Keywords

Introduction

Given the broad intention of global health (GH) research and practice to place a priority on improving health and achieving equity in health for all people worldwide, RWD has multiple opportunities to facilitate GH initiatives. The benefit of real-world data (RWD) to inform various aspects of global health research is well supported (Chan et al. 2010; Murray et al. 1996, 2015) with great expectation for expanded utilization (Nord 2013; Salomon et al. 2003; Salomon 2008, 2013). A great deal of recent attention has been given to its benefit in drug development with the potential to improve efficiency and lower costs (Breckenridge et al. 2019). Beyond the drug development setting, the use of RWD has other customers beyond big PhRMA. Provider and payer organizations utilize RWD to increase their knowledge about the effectiveness, safety, and costs associated with a treatment option, while healthcare practitioners benefit from RWD to help inform the real-world implications of their treatment decisions. Financial RWD is obviously also critically important in the assessment of needs and allocation of resources as administered by the WHO. Patients themselves can exploit RWD to help inform discussions with their healthcare professional about treatment options. The positive perspective for RWD and early evidence of improved decision-making is largely realized by development strategies focused for the developed world, i.e., advanced economies with advanced technological infrastructure or high-income countries (HICs). The path for RWD utilization in the global health space (low-/middle-income countries, LMICs) is not straightforward, and additional challenges exist.

Though the use of RWD to bridge populations to gauge various GH solutions works well in some instances, this bridging exercise is often not appropriate in a more global context. Reasons for this can be due to a variety of factors including differences in the standards of care, heterogeneous populations, societal structure/network, migration, and adherence. Some of these issues could be addressed by increasing the availability and utilization of RWD in the different regions of the world; however, the assumption that such data already exists or is accessible is often invalid. The trajectory of RWD utilization in and for developing or low-income countries (LICs) has been very different than in developed countries. Healthcare for the affluent world and the data streams that define strategies and indices of health are both more extensive and varied than those that map health in low-income countries with often little or no data in these populations. Likewise, when we look at global health indices as assessed by RWD, these are often weighted to HICs, and so the analyst must take care to adjust appropriately if the intention is to truly reflect a global health perspective.

An added complexity is the diversity of the stakeholders involved in various aspects of global health, not the least of which are the governmental interests in the management and monitoring of the health of the populations and economies in scope. Geopolitical organizations such as NATO and the World Bank have a vested interest as well as global health advisors and policy makers such as the WHO, various centers for disease control and philanthropic organizations such as the Bill & Melinda Gates Foundation, the Wellcome Trust, Gavi, IAVI, PATH, The Rockefeller Foundation, and others. All have had and continue to invest in generating and collecting RWD sources for various purposes related to GH. In this chapter we will explore the current landscape and opportunities for RWD to inform various aspects of global health, assess the data available in these situations, and examine the context for which the current applications can be sustained and extended to gain the full value of RWD for the benefit of global health.

Context for RWD to Support Global Health

If one considers the most common forms of RWD to include electronic medical records, electronic health records, claims databases, health surveys, patient registries, data from health-related applications and mobile devices, and data from social media, it is easy to conclude at first glance that most of these data sources simply do not exist in LICs. In reality, of course, many of them do exist, but the access and/or infrastructure supporting their organization/management is not in place, and often the quality of data is suspect. With respect to actual RWD generated in global health target populations, this is accomplished with limited capacity to date given the difficulty gathering the data for the most part. The expectation is that such RWD will increase dramatically as LICs’ economies improve, and governments have greater focus on healthcare and healthcare costs. Although not always optimal, most countries have some form of civil registration and vital statistics systems to record births, deaths, and causes of death. As these improve, they may represent another source of RWD that could be utilized for global health research. When we refer to GH research, we speak very broadly about initiatives that place a priority on improving health and achieving equity in health for all people worldwide, both LMIC and HIC economies.

