A panel of experts discussed advanced technological developments in telehealth, including patient-generated data, remote monitoring of the use of social networking, and wireless health. The following sections reflect the individual speaker’s comments and reflections.
David S. Muntz, M.B.A., CHCIO, FCHIME, FHIMSS
Office of the National Coordinator for Health Information Technology
Technology creates changing roles for and demands from patients. For example, technology can create unbounded expectations. There used to be short-term, mid-term, and long-term planning horizons, but today, the presence of mobile devices has developed the immediate-term expectation. Technology and its evidence base need to be developed more rapidly to get the technology into the hands of the people. In addition, people already have technologies that will have to be deployed in a number of areas.
In terms of the roles, who will be the primary coordinator of care? Is it the physician with PCMHs, the patient, or the patient support groups? How can different groups be connected? Meaningful use requirements will create meaningful conversations about the roles for the respective parties. Who will be the custodian of data? In virtually all states, the medical records person or the health information management person is the custodian of data, but that is likely going to change to be the purview of the patient. This would help solve some significant problems (e.g., privacy, confiden-
tiality, security, access). However, the country still has a significant digital divide. For those who are unable to be custodians of their own data, solutions such as health information exchanges will be needed, and somebody else will have to act as that primary coordinator of care. The Office of the National Coordinator for Health IT (ONC) has developed a consumer e-health group that focuses entirely on that issue. ONC talks about three “As”:
- Access: How do you get a patient to the data?
- Action: How you get a patient to take action on the data?
- Attitudes: How do you change the attitudes about care?
Products on the market today enable the monitoring of daily activities inside the home. How can they best be packaged and used? What reimbursement system will support their use? How will people be trained? How can information be gathered to promote particular actions to be taken? There are many challenges with patient-generated data. The first relates to the engagement of the patient and includes issues of culture, literacy, privacy, confidentiality, and security. For example, who is the custodian of an adolescent’s health information? How does it affect the doctor-patient relationship? How will individuals who do not know how to use computers be assisted with the use of these technologies? How will providers keep pace with the amount of information available to patients on the Internet? Will competition develop once patients have access to the same information their providers do?
Second, many have questioned the reliability of patient-generated data. In a face-to-face meeting, providers may use body language to determine truthfulness. However, the reliability of physician data may also be questioned. Some patients have discovered that some of the information in their medical record is not accurate. Having everybody look at the data ensures that the value of the data goes up and the integrity stays as high as possible.
Lastly, there is a need for the intuitive collection of data. If the use of a device requires special training, it probably will not be used. To this point, software will need to be designed better for intuitive usability. If a device is easy to use and is able to ensure that safety measures are put in place, it will deliver a better outcome. Finally, the technology needs to be incorporated meaningfully into the lives of the consumers.
Consumers are looking for a trusted source of data, but online searches produce a significant amount of inaccurate information. If we want people to use technology to gather data, and to be stewards of their own information, then we need to figure out how to improve the integrity of that data. There is a gap between the potential and the reality of telehealth. For example, only 15 percent of consumers have renewed a prescription online, only about one fourth want to adopt digital health records, and less than half think digital records will boost health care delivery. In response, the ONC developed several initiatives. In the Consumer e-Health Pledge Program (www.healthit.gov/pledge), more than 350 organizations (representing more than 100 million Americans) pledged to provide access to personal health information. Both data holders and non-data holders are encouraged to participate. ONC also looks to share real stories of people who have successfully used health information technology (e.g., the Million Hearts campaign, Beat Down Blood Pressure) with the intention of engaging more individuals. A current initiative rewards individuals for sharing stories of how having access to health records improved the quality of their care.
In the end, said Muntz, consumers need all the stakeholders to connect, communicate, and collaborate so they can better understand what an electronic health record can do.
