paediatric clinical trials and a considerable number had shortcomings.1–8 The quantity, quality and relevance of research data involving children are substantially lower than for adults,3–5 despite data demonstrating that inadequate testing of interventions in children can result in ineffective or harmful treatments being offered, or beneficial treatments being withheld.6

However, in the last few years, there has been an emerging culture of conducting clinical research with children, stimulated in England by dedicated research infrastructure and support through the NIHR.9 This infrastructure offers research design, expert review, trial management, research nurse, data support and dedicated facilities to conduct the studies. There is now the opportunity for paediatricians and other children's health professionals to conduct more and better research, a key point in a recent Royal College of Paediatrics and Child Health report.10 The report emphasised the power of research in children to turn the tide of the growing burden of non-communicable, chronic, adult diseases that have their origins in early life to benefit the health of an ageing population and future generations, and reduce healthcare costs. It made several recommendations to improve early-years research, including the formation of multidisciplinary, cross-institutional groups of clinical and non-clinical child health researchers; changes in regulatory assessment of research that are proportionate and based on consistent national criteria; support for parents’ and young people’s advocacy and improved research training for paediatric trainees. Recent work has highlighted that many parents and carers welcome discussion of research opportunities but has also reinforced that many children's health professionals have concerns about approaching families.11 ,12 In the context of a proven effective clinical research infrastructure for children, the challenge is how to design and conduct those trials that will make a real difference to children's health.

Standards for Research (StaR) in Child Health (http://www.starchildhealth.org) was founded in 2009 to address the paucity and shortcomings of paediatric clinical trials.13 This global initiative involves methodologists, clinicians, patient advocacy groups and policy makers dedicated to developing practical, evidence-based standards for enhancing the reliability and relevance of paediatric clinical research. Through a systematic ‘knowledge to action’ plan, StaR Child Health makes efforts to improve and expand the evidence base for child health across the world. In this overview, aimed at all children's health professionals, we describe how this work has begun and how StaR's future plans could be integrated with new and existing support for research.

The StaR group and process

The aims of StaR Child Health are to improve the quality, ethics and relevance of clinical research in children by developing and disseminating evidence-based standards and guidance for the design, conduct and reporting of clinical trials with children.14 It is achieving this using a systematic process which includes identifying relevant problems, reviewing existing appropriate knowledge, generating guidance where gaps exist, adapting knowledge to the relevant context, facilitating the implementation of new knowledge and promoting best practice. The impact of StaR's work is reviewed through evaluation of the knowledge uptake and the effect on practice.7 In 2009, a systematic review of available guidelines for the design, conduct and reporting of research in children found few relevant guidelines.14 For instance, guidance on appropriate stratification according to age, relevance of development and choice of child-specific outcomes was not available. The review14 found that guidance emphasised what should be done but had less focus on how to achieve this. StaR aims to provide a global forum to develop and facilitate the use of better guidance on methodology in children's research. Following a systematic review and extensive discussion at a multidisciplinary inaugural conference in 2010 (including child health professionals, methodologists and regulators), priorities relating to the design, conduct and reporting of paediatric clinical trials were agreed. The first six topics were: consent and recruitment; containing the risk of bias; data monitoring committees (DMCs); determining adequate sample sizes; selection, measurement and reporting of outcomes; and age groups for paediatric trials.15 ,16

These topics have been systematically addressed through standard development groups (SDGs). Led by a nominated convenor, these are multidisciplinary groups that act as a forum for individuals interested in improving evidence in the specific topic. Membership is by invitation and voluntary subscription, aiming for representation from developed and developing nations. SDGs perform a thorough literature review, summarise best and weaker evidence in the topic and develop draft reports in standard format. These reports are then reviewed by a wider group of researchers, as well as regulators and representatives from pharmaceutical industry and other interested organisations. A programme of revision and updating of current standards should ensure that guidance remains up-to-date. StaR has developed web-based tools, seminars on methodology at large academic meetings, specific trial design courses, and conferences open to all stakeholders such as patients and parents, researchers, research funders, regulators and journal editors. To date, StaR has hosted four global summits (video materials can be found on the StaR website), writing weekends and workshops as a result of which six priority issues have been addressed and published in open access, as summarised below.

