How Health Information and Digital Technology Are Reshaping Patient Engagement
Patients utilize vast amounts of information and technology tools to manage their health. When ill, people often turn to Google to learn about conditions associated with their symptoms. Newly diagnosed patients use online peer-to-peer communities to seek support and information. Patients log on to patient portals to access their lab results, review sections of their electronic health records, and communicate with providers through secure messaging. Wearable devices can be used to track activity and sleep and to monitor vital statistics (e.g., blood glucose) in real time, generating vast amounts of data that can be used to encourage healthful choices and manage certain chronic conditions. In recent years, communities of patients have shaped research agendas by donating their health data and prioritizing research questions that matter to their quality of life.
Advances in digital technologies have created new and more active roles for patients. Many patients have shifted from being mere recipients of information (a passive role) to being producers of knowledge and data that can be shared with their peers as well as with clinicians and researchers (an active role). The rapid expansion of health information and technology tools, from patient portals to self-monitoring systems and social networking websites, has created the potential to improve patient engagement and to facilitate collaboration between patients and providers. Health information technology may improve patients’ ability to follow care plans and self-monitor their health status, enjoy more productive discussions with their health care teams, and adopt preventive behaviors, such as regular health checkups and healthy choices.
The abundance of health information, powered by innovations in digital technology, should motivate many patients to become more engaged in their health, yet health information initiatives also face an array of obstacles. First, offering tools that facilitate the access and sharing of health information does not always translate into patient adoption, suggesting that the adage “if you build it, they will come” rarely applies. Second, poor-quality information and designs that are not user-centered may limit patients’ ability to act upon the information they receive. Third, health apps that are too complicated and inattentive to users’ needs and capabilities may leave vulnerable patients behind, perpetuating inequalities in health. Research on wearable devices has shown that less healthy and poorer individuals are the least likely to use such tools. Finally, low health literacy also may limit the impact of data-sharing platforms. Even as more health information becomes available, many patients may not be able to use it to guide their health care decisions.
To understand how these obstacles might be overcome, we examined some of the more innovative initiatives in health information engagement. Clarifying how these initiatives mobilize patients to play a more active role in their health, which features are most helpful for patients, and which features remain problematic may deepen the understanding of these tools and inform the design of future platforms.
Learning from Eight Leading Initiatives: Methodology and Objectives
We studied eight initiatives addressing varied health conditions and representing the leading frontier of patient-engagement health information platforms. The platforms that we chose met three selection criteria:
- They allow patients to access health information and data;
- They allow patients to produce and share their own health data; and
- They allow patients to connect with other patients, with their providers, or with research groups.
Initiatives that share these three characteristics constitute some of the most advanced examples of how health information and technology tools can be used to promote patient engagement. We used a multiple case study method to compare platforms and to infer differences about their designs and objectives. This type of qualitative research method is appropriate to investigate complex phenomena and develop analytic frameworks using inductive reasoning.
Our selection of case studies was driven by telephone interviews with 32 key informants that were carried out between May and July 2017. We used purposeful sampling to identify information-rich individuals who could help to orient our selection. Our informants included designers of health platforms, disease advocates, experts from health foundations, physicians, academics, and technology entrepreneurs. We also conducted independent reviews of platforms that met our research criteria. We submitted our findings to two experts on digital health, and one of them recommended adding a Facebook patient community. This opportunistic sampling yielded eight cases that became the focus of our study.
For each case, we conducted expert interviews with patients, physicians, researchers, technologists, patient advocacy groups, and other informants. We set the number of interviews at 5 to 12 per case and selected participants through snowball sampling. Between August 2017 and February 2018, we conducted a total of 68 semi-structured interviews by phone or via video conferencing services. We used different questionnaires depending on whether we were interviewing patients and caregivers or physicians and platform leaders. Interviews were recorded, transcribed, and analyzed to identify common themes and to construct a narrative of each case. For one of the cases, we participated in a conference attended by a network of patients, providers, and researchers and collected data using participant-observation techniques and in-person interviews. The Harvard University Committee on the Use of Human Subjects determined that this research met exemption criteria and did not require an institutional review board. Table I offers a synopsis of our case studies.
Even though these initiatives share similar traits and technology tools, they aim at quite different objectives. We identified the principal purpose of each health information initiative and clustered cases to understand how the central goals of these platforms inform their designs and methods for engaging patients.
