The workshop’s second panel presented three aspects of how consumer-facing technologies can catalyze widespread informed engagement with the health care system. Patricia Dykes senior nurse scientist and program director for research in the Center for Patient Safety Research and Practice and the Center for Nursing Excellence at Brigham and Women’s Hospital and assistant professor at Harvard Medical School, discussed the skills that health professionals need to interface with consumers regarding consumer-generated information. Lygeia Ricciardi, principal of Clear Voice Consulting, then spoke about how digital health tools can create incentives for consumer engagement in health care. Alison Rein, senior director for Evidence Generation and Translation at AcademyHealth, addressed some of the ethical, legal, and social issues that are associated with consumer eHealth and big data. An open discussion moderated by Bernard Rosof followed the three presentations.
Patient-generated health information, explained Patricia Dykes, is health information that is created, recorded, gathered, or inferred by or
1 This section is based on the presentation by Patricia Dykes, senior nurse scientist and program director for research in the Center for Patient Safety Research and Practice and the Center for Nursing Excellence at Brigham and Women’s Hospital and assistant professor at Harvard Medical School, and the statements are not endorsed or verified by the IOM.
from patients or their designees to help address a health concern (Derring, 2013). Patients can submit health history or treatment history information, or they might submit symptom logs so that their health care team members can assess how their patients are doing with self-management. At Partners HealthCare, patients routinely report outcome data related to cardiac surgery and orthopedic surgery. One of the fundamental characteristics of patient-generated health information, continued Dykes, is that it is captured by patients, recorded by patients, and shared by the patients if they decide to do so. Many patients, for example, collect biometric data and they may or may not share with their provider.
Patients believe that submitting these data is important because they can help keep the EHR up to date, and because they can provide a more complete picture of the patient’s status than the data typically collected in the context of a visit. For example, patients who report their blood pressure readings over time are giving their providers a more complete picture than the snapshot readings taken in the office. Updated medication lists give providers a view of not just what they ordered, but what patients say they are actually taking. Patient-generated data can also provide the care team with access to the patient’s problems, concerns, goals, and preferences, which can then improve the care plan concordance between patients and the care team, Dykes explained.
Given that research has shown that when providers pay attention to patient-generated information it improves engagement and activation (Hibbard and Lorig, 2012), what are the skills and attitudes needed to promote this kind of attention? One attitude, perhaps an overarching attitude, said Dykes, is the value proposition of patient-reported health information: that it can help solve problems and provide key benefits that improve care. “I think one of the challenges is communicating this to providers so that they are willing to interface with patient-reported health information,” said Dykes. “We think the value proposition is that with the rise of smartphone and mobile devices, patient interest in health information has really exploded. Patient attention to health information when the provider is paying attention, too, actually improves engagement, activation, health, and wellness.” In particular, Dykes noted the work of Judith Hibbard, who has shown that patients who are activated are more likely to undertake positive health behaviors, have lower costs, and have better outcomes than patients who are not (Hibbard and Greene, 2013; Hibbard et al., 2015).
In addition to the attitudes related to the value proposition, communication skills are essential for providers interfacing with patient-reported health information, something that Dykes noted was discussed in the workshop’s previous session. “Now, more than ever, providers need to have expert communication skills, both face to face and when using technological tools,” she said. “They need to know how to identify decision oppor-
tunities, recognize those opportunities, and take advantage of them. They need to then take the time to engage the patient using decision aids so that the patient can participate in the decision making and understand the information in the level of health literacy that is appropriate for them. This type of communication requires both a change in attitudes and training for many health care providers.” Dykes added that communicating with surrogate decision makers is another important concept that many clinicians need to learn. It comes into play when surrogates need to report patient information that the patients themselves are unable to report because of their medical condition or because of language barriers. At Brigham and Women’s Hospital, for example, physicians and nurses engage care partners who are communicating for patients who are in intensive care and oncology units.
Assessment skills are also important when it comes to determining how health literate a particular patient may be, and the outcomes of these on-the-spot assessments should determine the types of tools that the care team uses with a particular patient. Again referring to research by Hibbard, Dykes noted that disease management improves when matching a patient’s activation level with the care plan and using the activation level to determine the types of tools and the intensity of follow-up that is required for a patient (Hibbard et al., 2009).
Another important skill that the health care team requires is the ability to sort through patient-generated data. “There is a lot of fear about this,” said Dykes. “Today, relatively few patients are contributing health information to their electronic record and submitting it to their providers. We need to adopt our informatics strategy so that we can process big data, not only taking these data in, but having the ability to display them in a way that they are going to be usable for clinical decision making.” Currently, she explained, there is an active debate about whether or not data entered into the EHR require review. She noted that in preparing for this presentation, she spoke to many providers who told her that they do not want to see Fitbit data because they have no idea how they will process that information within the time constraints of the 15-minute office visit.