The fact that a disparity in both GH research and practice exists around the world is a surprise to no one. The dynamic nature of the disparity in the face of political and economic uncertainty makes the goal of “leveling the playing field” especially difficult however. Part of the challenge has been the generation of credible data documenting disease prevalence in LMICs where the global burden of disease (GBD) is highest. One needs only to recall the recent past to see the challenge and the effort it took to make progress. The GBD enterprise dates to the early 1990s, when the World Bank commissioned the original GBD study and featured it in the landmark World Development Report 1993: Investing in Health. Co-authored by Dr. Christopher Murray, who went on to become Director of the Institute for Health Metrics and Evaluation (IHME), the GBD study served as the most comprehensive effort up to that point to systematically measure the world’s health problems, generating estimates for 107 diseases and 483 sequelae. It covered eight regions and five age groups with estimates through 1990. The GBD 1990 impacted health policy and agenda-setting throughout the world; it brought global attention to otherwise hidden or neglected health challenges, such as mental illness and the burden of road injuries. GBD work was institutionalized at the World Health Organization (WHO), and the organization continued to update GBD findings. In 1998, the WHO created a Disease Burden Unit, which generated GBD estimates for 2000, 2001, and 2002, publishing the estimates in WHO’s annual World Health Reports. In 2008, WHO updated the GBD estimates for 2004. It took over two decades to expand the scope of the effort to its current form and establish a process for the accurate collection and distribution of this information. The WHO continues to recognize the strong interest in the data it collects in LMICs and is one of the few organizations in this area to provide access to such data in an open and transparent manner (see https://www.who.int/gho/publications/world_health_statistics/en/). However, this is often provided without much context, and the metadata that would facilitate the assembly of meaningful, analysis-ready data is simply not there. Likewise, much of the data are flat files (ASCII or Excel-based) and require significant effort to transform to be useful. Nonetheless, the collection and provision of vital sign data, patient comorbidities, and risk factors in LMICs is a significant contribution by the WHO and highlights an important facet of the GH ecosystem – the necessity of sharing in an informative and expedient manner is neither guaranteed nor always recognized. An important aspect of future progress will be the effort to organize/qualify data sources and provide governance over shared, accessible RWD data sources.

As globalization continues to evolve and health issues affecting countries and parts of the world become increasingly interconnected, so does the interest in global health and welfare. Not surprisingly many academic centers have also responded to the interest in and need for global health initiatives by creating innovative educational programs to address global health and welfare concerns. The interest is worldwide and diverse based on affiliation, focus, and research interests (Stone 2014; Sienkiewicz 2019). In addition, the Consortium of Universities for Global Health (CUGH) supports academic institutions and partners to improve the well-being of people and the planet through education, research, service, and advocacy (https://www.cugh.org/members). These programs offer multiple benefits, not the least of which is to give students the opportunity to study, research, and perform clinical rotations within domestic health as well as international health domains. Beyond the clinical/medical incentives, some programs offer specialization in the analysis and interpretation of RDW for global health issues (Withers et al. 2016) which train young scientists in the relevant approaches and methodologies but also give them a context for global health application. In addition, many academic medical centers not only provide teaching and research programs in various aspects of GH but also have arranged country-specific agreements with target LMICs where the GBD is high-yielding unique RWD source. In some cases, they are directly involved with WHO to propose labelling changes and make GH recommendations.

As global interest in RWE continues to grow, the databases and research methodologies used to collect and analyze these data have become more sophisticated as healthcare researchers are gaining access to new, previously unavailable data. With robust RWD sources, more insight-generating analyses can be conducted to help better inform healthcare decision-making based on everyday patient outcomes.

Definition Real-world data (RWD) refers to data related to patient health status and/or the delivery of healthcare that are routinely collected from a variety of sources, such as EHRs, claims and billing data, medical product and disease registries, patient-generated data including in home-use settings, and data gathered from other sources that can inform on health status (e.g., mobile devices).

Definition Real-world evidence (RWE) is the clinical evidence regarding the usage and potential benefits or risks of a medical product derived from analysis of RWD.

