Protecting data privacy

Ensuring that consent, confidentiality and patient privacy is protected is a prerequisite for building trust in big data and real world evidence. Sharing data across different legal systems is a challenge for big data research. DO->IT has produced an informed consent form (ICF) and guidelines to encourage data sharing with the protection of privacy.

What are the existing and emerging data privacy and informed consent practices in research?

  • There is substantial heterogeneity in preferences and practices surrounding informed consent.
  • There is a clear interest in developing new models of informed consent and adapting existing models, particularly driven by information and communications technology, data re-use, data linkage, and genome-based research.
  • Novel models of consent have the potential to add value in an emerging data environment and to facilitate research opportunities while preserving participant rights.

Which minimum data privacy standards should be included in Informed Consent Forms?

  • BD4BO partners have developed a proposal to demonstrate what EU-wide and even globally harmonised data protection clauses could look like under the new General Data Protection Regulation (GDPR). The ongoing heterogeneous interpretation of the GDPR across countries continues to make it difficult to develop universally accepted provisions.
  • The developed clauses have been discussed with patient focus groups and a panel of representatives of data protection authorities and ethics committees from various countries.
  • The Informed Consent Form template developed covers the use of study data collected in clinical trials for future research projects including biobanks.
  • An Explanatory document was developed to underline the rationale and give more background on the complex choices behind the proposed Informed Consent Form wording.

How can data privacy requirements best be explained to patients?

  • Patients have a diverse understanding of data privacy requirements. Finding the right ways to providing information and clarifications to them is essential.
  • DO->IT has developed training materials with key questions and answers on data privacy requirements and concerns.

Selecting relevant outcomes

The DO->IT project looked at how to identify, select and measure relevant outcomes in a transparent process involving a range of stakeholder perspectives.

Why is it important to select relevant outcomes?

  • “Outcomes are the results of treatment that patients care about most.” We can only optimise our healthcare systems and make them truly patient-centric by identifying, measuring and selecting the outcomes that matter to patients.
  • Organisations who make decisions on access to medicines have specific evidence needs. Therefore, it is essential to consider their perspectives as well.
  • In most disease areas there is a lack of consistency in the outcomes reported within routine care, registries or clinical trials. This makes it difficult to pool evidence to inform healthcare or policy decisions.

What process does DO->IT recommend for selecting relevant outcomes?

  • DO->IT has developed a practical guide for BD4BO projects, to identify, select and measure the core outcome sets for their disease area.
  • The developed clauses have been discussed with patient focus groups and a panel of representatives of data protection authorities and ethics committees from various countries.
  • The practical guide proposes six main steps to develop a core outcome set, from scoping to dissemination, with a focus on ensuring wide stakeholder input across all stages. It is also intended to stimulate standardisation of outcomes in Europe more widely.
  • The stages for development and implementation of core outcome sets are:
    1. Scoping
    2. Use of available core outcome sets
    3. Identification of outcomes
    4. Selection of outcomes
    5. Selection of outcome measurement instruments
    6. Implementation and uptake.

Using data effectively

Real world data can facilitate bridging the evidence gap between randomised controlled trial and real-world populations. This requires the correct tools for credible and acceptable evidence and conditions that enable data sharing.

How can we ensure big data and real world evidence is analysed robustly and used effectively in order to be acceptable to all stakeholders?

  • DO->IT has produced a review of econometric methods for real world data analysis and undertaken a case study review of core outcome sets and their availability in real world data.
  • The review assesses the strengths and limitations of several econometric methods in their application to real world data.

How can decision makers use outcome data more effectively to bridge the evidence gap?

  • Decision makers can provide early advice on study designs to ensure the right outcome data is captured.
  • Decision makers are looking at the use of methods such as propensity score matching to predict effects in real world populations.
  • Decision makers recognise that to improve the use of outcome data there a need for better coordination of initiatives and access to health data.

Can core outcome sets be applied to real world settings?

  • The use of core outcome sets is good practice and their application to real world evidence can help strengthen its credibility for decision makers and acceptability for stakeholders.
  • DO->IT has undertaken a case study review to identify if core outcome sets can be replicated in several disease areas.

Improving health outcomes

The BD4BO programme aims to improve health outcomes and healthcare systems in Europe by maximising the potential of big data. It does so by demonstrating big data’s potential benefits across a number of disease areas.

How can we maximise the potential of big data to improve health outcomes?

  • Help key stakeholders understand the value of existing and prospective data sets by improving the quality of the data collected and by robustly linking datasets.
  • Demonstrate pragmatic approaches to improving the availability, access and use of health data for secondary uses e.g. establishing a Big Data platform, a distributed model approach, creating common data models, and methodologies for analysing data.

What resources has DO->IT developed to communicate BD4BO programme outputs and maximise their potential to improve health outcomes?

  • The BD4BO Knowledge Hub provides an online platform for BD4BO projects to share their work with each other and the wider big data community.
  • A communications toolkit, including patient stories and testimonies, highlights the impact of big data initiatives on patients.

Overcoming big data challenges

There is huge potential for the use of big data and real world evidence in health research but they also create new challenges for industry, academia, regulators, health practitioners and patients.

What are the most pressing challenges in maximising the potential of Big Data to transform healthcare systems?

  • While there are many challenges to maximising the use of Big Data, it is important to frame the objective and the desired impact in order to identify the most pressing ones.
  • Alignment of big data initiatives, for example those under the BD4BO umbrella, can help identify and increase focus on the most important challenges.
  • DO->IT research highlighted five key challenges to big data usage:
    • Lack of alignment on the value of big data among decision makers
    • Lack of interoperability within and between healthcare systems
    • Lack of a data sharing culture between data custodians, data users and patients.
    • Significant legal barriers for researchers who want to access health data
    • Sustainability of big data initiatives.

What should future research in this area focus on?

  • A set of recommendations for future research priorities, intended to support the shift towards outcomes-focused, data-driven European health care systems, has been were developed in consultation with the DO->IT International Advisory Board. These recommendations are based on the findings from case studies exploring the potential value of big data for medicines, regulatory and health technology assessment decision-making.
  • The recommendations include:
    1. Ensure that existing and prospectively collected data is put to use
    2. Make patient-reported outcomes available in routine data sets
    3. Demonstrate the advantages and limitations of non-experimental methods for causal inference, and identify situations when non-experimental evidence can complement randomised controlled trials
    4. Use pragmatic trials to investigate effectiveness and safety using robust methodological standards in real-world populations.