What the field needs, said Dykes, is a taxonomy to differentiate those data that need to be reviewed versus those data that patients can just submit into the EHR. To deal with those data that the patients do submit, the field needs to develop a workflow to make use of those data. In the Brigham and Women’s Promoting Respect and Ongoing Safety through Patient-centeredness, Engagement, Communication, and Technology (PROSPECT) project, for example, Dykes and her colleagues thought carefully about the workflow needed to accommodate data from every piece of technology they implemented to ensure that any data a patient submitted would be considered by the health care team.
The PROSPECT project is one approach to addressing the provider skill
set that is needed for processing patient-reported information in an acute care setting, in this case for patients receiving a bone marrow transplant or in the intensive care unit. PROSPECT is a cluster-randomized controlled trial that began in June 2014 and was scheduled to run through May 2015. Its goal is to optimize the overall experience of patients and care partners by facilitating engagement, improving care plan concordance, promoting dignity and respect, and enhancing satisfaction. The first of two important components to the intervention is the Patient SatisfActive Model, a structured team communication model that tries to improve patient experience and satisfaction in real time by giving providers the skills, tools, and workflow to improve communication between clinicians and patients and their family members. This training aims to promote attentiveness to the patient’s needs, concerns, and expectations in real time, and to engage the patient and family in the decision making, paying attention to cultural diversity; and then looking closely at each patient’s health literacy levels and how that determines the tools that the care team will use with a given patient.
The second component is a patient-centered, Web-based toolkit and microblog, essentially a Web portal where patients can report their overall goal for hospitalization, state their daily goals, and rate how well they think the care team is doing with their care (see Figure 4-1). Patients can also view the care team’s goals, see problems that care team members have identified, and then provide feedback to the care team. Patients can also message their care team about their care plans at any time. PROSPECT also has a set of provider-facing tools that are designed to integrate the review of patient-reported information into the workflow. For example, the safety checklist component of this toolkit puts as its first item the ability of the nurse to present any new patient or family input from the Patient SatisfActive webpage or microblog. The provider-facing tools also serve as the interdisciplinary care plan platform where the care team reviews what the patient or family reports over the past 24 hours and then agrees on a strategy for addressing any concerns. “For example they discuss whether they are going to go in during rounds and talk to that patient and family, in a family meeting, or whether they are going to use the microblog and respond that way” Dykes asked.
Implementing the Patient SatisfActive Model has been successful and appears to be a reasonable approach, Dykes said, but there have been some challenges that PROSPECT had to address. Working with incapacitated patients and elderly, for example, requires engaging care partners to input information into the patient-facing Web portal. Workflow integration has gone well with teams based within the hospital, but has been less successful with non-local providers and consultants. PROSPECT has improved the ability to manage a multidisciplinary care plan by incorporating patient preferences, goals, and priorities and by reconciling siloed nurse and phy-
SOURCE: Dykes presentation, March 24, 2015.
sician documentation, but there is still room for improvement, she said. A major barrier when the project went live, and one that Dykes and her colleagues are still addressing, has been getting providers to understand the value of patient-generated information.
In conclusion, said Dykes, patient-generated health information can be used to inform care decisions and to promote shared decision making. “It is important that we first recognize the value of patient-reported health information and that we consider what the impact is going to be on workflow and address that as part of the project,” she said. “Attention to provider readiness and logistical skills are key to success, and implementation is not without real-world challenges. True stakeholder engagement in designing the data, information, and workflows was an important preparatory activity that has helped us quite a bit through this project and continues to help us.”
Ricciardi discussed the consumer or patient perspective on the use of digital health tools. She started by discussing the tools available for increasing consumer engagement in their health care and noted that there is a rapidly growing number of consumer health apps, mostly monitoring exercise and diet, with some tied to wearable sensors or biometric devices. Forecasts project there will be some 68 million users of wearable electronic fitness monitoring devices worldwide in 2015 (Gartner, 2014), and that more than 70 million personal health and wellness devices will be sold in the United States alone by 2018, generating more than $8 billion in product sales and service revenues.
People are beginning to use these apps and wearables, according to research from the Pew Research Center (Fox and Duggan, 2013), but the potential for growth remains substantial. According to the Pew report, 69 percent of adults track their health or the health of others in some way (whether digitally or in a non-digital format), with the majority recording weight, diet, or exercise, but only 21 percent of trackers are using technology to track health. People with chronic health conditions are most likely to be trackers. Another study found that 10 percent of adults own an activity tracker, but that 50 percent who have owned one stop using it and 30 percent stop within the first 6 months (Ledger and McCaffrey, 2014). Ricciardi counted herself among those who own activity trackers that now sit unused in a drawer, though she added the devices still impacted her being engaged in her health. As an example, she said that after she had her first child, she joined an online weight-reduction program for only a few months, but from which she took lessons she is still using today. “I do not think that a measure of success for an app or a tool has to necessarily be that I have it strapped to my body forever,” said Ricciardi.