The global health ecosystem includes a variety of valuable stakeholders and partners including, but not limited to, the following: the WHO, various decision-makers (e.g., ministries of health, government agencies, other government departments at the national level), health partnerships, foundations, intragovernmental and nongovernmental organizations, civil society, media, professional associations, and various academic collaborating centers. The United Nations at the global, regional, and country level is also a major player. In addition, there are industry partners seeking to market their goods to GH economies, not the least of which are the pharmaceutical and related life science industries. Table 1 describes some of the specific stakeholder agencies and organizations that collect, promote, provide, or analyze RWD for various GH initiatives.

Table 1 Stakeholder agencies investing in RWD as part of their global health initiatives
Full size table

Within the context of drug and vaccine development, regulatory authorities have weighed in and provided both opinions regarding the value and use of RWD to augment a sponsor’s submission and the conditions upon which they will review. In December 2018, the FDA published a framework for Real-World Evidence program. In 2018, the EMA published a discussion paper on the use of patient disease registries for regulatory purposes (methodological and operational considerations). A major driver for the integration of RWE into the clinical decision-making process is understanding that RWD may provide valuable information on both short- and long-term medication safety and effectiveness in patient populations that are not well represented in randomized controlled trials (RCTs). RWD may also be used to provide insights into other aspects of medication use in clinical practice, including how a medication compares with other therapies in terms of surrogate outcomes, biomarkers, clinical uptake, safety, and cost. Through this application, RWD plays a critical role in bridging the gaps in information that cannot be met by RCTs alone. In resource-scarce healthcare environments, RWD may increasingly be used to generate more cost- and time-efficient post-marketing evidence than RCTs. At the moment the submission of RWE derived from RWD is viewed by regulatory agencies as supportive only as opposed to being a substitute for the two well-controlled Phase 3 trials currently required to gain market access. As confidence in the data quality and the approach improves, this may change as many are hoping. Independent of pharmaceutical sponsors, the payer community (health insurance providers, etc.) also look to RWD to verify effectiveness of approved agents and to justify formulary listing and reimbursement policies (Katkade et al. 2018; Amirthalingam et al. 2014).

One of the frontier areas for RWD application to GH initiatives is the extent to which it may influence future policy decisions. To some extent this has already been accomplished though it is based on limited scenarios where data has been highly curated, and the decision-making risk was not extreme. Certainly, the use of RWD to estimate disease prevalence rates and GBD in LMIC is a milestone for RWD impact and provides guidance on funding initiatives likely to improve these situations (GBD Collaborators 2015). At what point however are we comfortable with making policy recommendations for LMIC based on RWE? As with the desire to use such data for regulatory approval in HICs, the issue quite often revolves around uncertainty in the quality of the underlying source data and the methodologies and approaches used to compare RWE derived from RWD with data obtained in a more rigorously controlled setting. There is also the issue of education of both regulators and policy makers. While legitimate concerns exist regarding the quality of the underlying source data, the interpretation of such data with those collected under more stringent conditions (e.g., RCTs), and the evolving methodologies which have not been wholly validated, the necessity to look beyond the short-term limitations to provide an appropriately critical assessment must be embraced. Before we can engage in how such data can be used to inform policy decisions, greater comfort with the aforementioned issues must be addressed. Even though PhRMA looks on with interest and support, most of their efforts are based on late-stage, commercial, or surveillance efforts. The lobby for use of RWE as part of regulatory requirements begins with those that cannot afford the cost of Phase 3 trials and/or patient populations where financial incentives are limited as well as industry interest.

With respect to actual RWD data sources that might support various aspects of GH research, these would include patient registries, healthcare databases including electronic health records, pharmacy and health insurance databases, social media, and patient-powered research networks. Some of these are available publicly, and some are sold typically in a fee-for-service model but also as a subscription service. The vast majority of RWD available for purchase is, of course, predominantly based on HIC patients given the infrastructure, that is, type required to collect, store, and manage such data. Most vendors (that sell RWD) will claim they have LICs represented in their RWD sources, but typically upon investigation, such data comes from affluent regions of these countries and potentially represents foreign influence as opposed to reflecting the true population characteristics of the representative LIC. While even remote LIC hospitals are beginning to collect HER data, there is often reluctance to provide these data to those interested in GH, especially when the request comes from parties outside the host country. When permission is granted, it is usually with the stipulation that the data does not leave the data repository, and either must be analyzed on site or with significant security assurances regarding access only privilege (Srivastava 2016). Many LICs struggle with the adoption of HER in general, and the mixed healthcare system in India (government, state and private sector influence, and ownership) has created a diversity in readiness and strategy. Likewise, the common issues of standards adoption, limited current use for clinical decision-making, lack of legal safeguards (laws in place) to ensure the appropriate sharing and data privacy, and liability concerns create what seems like an insurmountable obstacle for expedited HER access in the near future though the interest level is high.