The first six StaR child health reports

  1. Consent and recruitment:17 Participation in research is a voluntary act manifest by the explicit consent of the participant. Usually straightforward for adults, the situation is much more complex in children where the consent is usually given by the parent or carer. Investigators should also ensure that they have considered the child's view, often demonstrated through the assent process. There are factors common to all trials or studies that influence the likelihood of participation (such as the underlying health disorder, research intervention, intensity of the study, quality of the study information and the existing beliefs of participants and investigator). A key factor for children's studies is the appropriateness and format of the information and there is often compromise between the regulatory requirements and the need for simple, brief and clear text, produced at the correct reading age. Demonstrating a specific strategy on information and recruitment is a prerequisite for a successful trial and even more so for children. The standard uses clinical scenarios to discuss and provide guidance on (1) who should give consent for children, (2) what information is necessary to obtain consent, (3) ensuring equitable opportunity rather than exploitation, (4) discusses the issue of payment and compensation, (5) the potential conflict of a child's clinician also acting as the investigator and (6) issues on eligibility of patients invited to participate in a trial.

  2. Containing risk of bias:18 A high risk of bias in a trial can lead to an overestimate of treatment effects13 and reviews of paediatric trials have shown that many are at a high risk of bias.14 This StaR paper addresses how to reduce and limit bias, such as ensuring adequate (1) techniques and security of randomisation, (2) blinding, (3) follow-up and handling of missing outcome data and (4) outcome reporting. Guidance is discussed in the context of both trial registration and existing guidelines for development and reporting of randomised controlled trials. The report suggests more work is needed to improve the implementation and uptake of best methodological principles, especially in child health research.19

  3. DMCs:20 DMCs should be considered for trials with vulnerable populations such as children. One review of 739 paediatric trials indicated that 2% had a DMC, 71% reported an adverse event (AE) and 20% a serious AE.21 An independent DMC can support the safety of study participants, but decisions made by DMCs (e.g. interim analysis and early stopping of trials) can affect the scientific validity, results and clinical impact of a trial.22–24 Two recent systematic reviews showed that only 17% of 648 recent paediatric trials reported on DMC activities, interim analysis or early stopping.25 ,26 This StaR article defines a set of minimum requirements to which DMCs should adhere to best serve paediatric researchers as well as trial participants. It explains criteria to determine whether a DMC is required for a particular study, both clinical (e.g. trials addressing major morbidity or mortality end points, a novel intervention with little existing safety data) and methodological (e.g. planned interim analyses, large and/or multicentre trials). When convened, DMC membership should be broad enough to include individuals with clinical and methodological expertise and knowledge of local context, and the operations of these committees should be guided by a detailed charter. The authors provide recommendations on reporting on the DMC and its operations in manuscripts as this will inform the reader in interpretation of trial results.

  4. Determining adequate sample sizes:27 Unsurprisingly, most paediatric trials are smaller than adult trials, are more often single-centre and rarely report sample size calculations.28 A key reason for smaller sample sizes in paediatric research is the relative scarcity of many medical conditions in children. While it is unethical to recruit more participants than is necessary (no unnecessary experimentation should be done), it is also vital that any research aims to recruit an adequate number to answer the designated objective. Failure to achieve the required sample size may prejudice the chance of answering the original research question and thus potentially waste the participants’ voluntary efforts besides the costs and resources in the study. Best practice in achieving the right number of subjects in a trial may be achieved by conducting more pilot studies to establish variance in outcome parameters, and applying interim analyses and other methods like triangular tests in the context of an adaptive sequential design to make paediatric trials ostensibly more efficient than adult trials. This StaR guidance covers (1) optimal sample size (including simulation and modelling techniques), (2) stopping criteria and (3) novel approaches to conducting research when the number of available participants is small (e.g., adaptive protocols, use of historical data).