The take-away from these figures is that people are using these technologies, but that they have not yet met their full potential. One question that needs addressing is how to learn more about what technologies people do or do not adopt. “How do we choose?” she asked. “Right now, there is no real guidance on what to use, and I would suggest that a potential role for the IOM might be to create a rubric or a framework through which public- and private-sector entities could come together and help develop a code for what is desirable in these technologies.” In her opinion, the key characteristics of technologies are that they are useful, user-friendly, either good looking or invisible, comfortable, hardy, and affordable. They would also come with transparent privacy practices, integrate easily with other
2 This section is based on the presentation by Lygeia Ricciardi, principal of Clear Voice Consulting, and the statements are not endorsed or verified by the IOM.
products, and be of a trusted brand with good reviews. She noted that one app she recently tried sent a text message to one of her friends telling her that this app was now mapping her runs. As far as she knew, the app never asked permission to do this and she would not have known it was doing so if her friend had not alerted her to the fact. “I was unhappy and I still hold it against that product,” said Ricciardi. “I imagine that I’m not the only consumer who is going to feel incensed when we find our information is being used in ways that we were not informed about.”
Besides her personal and informal list of desirable characteristics for an adoptable technology, there are other motivators for people to adopt apps, tools, and technologies, such as the fun factor and the extent to which the technology can be tailored to the user’s personal health goals and preferences. Technologies that contribute to society, perhaps by enabling the user to contribute data to a pooled database for research or other purposes, are more likely to be adopted, too. As an example, she cited the fact that 80 percent of the clients of the genomics company 23andMe say they are willing to share their data for research.
There are also a number of extrinsic motivators that can encourage people to adopt or use some of these tools. Money is a big one, and potential sources include employers, payers, providers, and accountable care organizations. As an example, Ricciardi described the Welltok Health Platform, an app that employers can offer that gives people incentives such as coupons based on adoption of healthy behaviors. Castlight is an app that employers can provide to their employees to enable them to compare the cost of potential service providers, thus enabling them to realize financial savings. Other extrinsic motivators are social in nature, such as the app Stickk, which enables users to bet with their friends that they will meet some health-related goal, such as going to the gym a certain number of times per week. If the user does not meet the goal, the app donates money to a cause the user designates as one that he or she does not like.
With regard to what is on the horizon, Ricciardi believes there will be increasing use of extrinsic incentives. “I think we will see more of the major players, from providers to payers to affordable care organizations, who are motivated through policy changes to get consumers to be more engaged in their health,” she said. She noted that she would like to see more evidence about what kinds of apps and tools are effective because that will help consumers choose among the thousands of choices, and she predicted that this evidence base will contribute to a market consolidation. Because health is so integrated into the rest of life, she predicted that the lines between health apps and tools and generic ones would blur. “You wouldn’t think of having a separate designated health phone and a finance phone,” she said, and she expected that more generic information, such as what a person buys at the grocery store, who is in their social network, and how often they interact
with members of that network, might become part of a person’s health information. In short, she said in conclusion, she was hopeful there will be “increased adoption and use of these tools in a way that they really become a ubiquitous part of our lives in a way that helps us reach the health goals we define as individuals.”
Alison Rein started the final presentation in this session by explaining that she oversees a portfolio of work at AcademyHealth that straddles two major focus areas. The first is how to use these new data sources, coming from both within and outside of the health care system, to generate better evidence. The second is how to engage patients and consumers more meaningfully in that evidence-generation process. She also emphasized just how important the ethical, legal, and social dimensions of “this brave new world of consumer eHealth” are to getting the most out of these new data sources.
To help provide some background to her talk, Rein explained that AcademyHealth envisions a future where individuals and communities are made healthy by the use of evidence in decision making. To realize this vision, the organization’s mission is that, together with its members, it will work to improve the health and performance of the health system by supporting the production and use of evidence to inform policy and practice. The Electronic Data Methods Forum is one program that AcademyHealth believes will help it fulfill this mission. Funds for this project are part of a major infrastructure investment by the Agency for Healthcare Research and Quality (AHRQ) through the Health Information Technology for Economic and Clinical Health (HITECH) Act, enacted as part of the American Recovery and Reinvestment Act of 2009. The forum recognizes that with the increased use of EHRs, there is a potential gold mine of data and information in these systems. However, there are some thorny issues related to methods, governance, data quality, informatics, and the application to a learning health care system that must be addressed to tap the riches of this gold mine. The forum, which also recognizes that tackling each of these issues independently in their separate silos will take far too long to create a learning health care system, has created a learning community around these issues that engages people who are trying to use electronic health data for operational, quality improvement, and research purposes. “We use this as
3 This section is based on the presentation by Alison Rein, senior director for evidence generation and translation at AcademyHealth, and the statements are not endorsed or verified by the IOM.
an opportunity to do collaborative work, to share insights, and to work on shared challenges together,” said Rein.