Beyond HER there are also disease registries that provide longitudinal data in certain patient populations of interest. In the global health space, the availability of such registries coincides with disease prevalence, but here again that does not mean that rare diseases don’t have outreach into LICs. The underlying problem, as always, is the lack of infrastructure to identify and care for these patients. Table 2 provides a landscape view of available RWD sources that serve the global health community. It should be viewed as representative and not exhaustive though it attempts to provide a diverse view of varied data sources and data types as well as the extent to which it represents LMICs which are often the subject of global health research and outreach.

Table 2 Common RWD sources (registries, HER, etc.) available for access and/or purchase and the extent to which they represent data from LMICs and usefulness for GH research
Full size table

Patient disease registries are organized systems that use observational methods to collect uniform data on a population defined by a particular disease, condition, or exposure, and that is followed over time. They are also a critical RWD source utilized in various aspects of research including GH initiatives. Many registries are focused on providing longitudinal data and clinical experience in specific disease populations around the world; many of these are rare diseases. The vast majority of registry data exists in HICs that have established infrastructure for identifying and connecting patients to resources for treatment, data collection, and support. Likewise, the organizing group responsible for managing and maintaining registry data is often specific foundations representing these patient populations in HICs and more likely to be organized and managed by individual government organizations in LMICs (see Table 2).

Most would conclude that registries remain an underutilized resource with some obvious shortcomings including heterogeneity in registry design and in the data collected, unreliable data quality, and data sharing impediments (McGettigan et al. 2019). In an effort to address some of these deficiencies, the Patient Registries Initiative was established by the European Medicines Agency in 2015 to support registries in collecting data suitable to contribute to regulatory assessments, especially post-authorization safety and effectiveness studies. Table 2 contains several registries (e.g., sickle cell, cancer, and stroke) in LMICs targeted to diseases specific to their populations but also many in which there are HIC equivalent sources.

Other Data Sources

In the USA, the Agency for Healthcare Research and Quality (AHRQ) offers practical, research-based tools and other resources (Kronick 2016) to help a variety of healthcare organizations, providers, and others make care safer in all healthcare settings (https://www.ahrq.gov/data/resources/index.html). Other similar organizations around the world including the National Institute for Health and Clinical Excellence (NICE) in the UK, the Institute for Quality and Efficiency in Health Care (IQWiG) in Germany, the Haute Autorité de Santé (HAS) in France, and the Canadian Agency for Drugs and Technologies in Health (CADTH) provide similar resources for the GH community. Additionally, Health Technology Assessment International (HTAi) is an organization with global membership that promotes evidence-based technology assessments.

The contract research organization (CRO) industry has also started to make significant investments in the acquisition of HER data with the intention of providing it to pharmaceutical sponsors mostly for the purpose of aiding patient and site selection but also for Phase 4/ post-marketing surveillance studies and market access facilitation.

Data are for the USA:

  • Percent of office-based physicians using any EMR/EHR system: 85.9%

  • Percent of office-based physicians with a certified EMR/EHR system: 79.7%

These efforts are at various stages, and in most cases the vendors are transparent in the recognition that the underlying data quality in many of the resident systems is extremely variable. Given this fact every consumer of this data must recognize that they use the information at risk. Likewise, pooling this data across sites is not straightforward given the variation in HER systems and underlying data structures and standards (or lack of). Table 3 provides an assessment of the current CRO landscape with respect to their HER data provision coverage across key LMICs where mature HER experience does not exist but is growing. Some of the individual hospitals, particularly in Africa and India, are listed in Table 2 as well.