Focusing on Objectives and Understanding Patients’ Motivation to Engage
Generally, platforms feature several different ways to engage patients. Most initiatives allow patients to access their health data. For example, in the Swedish Rheumatology Quality Register, patients can log into a patient portal to review their historic data and learn more about their response to their treatment plan. Tidepool provides access to blood glucose data, and the Health eHeart study provides access to blood pressure and cholesterol data. Several initiatives also have online forums, or private messaging tools, that let patients connect with their peers, with their clinicians, or with research communities.
However, our findings indicate that, when it comes to health information initiatives, less is more. Initiatives should focus on the intersection between patient motivation to engage and platform objectives and then develop functionality that advances both. This approach is preferable to offering an array of functionalities with the hope of attracting more users.
Coproduction of Health Quality Improvement
Health initiatives that focus on engaging patients in long-term quality improvement (e.g., ImproveCareNow and the Swedish Rheumatology Quality Register) mobilize patients not so much through information dissemination or access to personal data, but by building strong and lasting connections that join patients, caregivers, and clinicians to work together to improve care. One patient who attended ImproveCareNow’s biannual conference, where patients, providers, and researchers exchange knowledge and ideas, explained:
There is empowerment in feeling connected. Health care often follows a paternalistic model, but having access to clinicians shortens the distance between patients and providers, putting them on more of an equal footing. Providers feel more touchable. They are people too, and they want to learn from us as well.
Cultivating such connections through periodic conferences, meetings, and webinars may help to sustain patient engagement. Similarly, patients in the Swedish Rheumatology Quality Register use an online portal to report their health outcomes (e.g., ability to perform daily activities, level of pain, etc.) prior to their clinical encounters. Their data “feed forward” to their clinicians and, together with data collected by their rheumatologists during clinical encounters, populate the rheumatology registry. These patients share their health data so that they can engage in more meaningful discussions with their physicians. They use the platform because they are asked to do so by their physicians and view it as part of their care routine rather than as a tool to use independently.
In this case, making it easier for patients to record their data and discuss symptoms and progress with their clinicians may stimulate engagement much more than simply adding a feature that allows patients to discuss their symptoms with other patients. Connecting with their peers is not their motivation to engage, and there may already be better tools that serve that purpose (e.g., patient social networking sites).
Peer-to-Peer Patient Connections
Peer-to-peer platforms (e.g., MyHealth Teams and Breast Cancer Straight Talk Support Facebook Community) engage large numbers of patients frequently because they respond to patients’ need to find information as well as emotional support. Members use peer-to-peer communities to learn not only about care options, but also about strategies to reduce the side effects of medications and hacks to improve their quality of life. This pooled patient knowledge is particularly valuable because it fulfills information needs that are often unmet by the health care system and builds patients’ confidence to become more engaged in their health.
With over 2 billion monthly active users worldwide, Facebook is the most common social networking site for health interventions and is one of the first resources that many patients turn to for help with their condition. First, these online communities thrive when their members are numerous and active. Members of the Facebook breast cancer community that we studied receive feedback on their posts in a matter of seconds because their group is so active and there are always multiple members online.
Second, moderation is key. Online conversations in both online communities that we examined are carefully moderated to eliminate content that is inappropriate or not backed by scientific evidence as well as to maintain respectful interactions. Whether the moderators are users, as in the Facebook group, or platform employees, as in MyHealthTeams, their role is essential for the productivity of the online communities. Third, peer-to-peer communities are appealing because they can be accessed where and when members need them, making them a fast and convenient option for patients. Offering communities that are accessible via mobile apps may be a good strategy to expand the base of smartphone-dependent users who do not have broadband service at home.
Finally, the answers that patients receive from their peers are tailored to their specific circumstances and needs. Peer-to-peer communities display higher levels of information actionability and personalization, characteristics that may increase patients’ ability to use the information and data that they acquire through the platform for their health decisions. MyHealthTeams takes actionability one step further by pooling and returning patient knowledge through infographics that offer simple steps to improve patients’ health and quality of life.
Facilitating Health Information Access and Use
Initiatives focusing on expanding access to information (e.g., Tidepool and OpenNotes) may motivate more savvy and sophisticated patients to visualize their data while leaving behind users who are less versed in technology or those with fewer literacy and numeracy skills. Tidepool combines diabetes data collected by devices from different manufacturers into an integrated device-agnostic platform. This significant advance has enabled patients to download and share all their diabetes data easily with their providers before clinical encounters.