Another initiative that AcademyHealth started some 4 years ago is called the Consumer Patient Researcher Roundtable, which again focuses on collaboration. “We recognized early that very often when we went to meetings, the interests of academics and researchers were presented as polar opposites from the views and interests of patients and consumers, which is actually odd when you think about it because I never met a researcher who said that he or she was trying to do things to the disadvantage of the patients he or she serves,” said Rein. As a result, the roundtable promotes conversations and looks for opportunities in the eHealth space where researchers and consumers can work together.
Rein noted, as did Ricciardi, that while there has been a great deal of focus on the robust data streams coming from EHRs, there is an explosion of data from other sources, including devices, apps, and websites. “I think this is a golden opportunity for people where I sit, for the research community, and for pretty much everybody who cares about health because we can now look at data that go beyond the 10 percent of health that is represented by or attributable to what happens in the care experience and look at the 90 percent of health that is affected by everything else,” said Rein.
The ability to use technology to better engage patients offers tremendous opportunities and potential benefits, said Rein, who noted several examples:
- Improved patient access to their own clinical information via providers and insurers.
- Engagement with health and wellness information via portals and websites.
- Enhanced quality and coordination of care.
- Expanded opportunity to contribute to and participate in research.
- Increased ability to “quantify self” via apps, wearables, and devices.
As Ricciardi and others mentioned, Rein said there is a tremendous appetite on the part of patients and consumers to engage in this space. A report from the PricewaterhouseCoopers Health Research Institute stated that 47 percent of consumers agree that mobile devices can be used more effectively to coordinate care, and that 65 percent of consumers with one or more health apps on their mobile devices use them at least once weekly. Half of the consumers surveyed said they would be likely to use consumer devices to perform self-evaluations and 56 percent said they would be comfortable having their health data shared if doing so would improve care coordination (PricewaterhouseCoopers, 2014). Rein noted that the ONC
and the National Partnership for Women and Families each released survey results over the past year that corroborate and extend these findings.
Given the appetite consumers have for eHealth applications, it is important to consider the risks and whether those risks are evenly distributed across all constituents. Loss of privacy is the risk that gets the most attention and it can take two forms: (1) personal identification and blind discrimination, or (2) discrimination by association. As an example of the former, Rein noted how purchasing an item online suddenly seems to trigger a host of related content appearing on other websites. Discrimination by association refers to a characteristic that a person might have that is associated with something else, such as prior incarceration. Another risk is loss of control of one’s own information, and to some extent, this may have happened already. Diminished access to potentially beneficial information or treatment is a risk that is associated with the fact that many vulnerable populations may not get access to care because they cannot purchase an app or device and have to relinquish some level of control in order to gain access to something they believe they need medically. Rein said these risks are not evenly distributed.
There is no comprehensive framework for health information privacy and security, Rein said. Although this has been said many times, it bears repeating, she explained. Societal norms informed by the Hippocratic Oath, together with legal protections afforded by the Health Insurance Portability and Accountability Act (HIPAA), the Genetic Information Nondiscrimination Act (GINA), and the Common Rule, are only in force for data generated by the health system. For patient- and consumer-generated data, there are few protections and many points of potential risk, a situation that Rein characterized as being like the Wild West.
“Given all of this,” said Rein, “I think the question becomes what things are health literacy sensitive? That is, if we apply basic health literacy concepts and practices, could we reasonably expect consumers to take some different action? Conversely, are there things that are so complex or so opaque or where there is no transparency that it is just really too much to expect that consumers would be able to engage or interact in that way?” Consumer use of mobile apps provides one example of something that is potentially health literacy sensitive. Mobile apps, websites, and devices all capture the users’ data, and in many cases are contributing data to a larger, aggregate dataset from others who happen to use those technologies. In large part, explained Rein, these data are aggregated, de-identified, and then used for a variety of purposes. “There is some commoditization happening on the back end that is important to recognize.”
She also noted that many of the vehicles used today for informed consent do not really do the matter justice. She mentioned a project that the Electronic Data Methods Forum conducted with John Wilbanks at Sage Bionetworks to create a set of wire frames and extractable components that can be used to obtain consent via a mobile device. Some of these components were included in Apple’s research kit. This participant-centered consent toolkit uses icons and words to walk the participant step by step through the consent process and test the user’s understanding of consent at each step.4
A less clear area of concern entails consumers searching for information on health-related websites. More than 70 percent of American adults look up health information online, but only 36 percent know that advertisers are allowed to track their visits to health-related websites. Moreover, some 77 percent of searches begin not at a health-related website such as WebMD, the National Institutes of Health (NIH), The Mayo Clinic, or AHRQ, but at Google, Bing, or Yahoo. Most of these searches track back to large corporations, such as Google and Facebook, and to data brokers, such as Experian and Acxiom. These are what is known as third party requests. First party requests, explained Rein, are those in which the user gets information directly from the website being visited, whereas third party requests are those for which the host website has to get information that it extracts from other sites. Those requests “leak” information on the user to the third parties, triggering privacy concerns.
