Understanding Value in Health Data Ecosystems

A Review of Current Evidence and Ways Forward

by Sonja Marjanovic, Ioana Ghiga, Miaoqing Yang, Anna Knack

This Article

RAND Health Quarterly, 2018; 7(2):3

Abstract

The potential of health data to improve the efficiency and effectiveness of health research and development, healthcare delivery, and health systems more widely is substantial. There are many initiatives across the EU that are experimenting with ways to capture value and address the nexus of technical, legal, ethics-related, governance and data protection-related, and cultural challenges to delivering potential benefits for society and the economy. The field of health data research and policy is highly dynamic and there is a need for further reflection, thematic learning and evaluation to better understand how to create and connect receptive places, to inform future interventions and to identify transferable lessons. Our research emphasises that realising the benefits of health data at scale will require: a simultaneous focus on the technological and structural conditions that are required; collaboration and coordination to transform working cultures and build health and care workforce and citizen capacity to engage with data; and efforts to ensure that policy, industry, and research communities respond to public concerns, needs, and expectations in a timely and sustained manner. The global community of individuals and organisations with a stake in health data will also need to consider how progress can benefit different populations across the world in an equitable manner.

For more information, see RAND RR-1972-EFPIA at https://www.rand.org/pubs/research_reports/RR1972.html

Full Text

A Need to Understand Changing Data Environments in the Health Sector

The evidence base that informs the life sciences and healthcare delivery is changing substantially. New types of data are informing the health sector, and advances in digital technologies are allowing for wider-scale and more diverse data gathering, processing and analysis. In its broadest sense, health data refers to any type of data that is useful for improved research, innovation and healthcare-related decision making. It can inform research and innovation, prevention and treatment strategies, health promotion efforts, self-care, health systems planning and wider public health activities, behaviours and decisions. As such, health data is relevant to a broad range of actors: academic researchers, industry, healthcare professionals and providers, patients and the public, payers and policymakers, regulators, charities and the third sector.

Health data stems from diverse sources and includes clinical trials data, clinician- or patient-reported data, and behavioural and health systems data. Some examples include: (i) electronic health records (EHR) data on patient symptoms, referrals, prescriptions and treatment outcomes; (ii) longer-term treatment outcomes data from real-world effectiveness studies; (iii) medicine performance data from randomised controlled trials (RCTs); (iv) genomic and proteomic data on individuals and associated biomarker data; (v) data from wearables and sensors (e.g. on vital signs); and (vi) data on individual preferences and health-seeking behaviours from social media. Although there is no consistent or unified definition of health data in the literature, it tends to be categorised according to who the providers are or what the data attributes or methods of collection are. In this context, literature devotes substantial attention to the potential of big data and real-world data to inform the health sector.1

In light of the changing evidence base underpinning the health sector, the European Federation of Pharmaceutical Industry Associations (EFPIA) commissioned RAND Europe to conduct a rapid evidence review of key insights on the value of health data. More specifically, the review aimed to: (i) identify and explain the potential and existing benefits that can stem from effective use of health data; and (ii) examine the key drivers of supportive health data ecosystems and their implications for future research, policy or practice, in light of the diverse challenges to be addressed. By health data ecosystems, we mean the technological and social arrangements underpinning the environments in which health data is generated, analysed, shared and used. The review was informed by an analysis of scholarly and grey literature, interviews with experts from diverse stakeholder groups and consultation at a European Union (EU)-level expert roundtable.

The Potential Benefits of Health Data Access, Sharing and Use

The literature on health data generation and use identifies a myriad of potential social and economic benefits from health data use, as overviewed below. In general, the evidence base focuses more on future than realised potential, reflecting the state of the field. Throughout the research we provide examples of initiatives seeking to capture value in three main areas: (i) research and development (R&D) and innovation; (ii) public health and pharmacovigilance; and (iii) healthcare delivery and the health system more widely. We also discuss how specific stakeholder groups could benefit from a data-rich health ecosystem.

Overview of Potential Benefits of Health Data

Potential Benefits for R&D and Innovation

Opportunities to explore new research areas could stem from access to richer data sources and new analysis techniques. For example: (i) linking datasets on genetic profiles with EHR on patient symptoms can help reveal patterns of association or disease causation that it was previously not possible to detect; (ii) access to real-world data from pragmatic trials and other real-world evidence can enable research not possible under an RCT model due to ethical issues (e.g. studies on narcotic abuse) or due to challenges of sample size (e.g. research on rare diseases).