Bonnie Britton, R.N., M.S.N., ATAF
Eastern North Carolina is very rural with a great deal of poverty, illiteracy, and chronic disease. Vidant Health is the largest health care system in North Carolina, serving 1.4 million people. Seven of the system’s 29 counties are among the top counties for chronic disease in all of North Carolina. As a result of previous experiences with telehealth in the state, Vidant Health developed a remote monitoring program for patients with cardiovascular disease and pulmonary disease for all 10 of its hospitals. Vidant Health’s goal for the program was to identify inpatients to be referred to a telehealth program that would monitor their blood pressure, pulse, weight, and oxygen saturation in their homes on a daily basis. Additionally, the program incorporated a patient activation measurement tool, which assesses the patient’s engagement in their own health care. The program is directed at patients who score low on this tool—the patients who are distrustful or fearful of health care, who believe care is the responsibility of the health
care provider and not the patient, and who tend to be noncompliant. The program also includes elderly, homebound patients with chronic disease.
Many people are initiating remote monitoring in the patient’s home. One of the big lessons is to develop a program based on the payment structure. To determine their target population, Vidant Health looked at their core measures, what will be reported publicly, and where they stood on the linear graph of value-based purchasing. Another lesson is that care needs to shift from the hospital to the home setting. With the patient at the center of care, there has been a lot of research about care coordination and transitions in care. However, the models for those programs have very high nurse-to-patient ratios, which make them very hard to scale up, and is unaffordable. With remote monitoring, the nurse-to-patient ratio is one nurse to between 85 and 100 patients. In addition, changes are needed in hospital care management. Another strategy is the PCMH; four Vidant Medical Groups are seeking certification as PCMHs, and they all have incorporated telehealth into their programs.
To be successful, remote monitoring programs should be based on best practices. For example, programs should focus on the top 5 percent of users—the high-risk patients with chronic illness. Assessments of engagement are needed so that providers can teach and coach based on the activation level. Goals should be patient-centered goals, not the goals of the provider. For most patients, these goals will be very small steps to encourage engagement.
Additionally, specific protocols are associated with success. Patient-selection criteria should be inclusive versus exclusive. For example, Vidant Health’s only exclusion criterion is that the patient does not have electricity. Programs need standardization for patient identification, screening, and enrollment. There also needs to be a provider plan of care in the ambulatory record, and a physician referral in the inpatient record. In the Vidant Health program, licensed practical nurses train patients on the use of the equipment and test patients’ competency. They also do medication reconciliations at discharge and the day after discharge, when they install the equipment in the patient’s home.
Programs need to be data driven. The Vidant Health program accumulates data on demographics and objective clinical data (e.g., height, weight, blood pressure, pulse, oxygen saturation, low-density lipoprotein, hemoglobin HbA1C). Financial data are also important. Finally, integration of the electronic health record is critical. Telehealth vendors are willing and able to do this integration, but the electronic health record vendors need to be pushed.
Finally, capital is the biggest roadblock for remote monitoring, especially lack of reimbursement. Providers need to see this as a cost avoidance and cost savings strategy.
For the 65 patients who completed Vidant Health’s remote monitoring program, data have been analyzed by pooling all of the data from 3 months prior to telehealth, during telehealth, and 3 months after intervention. About half of the patients were female (54 percent), African American (50 percent), and over age 70 (53 percent). Primary diagnoses were heart failure, diabetes, and hypertension and the primary insurance was Medicare. As seen in Figures 9-1 and 9-2, hospitalizations and bed days were both dramatically reduced after implementation of the program, both during the monitoring, and in the 3 months after monitoring had stopped. Among other benefits, this frees up hospital beds for surgical patients and other more critical cases. It also allows patients to return to their homes and communities.
FIGURE 9-1 Impact of remote monitoring on hospitalizations.
SOURCE: Reprinted with permission from Bonnie Britton (2012).
FIGURE 9-2 Impact of remote monitoring on hospital bed days.
SOURCE: Reprinted with permission from Bonnie Britton (2012).
PatientsLikeMe is an online platform for social networking for people with chronic illness. PatientsLikeMe is uniquely positioned as a generator of data. As we talk about what kind of data we can gather from outpatient life, platforms like this are a cheaper way to get at some patient outcomes. PatientsLikeMe was founded because Stephen Heywood, diagnosed with Amyotrophic Lateral Sclerosis (ALS),1 and his family (a family of engineers) became frustrated by the quality of the data available regarding individuals with ALS. Literature on ALS is limited to very small cohort trials, and there are not many information manuscripts about what day-to-day life is like. There was no place for someone to come and talk about what it is like to have that disease (or many other chronic diseases) with a data-driven perspective. Many places have narrative content (e.g., blogs), but the Heywood family wanted to supplement some of that narrative content with data content.