A study to test the usability of the Tidepool app on a small number of patients and caregivers showed that data download increased over time, with the number of users downloading data once per month going from 9% (before the introduction of the app) to 38% (3 months after the introduction of the app). This increased level of engagement has greatly facilitated discussions between patients and physicians, enabling physicians to better understand blood glucose fluctuations and adjust insulin dosing. Physicians also can use Tidepool to identify teachable moments and to improve patients’ ability to self-manage their condition.
However, although the evidence suggests that Tidepool may help patients download their data more often, downloading device data is not equivalent to examining the data independently. Although Tidepool offers the unique opportunity to integrate and store diabetes data in a single platform, its graphs may be too complex for most patients to use independently for the purposes of understanding their condition and calibrating their insulin dosing. Additionally, the data do not suggest specific adjustments that patients can adopt to better control their blood glucose and are not presented in ways that reflect patients’ circumstances and abilities. Patients have to connect the dots to make sense of their data, and some may not have the skills to do that on their own.
OpenNotes has persuaded >200 health systems to make their notes available via their portals to >40 million patients. Surveys of patients who accessed their notes showed that 77% to 87% of patients reported feeling more in control of their care and that 60% to 78% of patients reported improved medication adherence. Doctors who shared their notes reported strengthened relationships with their patients (as indicated by improved communication and shared decision-making) and noted that some of their patients seemed more activated and empowered.
Whereas OpenNotes has successfully expanded access to notes, it remains unclear how many patients actually set up and use portal accounts, which is a necessary step in order for them to read their notes. Research on patient portal adoption, defined as setting up an account, indicates a mean adoption rate of 52%. When it comes to actual use of notes, a survey of VA patients showed that almost a quarter reported viewing notes at least once. Research on the use of notes at two institutions showed that approximately 60% of notes were accessed within a month of their availability but that access was contingent on receiving email notifications. At a hospital where email notifications ceased, the percentage of notes that were viewed dropped from 60% to <20%.
In both cases (Tidepool and OpenNotes), engagement with the platforms is moderate and is driven by clinicians’ requests that patients upload their data or by alerts informing patients that notes are available. Independent and frequent use of the data may be modest and may be linked to the limited actionability and personalization of the data that patients can access. Discussing patient data or the content of notes during a visit may educate patients on how to reap the benefits of this information. Clinical notes may offer some actionable information, such as a patient’s treatment plan or a reminder to schedule an appointment with a specialist, but other information may be too complex and not tailored to patients’ abilities. A stronger focus on making the information simpler and easier to act on may expand the user base of these initiatives. Furthermore, offering the information via mobile apps may facilitate access by a more diverse population and increase convenience for users.
Coproduction of Research
The aim of initiatives focused on the coproduction of research (e.g., IBD Partners and the Health eHeart Study) is to strengthen patient participation in research by collecting large amounts of health data and research ideas produced by patients. The 2010 launch of the Patient-Centered Outcomes Research Institute (PCORI) and its Clinical Research Network (PCORNet), a government-sponsored initiative to build a vast “network of networks” of patient data to promote medical research, supported numerous participatory research networks, including both IBD Partners and the Health eHeart Study.
These initiatives thrive when a large number of patients contribute their health data regularly and for a long period of time. Such sustained engagement is necessary to build a repository of longitudinal data to support research. However, few patients update their data with regularity. For IBD Partners, for example, only about 4% of >15,000 registered users have updated their health information twice a year since the platform was launched. The Health eHeart Study has the potential to become the technology-powered version of the famous Framingham Heart Study, but tracking participants over time remains a challenge. This initiative was able to mobilize about 140,000 members to provide some health data to support cardiovascular research, but few patients returned to the platform to update their information. As an expert interviewed for this research reflected:
It’s sort of a double-edged sword to make recruitment and enrollment so easy for people…if it’s so easy that you can simply click on a link and put in your information and be in a study, then you don’t necessarily feel that attached to the study.
Patient-powered research networks are sustained by patients’ desire to support research on their condition. The connection that should be cultivated to favor such initiatives is the one between patients and researchers. Therefore, showing patients how their data and ideas contributed to research projects and ensuring robust communication and feedback among patients and researchers may improve engagement more than adding features that allow patients to access their data or that provide peer-to-peer social networking capabilities. Enlisting clinicians to educate their patients about the importance of these initiatives in advancing research may expand the base of regular users.