  5. Selection, measurement and reporting of outcomes:29 Clinical research should provide answers that in turn enable better information, optimal care and well-being. To achieve this, the study needs to have the right outcome measure. However, compromise may be needed (a surrogate endpoint) for reasons of efficiency and practical considerations. A recent systematic review of children's studies concluded that few discuss the appropriate choice of outcomes.30 In trials across a broad age range, where some children are recruited but the majority are adults, outcome measures may be relevant only to the adults or may be based on endpoints or tools not validated in children. Sometimes, different methods and tools are used to measure the same outcome (e.g. pain, behaviour, quality of life). Patient reported outcomes can be useful but also challenging for children (who completes them, are they validated in that age group?). Unless the outcomes are clear, consistent and relevant, it is difficult, even in systematic reviews, to assess the totality of evidence based on different trials and thus to enable decision-making regarding healthcare interventions. This StaR report offers guidance on the choice, measurement and reporting of outcomes.30 ,31 A model for developing core sets of outcomes to be included in trials is presented.

  6. Age groups for paediatric trials:32 The response (and thus the degree of benefit versus harm) of adults and children to the same medication may be similar but may also be very different (for instance, steroids in bacterial meningitis, long-acting β2-agonists in asthma, phenobarbitone in cerebral malaria). Extrapolating the results from adult data may therefore lead to very inappropriate conclusions for children. Even within childhood, there may be wide variation in response between a preterm neonate and an adolescent. However, there is large variability in the age ranges and subgroups of children considered appropriate in paediatric trials and meta-analyses. This StaR report offers guidance on (1) age groupings, based on existing best practice, (2) justification for subgroups or combining age groupings and (3) the need to ensure biological, developmental, psychological and social variables are appropriate for trial design for the included ages. The paper also provides an agenda for further research on age groupings.

Topics undergoing current guidance development include informed consent strategies in acute care situations, trials that evaluate complex interventions, generation of evidence for safety and efficacy in rare diseases, and selection of justified appropriate comparators in trials.

Future steps and integration with (inter)national developments

StaR Child Health's aim is to create opportunities to raise the standard of research in sick children through the development and implementation of methodological and practical guidance. Identifying and providing better guidance will not alone change practice; there is an equal need to enable dissemination, uptake and review of the standards. In the first years, a very broad alliance of stakeholder groups including clinicians, researchers, methodologists, industry partners, regulators and journal editors all participated in different activities of StaR Child Health. New opportunities for involvement will present as StaR Child Health identifies future priority areas and new SDGs are developed. Following publication of the StaR papers addressing the aforementioned priority areas, a knowledge translation strategy has now been developed to support the SDG recommendations. For example, a pilot study has been conducted to test the usability of a wiki designed to educate paediatric trialists and trainees in the principles involved in minimising risk of bias in RCTs.33 StaR is working in concert with other relevant initiatives such as Global Research in Pediatrics, a European consortium set up to stimulate and facilitate the development and safe use of medicine in children, with a key focus on training in clinical pharmacology,34 and in close collaboration with national research developments such as the MCRN in England and the Canadian Maternal Infant Child & Youth Research Network, now united with other national trial research networks in EnprEMA.35 StaR Child Health aims to evolve into a global child health research network with an ongoing research agenda of empirical research for informing decisions regarding design and conduct of clinical research in children; a repository of guidance documents and supporting materials; and engagement in knowledge transfer activities and training to optimise uptake and implementation.

What does all this mean to the clinician?

It is no longer ‘too difficult’ to conduct clinical research with children. Research support for children's studies is available in many countries. In England, the NIHR has contributed over £100 million to infrastructure for children's research (for instance, the MCRN (now CRN): Children, Biomedical Research Centre, and Clinical Research Facilities). In 2012–2013, children's research studies number third in a rank of all NIHR medical specialties, testifying to the upsurge of paediatric studies (WvH, personal communication). The results are impacting care and treatment decisions. We now have got the support, processes, infrastructure, better methodologies and tools for best research. When, in practice, one encounters clinical problems that should be researched, it is now time to use these resources, engage with the local or national child health research network (e.g. NIHR in England) and take part in improving children's and families’ health.

Acknowledgments

For parts of StaR's work, funding from Global Research in Pediatrics (http://www.grip-network.org/index.php/cms/en/home) was received.

References

View Abstract

Footnotes

  • Contributors WvH was invited to contribute this article and invited MO to contribute. Both authors contributed equally to the literature review and manuscript preparation. WvH acts as the guarantor for the article.

  • Competing interests WvH is a Joint (interim) Director for the NIHR Clinical Research Network: Children.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.