Today, PatientsLikeMe is an online network of more than 150,000
1 ALS is also known as Lou Gehrig’s disease.
patients who have signed up to use the website. These members can connect with others like themselves for personalized learning and support. They enter their own information over time, and the website graphs the information, which gives an immediate visual perspective on what their disease course has been like. They can use that information to dive into the richer community experience. For example, individuals can look for other people who are on the same medications, who have had the same disease for the same amount of time, or just share common symptoms. Most of the members have private profiles, which means that their information can only be accessed by other members of the PatientsLikeMe community. PatientsLikeMe knows the members’ e-mail addresses for adverse-event reporting, but do not know their real names. The majority of patients using the site have severe neurological diseases (e.g., ALS, fibromyalgia, Parkinson’s disease, epilepsy). These are not the diseases that are the big drivers of care in the remainder of the population.
In this community, people are surprisingly open about severe pathology. They score their quality of life and take disease-specific questionnaires to track things like medication dosages and side effects. Members talk about their primary diseases and their comorbidities, and share symptoms on a none/mild/moderate/severe scale. This enables others to get a sense of what it is like to actually have a specific disease based on a visual presentation of data. The site also has an area where people talk about what it is like to have a specific disease in more of a narrative view.
So, PatientsLikeMe is a place for people to come together and talk, but what is the benefit? One-third of the people with epilepsy in the online community say they had no one in the real world with whom to discuss their disease. Rural populations especially may have difficulty finding support groups. This site provides ongoing access to a support group anywhere at any time. For the majority of people with epilepsy, the site gives them a better understanding of their seizures. In addition, epilepsy patients frequently discontinue medications because they do not know if something is a side effect, because it makes them feel weird or strange. Most cases of uncontrolled epilepsy are due to non-adherence to medication. Information on the site helps them understand their side effects, which drives them to be more adherent to their medication in many cases. The site also leads to patient activation: Twenty-one percent of members with epilepsy said that as a result of using the site, they insisted on seeing a specialist.
Few care systems have adopted good places for patients to be meaningful custodians of their health care data. When people are trying to manage their health online, they are doing it in the context of other information. PatientsLikeMe allows patients to share their personal health information with others (e.g., peer groups, clinicians) in an isolated system. For providers and for care teams, PatientsLikeMe provides a clinically robust
understanding of the patients and real-world outcomes by aggregating the data and then supplying that data to interested parties. For example, PatientsLikeMe discovered that due to the side effects, patients with multiple sclerosis frequently take their biologic medications at night rather than during the day, but this is not something that clinicians frequently tell them to do.
Additionally, PatientsLikeMe provides the breadth and adaptability of social networking and crowdsourcing, but the presence of data is an underlying pin to these narrative threads. For example, if someone is considering a drastic therapy that is not well received in the peer-reviewed literature, a group of individuals on PatientsLikeMe may have personal experiences with the procedure and be able to share the data they tracked. So as opposed to health information that is placed onto the Internet without any fact checking, this allows for a data-based double check for some of this health information.
The goal of PatientsLikeMe is to build a world where every patient’s treatment is shaped by every other patients’ experiences. This is a complementary role that patients can bring to telehealth, especially in the context of things like remote patient monitoring. In order to zero the sensors in the PCMHs, we need to have some of this complementary data about what was going on in the patient’s life that day to make these sensors more robust and useful. Overall, PatientsLikeMe is trying to integrate with as many data streams as possible to make the entire system work well for patients.
Mohit Kaushal, M.D., M.B.A.
West Wireless Health Institute
The three main macroeconomic drivers of health care today are rising costs, the epidemiological transition, and shortages of health care professionals. First, approximately 18 percent of the gross domestic product is spent on health care, and this percentage is growing. Secondly, the population is aging. This is important because older adults have more chronic disease, which drives even more cost. Finally, to compound all of this, there are not going to be enough providers to care for everyone. Technology, especially mobile technologies, can solve some of these issues.
Health Care in Transition
Today there is a real shift from paying for transactions and volume of care to paying for outcomes. Essentially, the site of care needs to be shifted from expensive, centralized, bricks-and-mortar hospitals that are managed
by physicians and nurses, to sites outside the hospital, along with the “de-skilling” of health care. Labor productivity in many other industries has improved over the past decade, while it has lagged behind in health care.