Operational and cost-efficiencies could stem from: (i) better targeting of R&D investments and more appropriate clinical trial design due to improved patient stratification based on genetic traits and clinical records; (ii) reduced unnecessary duplication in research and enhanced confidence in results due to access to a richer and broader evidence base and enhanced data sharing.

Health data can also enhance the quality of research and innovation processes and outputs. For example: (i) real-world data from pragmatic trials can increase confidence in study results given that sample populations may be more representative of actual practice; (ii) using real-world data throughout the R&D and innovation cycle could also facilitate reimbursement for products that have a proven enhanced efficacy in a real-life setting and inform value-based and outcome-based payment approaches, as well as adaptive pathways; (iii) longitudinal data on treatment adherence and compliance creates prospects for new outcome measures in research.

Potential Benefits for Public Health and Pharmacovigilance

Prospects to scale up use of real-world health data in pharmacovigilance: Access to more diverse and greater amounts of real-world data than currently practised in pharmacovigilance (drug-safety monitoring), coupled with more granular information on patient profiles, could facilitate quicker and more rigorous learning about how drug safety relates to particular patient groups over time, including in the context of co-morbidities.

Prospects for enhanced, data-enabled public health promotion and prevention strategies: For example: (i) large and integrated environmental, genetic and socio-economic datasets could enable better prediction of risk factors for disease; (ii) data on health apps and portable devices could enable citizen empowerment and proactive behaviours in maintaining good health; (iii) computer algorithms and predictive analytics could assist in disease screening and early diagnosis.

Emergency-preparedness could be improved through more timely data matching disease outbreaks with covariates (such as environmental data from satellite sensors and data on symptoms from both health professionals and social media (although checks on reliability would be needed).

Potential Benefits for Healthcare Delivery and the Wider Health System

Benefits for healthcare quality: For example: (i) more personalised care and enhanced predictive analytics could be enabled by more comprehensive clinical datasets (e.g. improved screening algorithms and integration of imaging data, genomic and proteomic data on new biomarkers, and symptoms data from EHR); (ii) workforce access to more comprehensive evidence could facilitate better-informed care decisions (provided that evidence is presented in a user-friendly manner and trusted).

Operational and cost-efficiencies in healthcare delivery: For example: (i) easier comparability of outcomes data from different treatments across patient profiles could be enabled by large datasets from EHR, and could allow for more efficient decision-making, reducing wastage and costs associated with administration of inappropriate or inferior treatments; (ii) a reduction in unnecessary hospitalisations could be facilitated through data- and technology-enabled self-care and self-management of risk factors and through remote monitoring of adherence to treatments (this would require careful risk management).

Wider benefits for the health system: For example: (i) real-world outcomes data for treatments (e.g. clinical and patient experience data) could enable better-informed drug safety regulation, adaptive pathways and innovative reimbursement models; (ii) greater usage of EHR and costs data could facilitate more efficient health systems planning and resourcing, improved workflows and administration efficiency.

Creating Supportive Health Data Ecosystems

The landscape for health data generation, interpretation and use is still in a relatively early phase of evolution, with potential for much further growth. A supportive health data ecosystem, which would harness the full potential of health data and scale up existing benefits, is yet to be established. To achieve this, further research and policy consideration in a range of areas is required (as highlighted in Figure 1)—reflecting both the opportunities and the inevitable challenges that accompany transformational initiatives of the scale presented by a health data-rich society. We discuss each of these key areas below.

Figure 1. Building Blocks of Supportive Health Data Ecosystems

Figure 1. Building Blocks of Supportive Health Data Ecosystems

Collaboration and coordination in a health data ecosystem: Initiatives which seek to capture value from health data need to take account of the interdependencies and interactions between both stakeholders (e.g. research, industry, healthcare providers, patients, policy) and sectors (e.g. health and social care), given that health R&D and the delivery of healthcare are increasingly cross-sectoral activities. There is much more focus in the literature on potential benefits for a particular stakeholder group than on how value capture by any one group depends on its interactions with others. Similarly, there is little focus on how collective collaboration and coordination could work to address key structural, technical, legal and social boundaries to data sharing and access—including ownership and economic issues. In this context, the alignment of interests and the effective management of trade-offs is critical and individuals and organisations with boundary-spanning roles across different communities will be instrumental in addressing barriers and scaling up existing efforts.