“Is there a health literacy–sensitive solution to these concerns?” asked Rein. Her answer was “maybe.” Her pessimism stems from the difficulty that the average person, including herself, would have in protecting his or her own data without help and the fact that add-on software solutions are themselves burdensome to end users. Even when done well, data protection is only moderately effective, claimed Rein. “Again, going out on a limb and being a bit provocative, this may be an opportunity where regulatory and/or policy solutions are needed.”
Bernard Rosof started the discussion by asking panelists to comment on the key challenges they see in getting consumers to sustain their use of consumer-facing technology and on getting providers to use the data generated by their patients’ use of the technology. Rein said that one general rule regarding consumer sustainability is that the technology needs to be easy to use without changing the consumer’s workflow. If a device runs out of batteries, needs recharging constantly, or is hard to set up, the consumer is much more likely to drop it. She also thought that just as physicians do not want all of the data that a Fitbit generates, neither do consumers. “What we need is smart software that helps us interpret key changes in the data or that alerts us to pay attention to things because we are going to be increasingly measuring and recording more data about every aspect of our lives, and none of us is going to want to review it all,” said Rein. “What are some ways that we can use intelligent technology to make sense of the data and highlight useful key points? That will be worthwhile.”
A workshop participant agreed with those points and added that there needs to be a return on investment for both the patient and the provider; one way to increase the return on investment would be with the kind of software that Rein described. Such software, said the participant, needs to support decision making, depending on who is using the data. Return on investment increases as the data become more usable and useful, and there is evidence, said the participant, that patients become activated and engaged in their health care when the data they provide are useful to their providers. “I do not think all providers know that and I do not think all providers understand exactly what patient activation is,” said the participant. “These are all skill sets that I think we have to reinforce and teach.” Rein then added that PatientsLikeMe reinforces the idea that patients give something and they get something back. “That has certainly helped keep a steady stream of people engaged in that site,” she said.
Rein also noted that providers have a high threshold for evidence. “I have talked to many who say ‘I want to know before I totally redesign my systems and my processes that I am going to get value from doing this,’”
she explained. She also noted that the process of showing that particular types of patient-generated data are useful is going to be long and fraught with failure. The field needs to understand that some data will be useful, and some will be useless. Ricciardi added that there is an assumption that the categories or types of patient-generated health data will need to be approved before they goes into the EHR, but she and her colleagues do not think this is necessarily the case. “If patients want to report their goals or particular data sets that they believe contributed to meeting those goals, they should be free to do so. Discussing the goals and the data with health care providers will help patients and providers to better understand one another,” said Ricciardi.
Bringing up the concept of behavioral economics and the way in which incentives were used with meaningful use, Rosof asked if a similar approach was going to be taken to encourage uptake of consumer-facing technology. Ricciardi predicted there will not be a meaningful use program for consumer health information technology. What she hopes will happen is that payment reform, particularly with regard to changing the underlying financial incentives in the health care system, will align the patient and provider in terms of having the goal of doing whatever is necessary to keep the patient healthy in an efficient way. “That, I think, will necessitate the engagement of consumers and individuals in their own health, which means relying on technology,” said Ricciardi. “We rely on digital tools and digital information in every other industry and it just makes sense. More information enables us to make better decisions.” She recounted that when she worked at ONC she found that not all stakeholders were afraid of patient-generated data. Many physicians, she said, wanted nothing to do with these data because they were not sure how to process it or how to be reimbursed for their time in doing so, while some of the major retail pharmacy chains were eager to have access to patient data and to use it to build stronger relationships with consumers.
Rosof then asked Rein to comment on how EHR data will be used. Her response was that they are being used today for quality improvement initiatives, comparative effectiveness research, and population health management. “A lot of characterizing of populations, of segmenting them, of trying to assign some type of risk, and of developing a proactive plan for mitigating risk is happening,” she said. In the past, the need to wade through thousands of paper records to find the necessary information for such studies made them unreasonable to conduct. Rosof noted that “patients are actually contributing to society without really knowing about it,” to which Rein said, “Sometimes, yes, unless they fully understand the HIPAA waiver.” She suggested that for health literacy reasons, it might be a good idea to change that waiver so that it is more understandable to more people. Ricciardi added that while many people—both providers and
patients alike—think that HIPAA prevents individuals from getting access to their own health data, the fact is that HIPAA guarantees everyone’s right to access their own health information. That component of HIPAA, she said, should be just as well known as the privacy piece of HIPAA, but unfortunately it is widely misunderstood.