Understanding the Scale and Depth of Engagement
When considering patient engagement, it is important to distinguish between scale (defined as the number of patients who interact with a platform) and depth (defined as the frequency with which patients interact with a platform). Whereas most initiatives in this study benefit from large-scale engagement, different objectives may require different depths of engagement. For example, in order to meet members’ informational and emotional support needs, peer-to-peer communities require the highest frequency of engagement. Without such deep engagement, members may have to wait too long for feedback and may decide to abandon their community, thereby depleting the pool of available knowledge and creating negative repercussions for the remaining members.
Other initiatives require periodic engagement from members to update their patient data (e.g., Swedish Rheumatology Quality Register, Tidepool, IBD Partners, Health eHeart Study) or to view their data (e.g., Tidepool, OpenNotes).
Finally, efforts to include patients in quality-improvement projects or in collaborative research initiatives require sustained, deep engagement by fewer patients. ImproveCareNow, for example, strives to create a learning health system, characterized by the motto “all teach, all learn,” in which patients, caregivers, and providers share information and cocreate and test ideas to improve the quality of IBD care. This collaborative model is based on intense and sustained patient engagement.
Our evidence indicates that a small group of dedicated patients and caregivers play an active role in quality improvement by participating in volunteer work at their care center (the local level) as well as at periodic community conferences where they share their knowledge and priorities with doctors and researchers (the network level). Such activist patients and caregivers develop resources for other patients and their families, from manuals for parents of children who have been recently diagnosed with IBD to toolkits for young patients living with an ostomy. The deep engagement of a small number of activists has positive reverberations for the 30,000 patients who receive care at centers that belong to the ImproveCareNow network.
Similarly, in initiatives that aim to involve patients as cocreators of research, a relatively small number of active patients are highly engaged in platform governance and in shaping research priorities. IBD Partners, for example, was able to generate a fruitful and ongoing collaboration between patients and researchers to identify patient-centered priorities and incorporate them into research studies. Patients have proposed 170 research questions and cast nearly 1,900 votes in a lively online forum. The research prioritization is dynamic and involves online discussions among patients but also regular feedback from researchers, who summarize existing research articles in simple terms and answer patients’ questions. Patient ideas have resulted in studies that have clarified controversial topics, such as the roles of fiber consumption and exercise in IBD, generating highly relevant evidence that patients can use in their disease management. Furthermore, research based on member data is returned to the community, closing the circle between contributing to research and receiving research findings in return.
In the same vein, a small number of very active patients help to shape the Health eHeart Study research agenda by contributing ideas that are meaningful to patients. As observed by one activist patient, “mitochondrial DNA is really exciting to a researcher, but it means nothing to a patient’s quality of life.” Some patients serve as coprincipal investigators in studies, review proposals to assess their patient-centeredness, and participate in platform governance.
Not all patients have the time and resources to engage at the level of the motivated volunteers who make ImproveCareNow, IBD Partners, and the Health eHeart Study possible. Therefore, for initiatives that rely on such frequent engagement, targeted recruitment of patients with potential for sustained engagement may be more fruitful than generic campaigns to expand membership.
Making Health Information Actionable
Several initiatives have shown limited use of the data by patients. For example, many members whose data are included the ImproveCareNow registry are unaware of the initiative and the information and tools that it disseminates. Similarly, independent use of data among members of the Swedish Rheumatology Quality Register, Tidepool, IBD Partners, and the Health eHeart Study seems scarce. A possible explanation is that patients use platforms such as the Swedish Rheumatology Quality Register and Tidepool because they are asked to do so by their clinicians. As a result, some patients may consider these platforms a way to share their data with their providers before a visit rather than as instruments that they can use to actively manage their own conditions.
Our interviews have indicated that patients who are particularly educated, technology savvy, and data driven are able to use their data independently but that, for other patients, the information is too complex. Too often, the information that is disclosed is dictated by the data that are available rather than by the needs of patients. For example, in the cases of OpenNotes and the Swedish Rheumatology Quality Register, the information that is disclosed was originally intended for clinical use and may not be fully relevant or intelligible to patients. As an expert interviewed for the Swedish Rheumatology Quality Register noted:
To make this tool work for patients, we need a better patient overview. What patients get is a visualization similar to the one that doctors get. For example, if patients see their lab results, they don’t really see what the normal values are. They just get a number; it doesn’t say much. Also, the graphical overview is difficult to interpret if you don’t have anyone to ask…. Doctors’ guidance is important. Patients who have been sick for long, and have seen data lots of times, may be more able to make sense of data. But it should work in a way that you don’t need to be with a clinician for 20 years to understand it.