Table 3 Contract research organizations (CROs) engaged in HER access and provision in key GH target geographic areas (e.g., LMICs)
Full size table

An important capability for HER providers is the ability to ensure patient privacy and maintain some level of data standards to confirm that the data is usable and interpretable. Clinerion and other vendors use proprietary technology to accomplish this and confirm that only de-identified patient data unlinked from identifiers is used, ensuring that identifiable patient data does not leave the hospital. Data security is enforced by multiple firewalls, one-way data connections, and user authentication. Compliance with data privacy regulations in the USA (HIPAA) and the European General Data Protection Regulation (GDPR) (EU 2016/679) is a must along with the express permission of participating hospitals and is done in accordance with relevant local/federal legislation.

An important resource for those getting trained to work in this field is the access to mentors and other resources that would benefit young data scientists. For this purpose, professional societies have recognized the potential of RWD to influence healthcare from a variety of vantage points and provided programming, networking, and educational services. A short list of professional societies along with their primary focus is provided in Table 4 below.

Table 4 Professional societies offering RWD programming and educational/networking opportunities
Full size table

Importance of Data Sharing

Global health research and policy recommendations face increasing demands for transparency at every stage, including an emphasis on improvements in data sharing (Taichman et al. 2016; Longo and Drazen 2016). Data sharing is not without challenge and controversy, particularly as it incurs additional costs, in both time and money, for researchers with limited budgets and timelines. Nevertheless, the likelihood that most or nearly all data pertinent to clinical trials and RWE analysis may be made public may reinforce the commitment of stakeholders to adhere to the highest standards of RWE analysis. More diverse RWD serving the GH ecosystem suffers similarly. Transparency and data sharing may also deter those tempted to release studies that are inadequate, misleading, or even falsified.

Important: The potential for RWD to inform GH initiatives will only be realized with strong and transparent sharing policies with expectation on data quality and recognized governance procedures.

Stakeholders need to commit themselves to achieving a standard in which researchers will post study designs, assumptions, and data sources prior to beginning the research process, at least in cases where analysis is likely to have major implications for public health. Some sources of RWD are likely to enjoy a higher level of trust than others. Data sets of proven quality will typically include well-established and long-running, or longitudinal, data sets that encompass data elements that are well accepted in clinical practice and clinical research, allowing for comparative analysis and for generalizability of results. Newer, less familiar, or novel data sets should be accessible for review and for comparison to similar data sets that meet the test of reliability and comparability.

Patients must not be left behind either. The last decade has witnessed increased health-related research in resource-poor settings mainly in Africa, Asia, and South America. While this increase is a positive outcome for addressing neglected diseases, it has also raised concerns over potential for exploitation through unfair distribution of risks and benefits among the parties involved. Several strategies to address these concerns have been promoted in the ethics literature including calls for universal standards of care, reasonable availability of proven interventions, and, more recently, promotion of the overall social value of research (Lairumbi et al. 2012). Closely related to the idea of promoting the social value of research is the determination of fair benefits through the consideration of what is owed to those participating in research and their communities, a process that has come to be referred to as benefit sharing. Although there now seems to be an international consensus on the need to share benefits arising from global health research, there has been continued debate over what constitutes a fair benefit, whether those that address the micro level issues of justice (those relating to individual circumstances of those participating in research) or those focusing on the macro level (the broader issues that might predispose participants to exploitation). The benefit of recent cross-sectional sharing on stimulating rapid innovation is indisputable so one would hope that the ethical considerations can be appropriately dealt with without stagnating the progress made thus far.

RWD Use Cases

The challenges of generating, analyzing, and applying RWD are particularly problematic in low- and middle-income countries. Given the size of the populations in question and the complexity of current healthcare delivery in these geographic areas (many diseases and thousands of medications and interventions), the reconciliation of data-driven improvements in clinical strategies with good population health is complex. The recent development of new methods to collect, analyze, and apply data has narrowed the gap between healthcare delivery and population health and created a vision of how these two settings can be bridged and with the potential for improved health outcomes. These new methods of collecting, curating, and conceptualizing data offer advantages to populations to see into their health trajectories with greater precision (Wyber et al. 2015).