Ruth Parker added that she wanted to make one thing clear about how EHRs are being used: “Billings are improving and that is a big issue for patients. The EHRs that we have and that we use came to be because we addressed that part of it. I think we are in the embryonic phases of figuring out how they can be good and useful.” Rosof asked Parker if EHRs could identify gaps in care that could lead to performance improvement and better outcomes. Parker replied that she believes this is possible and is being done is some places. She also stressed that while many institutions have built out their EHRs for the principles related to outcome and performance improvement, it is important to understand that the original interfaces were designed primarily for billings. Rosof remarked that there are four drivers of provider behavior: professionalism, performance, payment, and patients. “How you prioritize those seems to set institutions apart, and I think EHRs need to get up to that task.”
Rosof asked the panelists to comment on the Apple research kit and how it might affect research. Will it skew the data? Will it lead to increased health disparities? What are the ethical and privacy implications of such a system? Rein replied that Apple has established a baseline level of setting expectations in terms of disclosures and privacy concerns for those who want to use the research kit. She also said that the question of skewing is an important one, but noted that she had heard some 48,000 people had signed up to participate in research projects in just the first few days the research kit was available. This suggests that this system might broaden participation. What could be skewed, she said, is that this system excludes users of the Android operating system, which happens to serve many potentially vulnerable populations. “There is now this pressure to build a comparable mechanism to conduct research and apply the same tools and framework to the Android space. I do not think it is going to be exclusively an Apple endeavor for long, at least based on the conversations that I am hearing,” said Rein. “Then there are endless opportunities for far larger sample sizes than we have enjoyed in prior research efforts. Obviously, that means there is another set of methodological and other issues that are going to be triggered.”
Ricciardi commented that Apple, because of its cache among consumers, is helping to put health on the map for the average person even though the average Apple user tends to have a higher income because the products are more expensive. “But I think their entrance into this space, combined with their skills at marketing and mainstreaming things,
is changing mindsets so that people will associate health with their phone or watch and start feeling like these devices and apps and tools are part of the mainstream experience of their lives.” Although traditional research will certainly continue, the opportunity is growing to augment traditional research with real-time data from people in their natural environments.
Rima Rudd, senior lecturer on health literacy, education, and policy at the Harvard School of Public Health, noted that there was a roundtable forum 5 or 6 years ago at which there was a discussion on ownership and control of data, particularly with regard to traveling population groups such as the homeless who might need to make a decision on what data they want to share with a clinician. That control and ownership were absent from this workshop’s discussions so far, she said. Dykes said the question of ownership comes up frequently in the comments that patients submit to their microblog. Currently, in the research context, patients are not getting their data. She and her colleagues are exploring the recommendations and infrastructure needed to be able to provide patients with access to these kinds of data.
Ricciardi said she likes the model in which patients collect their health data and choose to whom to give access and how it will be used. “I tend to dislike the term ‘ownership’ in this particular context,” she said. “I do not think anyone really owns data in an electronic environment in which it can be easily copied and dispersed.” Instead, she prefers the idea of control, of giving the patient more control over how information is used rather than “going down the rabbit hole of discussing who owns something. It’s a small shift in how we talk about it, but I would stay away from the term ‘ownership’ per se.” Rein said the issue of who is allowed to do what, why, when, and in what context remains an overarching theme in much of the work she does. Health systems, she said, act as stewards of the data because they have a legal record that they are responsible for maintaining and they do maintain a higher standard for access and sharing through the Institutional Review Board mechanism. Outside of the health system, though, there are far fewer controls and probably less knowledge about controlling access to data. What Rein would like to see is that all data are returned to the individual and that some tool, which has yet to be developed, nicely organizes an individual’s data and enables that person to authorize who gets access to those data.
Rudd then commented that health literacy is a two-sided coin. On one side is an individual’s listening and comprehension skills, while on the other side is the health care provider’s speaking and explanatory skills. “One cannot measure someone’s listening skills without measuring the speaking skills, and one cannot measure literacy skills without understanding the difficulty and complexity of the various texts,” she explained. In that context, she asked Dykes if her project includes an evaluation of both patient
literacy skills and some measure that can differentiate if one team has better communication skills than another. Dykes replied that the research team did create a standard for communication and that providers are documenting when they address patient’s needs, concerns, and expectations. What her team does, then, is track the frequency with which they document those factors during different phases of care, such as when they first enter the intensive care unit, when they are being treated, and when they are transitioning out of the unit. She noted that getting providers onboard with this requirement was a paradigm shift of sorts because the clinicians did not think this kind of information was valuable.