Because the information that is made available is supply driven, rather than centered on the needs of patients, we should not be surprised by limited access and use. Research confirms that tools that do not take into account patients’ “needs and wants” are characterized by low adoption.
Additionally, it is often not clear to patients how data on their disease activity or overall health can be used to achieve better health outcomes. Limited actionability and personalization may explain why patients fail to use certain data independently. Initiatives such as ImproveCareNow and the Health eHeart Study are piloting apps that have the potential to facilitate data use and participation in research. In the future, more sophisticated tools, such as apps to conduct N of 1 trials to understand how medications, diets, or behaviors affect patients’ health, may encourage more patients to pay attention to their health data and prepare the ground for personalized approaches to care.
Being Aware of Institutional Barriers to Incorporate Patient Input
Challenges to engagement may derive from the way in which clinical work is organized. For example, some physicians may fear that giving patients access to their clinical notes could generate confusion and place an additional burden on overworked doctors. However, evidence indicates that the use of open notes did not increase clinicians’ workload. When pilot hospitals invited patients not only to read the notes, but also to provide feedback on them, some clinicians worried they would get a “laundry list” of concerns from their patients. Instead, patients demonstrated competence in identifying problems with their notes, and clinician reviewers determined that the majority of safety concerns warranted follow-up and resulted in changes in records or in care in a majority of cases.
Our interviews revealed that some physicians may be reluctant to alter their routines and workflows to accommodate patient-produced data and increased interaction with patients. For example, ImproveCareNow experimented with different apps to capture patient-produced data; however, a full implementation of such apps on a larger scale would require preparing physicians to receive patient data as well as patient requests that may be motivated by the data. In the case of Tidepool, even though some endocrinologists may welcome the flow of patient data and frequent interactions with patients, other physicians may not have the time, skills, or motivation to absorb the data.
Many physicians are trained to provide care that is disease-focused and may be unprepared to meet patients’ psychosocial needs. Additionally, data on patients’ well-being or quality of life may leave clinicians puzzled on how to react. Finally, incorporating patient-produced data may be challenging because of technological barriers related to linking the data to the electronic health record and regulatory barriers related to protecting patient privacy.
Preparing for Unintended Consequences
Transparency may create perverse incentives that make people less candid and so restrict valuable information flows. For example, clinicians may fear offending patients with their notes and choose to omit information on substance abuse, mental illness, or other sensitive topics, resulting in less-informative assessments of a patient’s health. The OpenNotes pilot found that some clinicians did restrict the content of their notes by omitting information on alternative diagnoses to consider in future visits, personal reminders, or content used to communicate with other clinicians, and some struggled with whether to hide notes from certain patients. Similarly, when reporting on their health outcomes, some patients may tend to overreport positive behavior (such as exercising or following a healthful diet) while minimizing negative ones.
Health Information Validity
Whenever nonexperts are called to the table, doubts about their competence and their ability to contribute arise. Some worry that patients may fall prey to misinformation and neglect the advice of those with hard-earned medical degrees. Research suggests that online health communities suffer from problems related to information quality and reliability. Our evidence indicates that some of the information exchanged in peer-to-peer platforms may not be evidence-based. Members of patient communities are reminded that opinions expressed online should not replace medical advice, yet certain patients can be misguided by unfounded information. For example, members of the Facebook breast cancer community that we examined reported that there are recurring posts on how soy may be harmful to patients or on the benefits of lemons for fighting cancer.
Often, however, other members or moderators are able to identify information that is not scientifically backed and offer evidence to disprove it. Other times, members share information without carefully verifying its accuracy or scientific backing (for example, posting sensational headlines of what causes or cures their condition), which may generate confusion. Finally, illegitimate members may post links with non-evidence–based information with the sole goal of getting clicks and increased Internet traffic. With so many voices and ongoing conversations, some misinformation is inevitable, no matter how fast moderators intervene.
Furthermore, the way the discussion is organized — in threads with sometimes hundreds of replies — may make it challenging to find information. Sometimes original posts are followed by a long trail of responses and members chime in without reading the previous content, which can generate confusion and needless worry, making some threads counterproductive.
