Article Text

Download PDFPDF

Letters
Ethical and regulatory considerations in the use of individual participant data for studies of disease prediction
  1. Bob Phillips,
  2. Neil Ranasinghe,
  3. Lesley A Stewart,
  4. on behalf of the PICNICC Collaboration
  1. Centre for Reviews and Dissemination, University of York, York, UK
  1. Correspondence to Dr Bob Phillips, Centre for Reviews and Dissemination, University of York, York YO10 5DD, UK; bob.phillips{at}york.ac.uk

Statistics from Altmetric.com

It has been suggested that individual participant data (IPD) meta-analyses of randomised controlled trials are the ‘gold standard’ approach to systematic review, allowing the most robust and unbiased assessment of research evidence to provide the most accurate information regarding the efficacy of a therapy.1 Such projects aim to collect raw line by line participant data from a range of relevant studies, often with updates of their outcome measures. The value of the IPD approach to review predictive studies, those assessing prognostic markers, clinical prediction rules or diagnostic tests, has also been stated.2

In forming an international collaborative to find clinical variables that predict the outcome of children and young people presenting with febrile neutropenia, we undertook an investigation into the ethical and regulatory considerations involved in sharing such information for our projects. This paper reports the outcome of our work to help inform the practice of others.

It has been suggested that the reuse of IPD from randomised trials within meta-analyses that address the same clinical questions should be exempt from further ethical review requirements. This is because the data are from studies which have already obtained individual consent.3 ,4 The use of data that had been obtained outside specific research studies, or where the meta-analysis has different aims, remains unclear. A consultation exercise undertaken in 2008 by the National Cancer Research Institute (NCRI) demonstrated the belief of most respondents that material and data collected from patients with cancer should be used, without identifiable information, as broadly as possible and that retrospectively seeking consent was inappropriate.5 The European Treaty on Biomedical Ethics permits the use of data without specific consent6 (15.2.i/ii) where there is minimal risk and potential benefit to similar persons.

Within the UK, legislation controls the use of patient data for the purposes of research, most recently the National Health Service (NHS) Act 2006. This has been interpreted by the UK Medical Research Council and summarised in a guidance document. These guidelines state that where possible, data should be released under specific consent. Where this is impractical, anonymised data should be used, and if this is impossible then an application to the Ethics and Confidentiality Committee of the National Information Governance Board for Health and Social Care is required to obtain access.7 Wherever data is used that has not had specific consent, consent should be sought from an appropriate research ethics committee (REC).8

The data sought for the ‘Predicting Infectious ComplicatioNs In Children with Cancer’ (PICNICC) Collaborative IPD review was anonymous (ie, the collaborative could not identify the patient from their data) and unlinked (ie, their data could not be mapped onto a subsequent dataset, with the potential for breaking anonymity). The project advisory group could not conceive of any harm that may have been occasioned by the use of such anonymous, unlinked data, and that there was a considerable benefit of an improved risk stratification system for episodes of febrile neutropenia for children and young people with cancer. This view was also supported by the parent representatives in the collaboration. Data were sought from formal randomised controlled trials and prospective observational studies, and also informal studies of data routinely collected in clinical practice or as part of quality improvement projects. The transfer of the information from the original researchers to the collaborative was requested by secure, encrypted electronic methods.

Within the UK, it was considered the project would require NHS REC approval for the use of patient data that had been recorded without specific research consent. Similar processes were discovered to apply in Australia9 and New Zealand.10 In other locations (such as Germany or the USA11) such data are exempt from the need for formal REC approval, but researchers are advised to have such protocols reviewed by ethics boards to assure quality and ease of publication.6

The PICNICC collaboration includes members from 12 countries. The collaborators, and those who expressed an interest but did not submit data, were surveyed to find their experiences of the process. These results are summarised in table 1. Within the UK, we received approval for the PICNICC IPD protocol from the University of York Health Services Research Ethics and Research Governance Committees, and from the NHS REC. In some European countries and USA, specific applications were made and consent obtained to share the information. Other groups were able to share their data from previous investigations without further formal approval. To our knowledge, no potential collaborative group had their request to share such data declined.

Table 1

Consent sought to collaborate in an IPD analysis of predictive features

Despite a perception from some clinicians that regulatory and ethical frameworks were blocking the work of clinical academics, it is important to report that large-scale international collaborations can effectively share data without obstruction or significant delay. Most service users wish to see care improved, and science advanced, and desire that the broadest possible use be made of their information. With a clear protocol, a sound ethical argument and appropriate requests to regulatory authorities, we believe that other studies should also be able to progress their objectives and may seek to use our experience and data to support their work.

References

View Abstract

Footnotes

  • Contributors BP co-conceived the project, designed the survey methods, collated and synthesised the results, and drafted the report. NR contributed to the development of the ideas and writing the study report. LAS co-conceived the project, reviewed the results and revised the report. The other members of the collaborative (see below) have shared their experiences, read and reviewed the paper and contributed to the discussion of ethical concepts underlying the report.

  • Collaborators The PICNICC collaboration is formed by those who have contributed data, or for patient/carer partners, significantly developed the project. The members are currently: authors, Roland A Ammann, Thomas Kuehne, Felix Niggli and David Nadal (Switzerland), Ian Hann (Ireland), Lillian Sung, Robert Klaassen and Sarah Alexander (Canada), Thomas Lehrnbecher and Arne Simon (Germany), Karin Meidema and Wim JE Tissing (Netherlands), Alex J Sutton, Richard Riley, Julia Chisholm and Rachel Dommett (GB), Elio Castagnola (Italy), Pamela Silva and Juan Tordecilla (Chile), Maria Spassova (Bulgaria), Hana Hakim and Glen Stryjewski (US), Gulsun Tezcan (Turkey), Lidija Kitanovski (Slovenia), Tiene Bauters and Geneviève Laureys (Belgium), Marianne Paesmann and J Peter Donnelly (EORTC).

  • Funding Grant number is MRC Research Training Fellowship G0800472.

  • Competing interests None.

  • Provenance and peer review Not commissioned; internally 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.