Parker asked the panelists if anyone has asked patients and consumers what they really care about with regard to privacy and data access. Rein said the only thing she knows is how much she does not know about what consumers want because it is wrong to assume that all consumers want the same thing. “I think there is a tendency to say patients and consumers all need and want the same thing, and my observation is just that there is a continuum. Depending on where you are in the continuum of your life and your health, your needs and expectations are going to be profoundly different,” said Rein. “We need to be able to acknowledge that and design for that at the care level, the research level, and every level.” She also acknowledged that this is not a satisfying answer to the question, but that it is the best that she has been able to come up with over the past 10 years.
Ricciardi said she agreed with Rein’s assessment, but added that there are some generalizations to be made based on surveys of what people expect with regard to health care. For most people, their primary concern is to be well, and many do not want to think about their health or health care, especially if there is not a problem. Many individuals with a chronic condition do not want to identify as someone with a chronic condition. People also express real interest in having a greater connection to their physician and they want to communicate more. “They want to establish a closer relationship, rather than one based on a 15-minute visit once a year,” said Ricciardi. Another obvious want is for health care to be easier, something that she identifies with as a mother of two children. “I hate that I have to drive in a rain storm, park my car, and bring my kid into the pediatrician just to show that she has that same eye infection that I knew she had because we’ve treated it in the past. If I could have just taken a picture with my smartphone, and if my doctor would trust me and would be reimbursed for prescribing whatever I needed to deal with the fact that she has pinkeye, everyone would be so much better off,” said Ricciardi. “Make this much less difficult for us please!”
Earnestine Willis, the Kellner Professor in Pediatrics and director of the Center for the Advancement of Underserved Children and of Health Equity and Urban Clinical Care Partnerships at the Medical College of
Wisconsin, asked Dykes if the PROSPECT project was changing both what physicians were doing and how the institution was getting information into the community. Dykes replied that PROSPECT and other efforts are changing how the institution communicates with patients by providing the community with the tools that allow them to communicate with and purposefully engage care partners. She noted, too, that it is important to engage people in the community because their capacity to engage and be activated is much less when they are in the hospital. Moreover, unless the community is engaged, and that includes a patient’s family and other potential caregivers, they are not going to be activated when they leave the hospital and return to the home and community setting. “In this project, we are trying not just to enroll the patient, but also to ask the patient who else can work with you to make sure that when you are not feeling well that we can still further your care plan and make sure that your goals are achieved,” explained Dykes. She added that this is an attitude shift that she hopes is being communicated to the community, that Brigham and Women’s Hospital is taking the time to not only engage the patient, but to make sure that others in the community are working with patients when they go home to further their care plan.
Willis also asked if the panelists have given any thought in this world of technology about issues of liability, given the recent data thefts that had occurred at two large health systems. Rein said this is certainly a concern and one that may be contributing to the reluctance of providers to absorb patient-generated data. She noted that Kaiser has been looking at what happens to patient-generated data once they have been integrated into the workflow and accepted into the EHR, but that this is an area that needs further exploration.
Suzanne Bakken asked the panel to speak about the idea of preparing the public, starting in kindergarten and running through high school, to participate in research studies using consumer-facing technologies. “Thinking about what we have been discussing today, what are the implications for clinician education in particular so that they can be integrating these tools into their practice more effectively?” she asked. Ricciardi strongly supported the idea of starting in the K-12 environment to create the mindset that everyone is welcome to participate in research and must be responsible for being more engaged in their health, and technology provides a set of tools to do that. She noted that when she was at ONC, there was a great deal of discussion about a shift in mindset being one of the biggest challenges facing the health care field. “What is required is not only changes in technology or payment reform—those are important pieces, too—but a shift in attitudes is really a critical piece of what we are talking about, and you really cannot start too early with that,” said Ricciardi. She noted that ONC is conducting a big public relations campaign related to the Blue
Button initiative in partnership with a number of nonprofit organizations to get the word out and get people in the mindset that they need to be involved in maintaining and managing their own health. “But I would love to start talking about the individual’s role relative to personal health earlier in schools as well,” she said.
Dykes agreed that attitude change is important and that changing provider attitudes was an important early piece of her work. When she and her colleagues first started implementing the patient-centered toolkit, providers expressed concern that patients were going to be communicating with them. “They did not see the value of giving patients this much information, that it would make the patient anxious,” said Dykes. It was necessary, she said, to make the case that engaging with patients was best done by providing them with the same information that the clinicians had, but in a way that would be useful to the patients, using resources such as Medline Plus. She also stressed the importance of obtaining provider input on workflow. “They have to be part of this. You have to hear their concerns and move with them,” said Dykes.