Challenges notwithstanding, there are good examples which illustrate the diversity of RWD exploitation to the benefit of global health initiatives. Table 4 provides representative, recent examples which illustrate the growing confidence in the approach as well as the diversity of the applications. What is apparent in the few case studies shown is not only the diversity of the application of RWD but the variation in RWD sources and the necessity of merging and interpreting disparate data types, not to mention the differences in stakeholders requesting the analyses.

The health economics example (Campioni et al. 2019) provides an excellent example of the recent trend by pharmaceutical sponsors to appeal to healthcare decision-makers in targeted geographic areas of interest. The study is sponsored by the pharmaceutical industry but conducted and analyzed in collaboration with academic investigators treating the target population. RWD sources represent a combination of RCT data with observational, cost-effectiveness databases. While the focus of the analysis and models developed are based on the payers perspective, the cost-effectiveness and budget constraint recognition is clearly targeted to position the sponsor’s contribution to combination therapy in this case as superior to alternative choices. Still, as the world economy gravitates to a more value-based healthcare system overall, these studies and analyses conducted in the target patient populations (multiple myeloma patients in the Czech Republic in this case) using relevant real-world metrics (quality-adjusted life year (QALY) gained) are essential to generated informed healthcare policy decisions. One hopes that these types of studies are also directed to LICs in the future (Table 5).

Table 5 RWD use case examples with global health implications
Full size table

Similar in intention, the regulatory example (Wedam et al. 2019) illustrates the merger of RWD from electronic health records and insurance claims with safety data (from global safety databases that sponsors are obligated to collect as part of their surveillance programs), RCT data, and early Phase 1 safety data. This analysis provides an early example of what is likely to be an increasing trend in the pharmaceutical industry and allows regulators an opportunity to judge the value of the information in the approval process without lowering the regulatory requirements (thus far). Most importantly, from the GH perspective, this allows a baseline assessment from which one can judge the impact of data quality, reliability, and methodologies from well-defined and curated sources, all of which will be of a concern for data coming from LICs.

The social determinants example (Douthit and Seema 2018) is a somewhat unconventional choice for inclusion in the table to the extent that the source data is not the “big data” RWD type that we typically associate with these use cases. Rather, it refers to 70 British Medical Journal (BMJ) case reports from five continents, written by doctors, nurses, students, and allied health professionals. These cases, a burgeoning repository of evidence of how real patients are affected by disease, trauma, violence, sexual assault, conflict, migration, adverse living and working conditions, and poor access to healthcare, discuss, in addition to clinicopathological findings, need to be provided to the patient populations that would stand to benefit from the shared experience. The example highlights the necessity of patient and provider engagement in GH solutions and illustrates the current problems with poor dissemination of critical knowledge that would inform decision-making for those that often lack advocates. It also provides a roadmap of how culturally and medically appropriate care was supplemented by a strong universal healthcare system which included ethnic minorities, regular visits from culturally competent nurses and physicians, and strong social support.

Finally, the disease containment and information campaign strategy example (Lima et al. 2013) provides a means of establishing patterns in cellular data to examine the impact of habits linked to healthcare outcomes. Specifically, the example illustrates how human mobility is linked with the spread of disease. Four different data sets containing information about user mobility and call patterns at various levels of granularity and time duration were used to build a model which projects mobility and information spreading. The spreading model was then used in a simulation context to describe how the process progresses in the presence of contagion and mobility. Such examples also provide a mechanism of how such approaches can be used in diverse parts of the world where the mechanisms for the spread of disease are not the same and where mobility is an important factor in disease containment.