Overall, even though some information should be taken with caution, information generated by peer-to-peer communities can be quite helpful. Several community members with medical training reported being satisfied with the reliability of the information that they receive. Finally, it is worth remembering that members often turn to peer-to-peer communities to find information that is not available from the health care system (for example, what clothes are most comfortable after a mastectomy or what breathing exercises are practiced by patients in other countries). As one expert observed, newly diagnosed patients often feel that they have to reinvent the wheel, when all they need is to connect with their peers:
Eighty-five percent of questions that patients have do not require medical attention. In many cases, MDs are not qualified to answer those questions…. Parents of autism-spectrum kids want to know how to get their son with sensory processing disorder to the barber for a haircut…. In the health care system, you see a doctor periodically, but what happens the rest of the time, when you don’t see the doctor and have questions?
The question of information validity, albeit legitimate, may apply only to a fraction of the knowledge that is exchanged in these communities. For many matters, patients appear to have sufficient expertise.
Technology Platforms Are Not for Everybody — but in Some Cases Can Help to Redistribute Knowledge
Several platforms appear to attract more sophisticated users. Research conducted on the Health eHeart Study and IBD Partners showed that the majority of participants tend to be well-educated Caucasian women. In the case of clinical notes, even though only a minority of patients (3%) found notes confusing, those patients tended to be >70 years of age, less educated, and unemployed. Such patients were less likely to find notes helpful for understanding their conditions and their care plan. Another study of patient access to clinical notes found that minority patients were less likely to access notes than white patients. Research has confirmed that patients from certain racial and ethnic backgrounds and of disadvantaged socioeconomic status are less likely to sign up for patient portals, which may explain their more limited access to notes. Patients of nonwhite race and ethnicity and patients who are less educated or have less trust in their doctors tend to be more concerned with privacy when accessing their notes.
At the same time, the benefits of accessing notes may be especially relevant for underserved populations. In one study, for example, patients who were older and less educated expected benefits from accessing notes, just like younger, healthier patient populations, indicating that the interest in accessing notes is generalized. In the cases of ImproveCareNow and the Swedish Rheumatology Quality Register, those who engage with information and data provided by the initiative are likely to be individuals with higher education levels and better literacy and numeracy skills. Similarly advanced skills are required to use Tidepool diabetes data independently.
The two online peer-to-peer communities (MyHealthTeams and the Facebook breast cancer community) may be the only cases with fewer distributive effects, for several reasons. First, both initiatives employ Web and mobile apps. Some minority communities have relatively limited access to desktop and laptop computers. Gaps diminish for smartphone ownership, suggesting that patient communities available via mobile apps may be able to reach a more diverse population. Furthermore, the patients we interviewed observed that online communities are convenient and offer an alternative for patients who may not have access to in-person groups in their area or who may face cultural or financial barriers to attend face-to-face meetings. Some patients discussed how members from remote areas are able to learn about treatment options from users who receive care at high-quality teaching hospitals, suggesting that participating in an online patient community may be a good way to achieve greater levels of bridging social capital and redistribute knowledge among patients.
Intentional Design for Patient Engagement
With the increasing quantity and variety of health information platforms for patient engagement, platforms have evolved from general-purpose tools to become more differentiated and specialized resources. Although they all create new roles for more active patients, many now focus more explicitly on how different patient contributions can improve health and health care. We identified four kinds of platforms in the current landscape: (1) those that join patients more closely to health care providers to focus on health care quality improvement; (2) those that connect patients to each other for support and peer advice; (3) those that make clinical information available to patients; and (4) those that mobilize the production of data and ideas from patients in order to advance research.
These various objectives require very different kinds of participation from patients. Organizations that seek to build effective patient-engagement platforms first should be attentive to the nature of the health contribution they seek from users and then should focus on developing platform designs that will generate and facilitate these specific contributions. This field is developing very rapidly, and it is likely that new health-engagement objectives beyond those examined in this article will emerge. But we believe that three general findings will persist: first, that the days of one-size-fits all patient-engagement platforms are past; second, that health information platforms will continue to differentiate; and third, that the most effective platforms will be those that discover the best ways to engage patients by empowering them to make specific improvements such as enriching their relationships with health care providers, creating mutually informative peer networks, or contributing to the medical research enterprise.
Disclosures: This research is funded by Commonwealth Fund Grant No. 20170973. The authors do not have any conflicts of interest with the initiatives they studied for this research. We are grateful to all the individuals who accepted to be interviewed for this project and to Mary Graham and David Weil, Transparency Policy Project Co-Directors, for their leadership and ideas. We are also indebted to the Ash Center for Democratic Governance and Innovation at Harvard Kennedy School and to The Commonwealth Fund for their support. Learn more about this research and our case studies at https://transparencyforhealth.ash.harvard.edu/.