Public acceptability and engagement: Literature highlights a need for learning from prior efforts to inform future engagement and communication campaigns with the public and to enhance public acceptability and involvement with health data sharing and use. The evidence emphasises that patients and the public expect to be informed. They need to understand the opportunities, risks and safeguards associated with health data use, individual and wider public-good benefits, potential impacts on doctor–patient relationships, and how inequalities will be prevented. The public expects to be informed about who has access to which data and why, how sensitive data and data portability will be governed and managed, under what circumstances they may opt out, and what legal provisions and citizen rights are in place should unintended uses or data hacking take place. Future research needs to consider what trade-offs patients and the public are willing to make when it comes to managing benefits and risks.

Making the most of recent data protection regulation and considering data sharing models: Developments such as the EU's General Data Protection Regulation (GDPR) are paving the way for a clearer governance framework for anonymised health data use in research and care delivery. The GDPR seeks to provide simplified consent requirements and requires rigorous technical and organisational safeguards for data sharing and use. Although the GDPR is a promising development in many ways, it is not without its challenges. For example, the boundaries of what is considered research use (and what constitutes commercial activity) and associated eligibility criteria for access to data are not yet clear. How the GDPR will be implemented at both EU and national levels remains to be seen. There is a need for further clarity on how different EU member states will interpret and act on the guidance, and what this implies for markets for health and analytics, going forward. An active campaign to ensure public understanding and acceptability of the GDPR governance framework will also be important. Further research will be needed to understand the regulatory capacity that has to be in place to ensure sufficient oversight of data access and sharing practices, public trust and commercial confidence in data validity. Related to data access and data sharing practices (but not the GDPR exclusively), various privacy and security safeguards are being piloted globally and span technological solutions (e.g. anonymisation, block chain, bloom filters) and more social and institutional responses (e.g. the GDPR, dynamic consent models, trusted curators).

Data quality and technical considerations: The literature on health data access and use also discusses a series of technical issues that initiatives in the health data field face. These include issues related to data quality assurance, a need for harmonised standards, issues related to the compatibility of IT infrastructure, a need for secure forms of data storage and transfer and enhanced data aggregation and sense-making capacities, and the governance of data use through technology and social interventions. There is already a diversity of EU efforts under way to try to address these aspects of health data ecosystems. Some examples include various cloud computing efforts for data storage solutions, new data mining techniques such as machine learning, opportunities for producing more complex databases with more complex data relationships, techniques for speedier data processing, and new probabilistic matching techniques to manage challenges associated with data linkage and anonymisation.

Workforce skills and capacities to engage with health data: Improving the ability of healthcare professionals to engage with health data will require capacity-building in both technical skills (e.g. clinical informatics) and in softer skills related to leadership and communication of new sources and types of evidence. Healthcare professionals will need to know how to access, interpret and act on diverse types and vast amounts of health data. Capacity-building among the workforce will also require the addressing of: concerns related to liability associated with data use and decision-making; time demands associated with engagement; uncertainty around how the doctor–patient relationship could be affected; and wider workforce implications. Capability-building needs to be considered in educational curriculums and continual professional development programmes alike. The research community, regulatory bodies and the general public also need to strengthen capacity to engage with health data in education and communication activities.

In Reflection

The potential of health data to improve the efficiency and effectiveness of health R&D, healthcare delivery and health systems more widely is substantial. There are many initiatives across the EU that are experimenting with ways to capture value and address the nexus of technical, legal, ethics-related, governance and data protection–related and cultural challenges to delivering potential benefits for society and the economy. The field of health data research and policy is highly dynamic and there is a need for further reflection, thematic learning and evaluation to better understand how to create and connect receptive places, to inform future interventions and to identify transferable lessons. Our research emphasises that realising the benefits of health data at scale will require: a simultaneous focus on the technological and structural conditions that are required; collaboration and coordination to transform working cultures and build health and care workforce and citizen capacity to engage with data; and efforts to ensure that policy, industry and research communities respond to public concerns, needs and expectations in a timely and sustained manner. The global community of individuals and organisations with a stake in health data will also need to consider how progress can benefit different populations across the world, in an equitable manner.

Note

  • 1 Big data is generally defined based on attributes of volume, velocity of data processing, variety, veracity and value. Real-world data refers to data collected (prospectively or retrospectively) outside the context of traditional RCTs and from diverse sources, for example data from EHR.

The research described in this article was conducted by RAND Europe.

RAND Health Quarterly is produced by the RAND Corporation. ISSN 2162-8254.