Conclusion

There are many stakeholders who stand to benefit from continued investment in RWD to advance GH solutions. The pharmaceutical industry faces challenges to find ways to make its innovative medicines available to patients. The interests of the global healthcare system including the interests of patients, prescribers, payers, and regulators, and the need to measure disease burden, create a complex environment for quantifying clinical value in diverse socioeconomic settings with varying support on the infrastructure side. Likewise, the healthcare industry will continue to be both a generator and exploiter of RWD. The delivery of healthcare is a complex endeavor at both individual and population levels. At the clinical level, the provision of care to individuals is guided, in part, by medical history, examination, vital signs, and evidence. More recently, these traditional tenets have been supplemented by a focus on learning, metrics, and quality improvement. The collection and analysis of data of good quality are critical to improvements in the effectiveness and efficiency of healthcare delivery.

There are also new frontiers for RWD. Specifically, there is a compelling case for the value of greater use of person-generated health data (PGHD) for regulatory and other policy decisions, and especially for integrating these data with other real-world data sources (Taylor et al. 2016; WHO 2015).

Definition Patient-generated health data (PGHD) are health-related data created, recorded, or gathered by or from patients (or family members or other caregivers) to help address a health concern.

Key areas considered where PGHD could add value to clinical research included defining patient subtypes, allowing more frequent measurements of outcomes, and creating more meaningful outcome metrics that measure patient preferences and improvement in their lived experiences. As we move toward a patient-centered health system in the age of personalized medicine, data from patients themselves will be needed to establish more robust patient profiles on how medications with the complexity of various comorbidities are influencing specific outcomes. Having these data available allows researchers and caregivers to ensure the right patient populations are being studied and treated appropriately for specific outcomes.

When PGHD are linked to other RWD sources such as curated, disease-specific registries, they can provide important insight into the patient’s clinical care journey and help accelerate biomedical breakthroughs. Linked with multiple RWD sources, PGHD could provide value when developing and validating patient-centered outcomes, particularly those outcomes that may not be easily identified through algorithms that rely on a single data source. These opportunities could be highlighted in a future set of “use cases” – potential applications to show their feasibility and impact on clinical evidence.

Data curation processes will be critical for improving the reliability and characterizing the validity of these data to ensure fitness for use to answer study questions. While there is a growing diversity of PGHD being collected by different systems, many questions remain about what data curation initiatives could improve the quality and assess the validity of PGHD. As the FDA and other regulatory agencies consider how to leverage RWD to inform key decisions, standardized approaches for evaluating these data and curation processes to ensure their fitness for the intended use are necessary. This will require consistent and scalable approaches that can be systematically applied to track the life cycle of PGHD.

Ultimately, policy makers and data users will need to answer the question, “Does this data meet expectations for regulatory use?.” This will require well-documented processes related to data generation and provenance, as well as understanding what data quality checks were used to assess the data set’s fitness. In addition, an appropriate governance framework must be developed and enforced to protect individuals and ensure that healthcare delivery is tailored to the characteristics and values of the target communities. The big data approach (Wyber et al. 2015) could be associated with health data that are owned by patients; robust governance processes that have been developed to ensure respect of values and principles in the use of data, with an emphasis on risk minimization; data that are aggregated automatically, with little effort and decreasing cost; interoperability standards that allow data to be seamlessly pooled and connected; laws that, while establishing adequate safeguards, allow the sharing and pooling of anonymized data in real time; and data that are presented in a usable format to patients, healthcare providers, entrepreneurs, and policy makers.

Clearly the path to ensuring the sustainability of RWD usage to address challenging GH problems is not straightforward. Nonetheless, there should be great optimism that this effort can be moved forward with confidence. It will require leadership from key GH stakeholders and a willingness from relevant governments to invest in infrastructure improvements while providing shared access to RWD sources. Data standards and broad transparency consideration are essential along with good communication. Some centralization of governance of such data with an “honest broker” model would also be helpful. The commercialization of RWD is well on the way to providing high-quality sources for commercial interests. While much of these efforts are currently dominated with data from HICs, one would hope that LMIC data will eventually represent a meaningful portion of these sources in the future. It should also be appreciated that these companies (CROs mostly) should move beyond the “use at your own risk” moniker and facilitate realistic solutions to data quality concerns. These are all solvable issues, and there is every expectation that future review of RWD analyses for GH solutions represents a significant advance to current efforts.