Winston Wong noted that the Roundtable on the Promotion of Health Equity and the Elimination of Health Disparities was going to be holding a workshop that would discuss how to increase minority participation in clinical trials. He wondered if the panelists had given any thought about how health information technology could be used to facilitate the involvement of underserved populations in clinical trials. Ricciardi responded that ONC held a roundtable on this topic when she was working there. “One big take-away for me was that you need to involve a lot of trusted community organizations or other representatives of the communities that you are trying to reach. It could be a house of worship, a health care organization, or a school, anything that the folks you are looking to as a group consider a home or resource that they trust. They say health care is local, and it does revolve around particular cultural communities as well as geographic ones. Members of a given community will help you figure out how specifically to employ technology—it is going to vary based on what group you are talking with and their particular norms and expectations,” said Ricciardi.
In a PCORI-funded project aimed at preventing falls, Dykes and her colleagues are running a clinical trial at 10 sites across the United States that involves a health information technology intervention. Nurses serve as fall care managers and make sure that patients receive and understand their assessments and tailored plans. The details for this intervention were worked out with the help of stakeholder councils that represented the different areas in which this research would be conducted. In some regions, particularly the Southwest, minorities are the majority and they formed these councils. “It is making sure that you are touching base with all the stakeholders that are going to be representing users of the technology and
participants in the trial and addressing their concerns,” explained Dykes. “We did that in preparation for submitting the application, but every month at every clinical trial site we continue to meet with these stakeholders and do refinements when we run into trouble.”
Michael Paasche-Orlow said he helped to build an interactive voice response system that called parents before they brought their children in for an appointment. He asked the physicians how often they wanted to see the data that the parents provided to the system. The answer was never, he said. Nonetheless, the investigators automatically input the data into the clinic note and the physicians were thrilled because it enabled them to easily upcode their billing. He also commented that in his opinion, nothing will make the HIPAA provisions understandable to the general public because they are too complex. He and his colleagues conducted focus groups and tried to explain to people that the health care provider would be responsible for protecting their information with a certain set of exceptions. The response of the focus group participants was to ask if they should trust the health care provider or not. “I do not think readability is going to be the answer with HIPAA,” he said. He asked the panelists if they see HIPAA as a barrier to research and as an empowerment for patients.
Ricciardi pointed out that the HHS Office for Civil Rights oversees enforcement of HIPAA and that the HIPAA notice was redrafted 1 year ago with significant input from ONC to be relatively clean, readable, and understandable. She said that she believes the new draft underscores the important point that patients have access to their records and that there are protections in place. Paasche-Orlow responded that he does not think HIPAA is inherently understandable to the patient, to which Ricciardi replied that perhaps it was possible to explain HIPAA generally, albeit not in every detail for every encounter with the health sstem. In response to a comment from Paasche-Orlow that HIPAA seems to be a barrier to the flow of information, Ricciardi added that the sad thing about HIPAA is that it was built with the best intentions and for good reasons, but that it has been implemented terribly. “If we could just get out the word that HIPAA at a minimum guarantees you the right to get your own health information and that you can then share it with anyone you want,” said Ricciardi. “I think if that understanding were more prevalent, HIPAA need not be such a barrier to the flow of information.”
Rein noted that there is little reason for any institution not to share information with another HIPAA-covered entity, but that is not how the HIPAA regulations are interpreted. She added that there are business interests that exacerbate this problem. She suggested that if the market conditions changed, data liquidity would result. Where she is spending time, she said, is thinking about the non-HIPAA-covered entities, how data are flowing among commercial interests that may not be covered by HIPAA, and
what reasonable protections might be. “I think that terrain is wide open for further exploration,” said Rein.
The final question in the discussion came from Jennifer Dillaha, who asked the panelists if they were aware of the Family Educational Rights and Privacy Act (FERPA). That Act limits the health care system’s access to the tremendous amount of health data in the educational system. Rein said she did not have a solution to this problem and that it was another example of where there are data that have to do with health that are not held by traditional health entities. “It is not until we have a different model of payment where people are actually responsible and accountable for the health of a population that we are going to see data flowing between and across all of those entities because they realize that just doing little things around the 10 percent of your health that is attributable to care does not actually make people healthier,” said Rein. She added that in the interim, the federal government has been working over the past 5 or 6 years on open data, establishing plans for exposing more data resources and, hopefully, having greater communication and collaboration among federal agencies. “Maybe there is an opportunity to flag this issue at that level and see if there could be some interim attraction before we have solved the payment reform problem,” said Rein. Ricciardi, while acknowledging that she is not a FERPA expert, said she believes FERPA has right-of-access provisions similar to those in HIPAA—that people have a right to their data. “Maybe, just like with Blue Button and health information in general, the individual is part of the solution here,” said Ricciardi. “You get your own data and you are able to combine them with data from other sources and share them with whomever you please. I would think about that as a solution—empowering the individual to be the one who gets the data and shares them.”