The current model of health care is reactive: It is characterized by low-frequency visits based on when the physician can see the patient; it is location-centric and high-cost. West Wireless Health Institute coined the term “infrastructure independence” for a new model of health care—a proactive system that provides the right treatment at the right time wherever the patient is and for a lower cost. For example, the earlier detection of patients’ illnesses and management in home-based settings rather than hospitals (or in independent living facilities rather than nursing homes) would result in huge savings.
Health care reform has had several inflection points. First, payment reform will likely reward best clinical practices. Second is the focus on technology, such as the digitization of health care. In other industries, digital data and analytics have transformed productivity and outcomes. More data are needed to help improve the health care system. The third practical effect of health care reform is that many physicians are becoming salaried employees.
The taxonomy of wireless health is extensive and includes terms like mHealth, wireless health, and telehealth. The system of wireless health includes various components. First there is data input with mechanisms (e.g., sensors) to capture physiological parameters. Then there is data transmission wherein data are moved via wired and wireless networks. Next, data need to be stored and analyzed so that they can be changed from raw data to meaningful information for health care practitioners. Finally, an appropriate user interface is needed to empower the final user with all this information.
Mobility allows the capture of data and transmission of information anytime, anywhere, and anyplace. However, that alone is not enough. The wireless health industry has moved away from just mobility toward creating end-to-end solutions that include analytics and user interface. In essence, technology must be implemented within the right clinical process. Furthermore, technology may help develop new clinical processes to care for patients for a fraction of the cost and with better outcomes.
On the data input side, sensors are becoming cheaper and more ubiquitous. Beyond capturing basic vital statistics (e.g., blood pressure, pulse, weight), technology is beginning to capture more valuable data. For example, when treating congestive heart failure, sensors in a patient’s bed might be able to pick up signs of decompensation 2 weeks before that patient
becomes symptomatic. This represents a convergence of health information technology, service delivery, and user interface and design. The amalgamation of all these very different disciplines will contribute to positive final outcomes. It is not just about making a device wirelessly enabled; rather, it is about how this device better manages patients for a fraction of the cost.
Optimism and Barriers
There are many reasons to be optimistic about wireless health. First, networks are ubiquitous, whether they are wired or wireless, and can transmit more and more information. Unfortunately, huge parts of the country still lag behind in the connectivity piece, which needs to be solved. Second, the consumer-scale production of smartphones and other devices has led to a proliferation of applications for consumers, providers, and other caregivers. The ones that are creating a lot of value are appearing slowly, but will continue to grow. Next, decision support may be the most important piece. There is a real generation gap now in analytics in health care versus other industries. The challenge in health care is how to capture multisource raw data (e.g., multiple sensors, medication compliance, the electronic medical record, social factors) and turn it into something meaningful, especially to determine unknown levers. Finally, the current rate of inflation in health care cannot continue, and will force the redesign of the health care system.
The VA has demonstrated compelling data with a home telehealth program. They showed a 19 percent reduction in hospital readmissions for people within that program and, for the patients who are admitted, a 25 percent reduction in bed days (Darkins et al., 2008). However, this is not just about a technology. Rather, it is about the right payment model, the right culture, the standardization of process, the use of care coordinators, and then the right technology to help augment and accelerate all of that.
Unfortunately, there are still significant barriers to the use of wireless health. First, baseline connectivity is the prerequisite to empower everything. Next are issues of interoperability and liquidity of data. Data are currently isolated in non-interoperable systems. Within the wireless health space, the concern is that front-end devices cannot talk to the back-end data warehouses or the electronic medical records. The next area of concern is how the different technologies integrate. Finally, clinical evidence needs to be perpetuated around the final value proposition of all of this. Again, the key point is that the technology has to be implemented in the right care processes to get the best outcomes, rather than developing technology just for technology’s sake.
Moderator: Kamal Jethwani, M.D., M.P.H.
Partners Healthcare Center for Connected Health;
Harvard Medical School
An open discussion followed the panelists’ presentations. Jethwani began the session by noting that several speakers referred to technology as a tool—that telehealth is not about the technology itself, but about the people and the processes. He also referred to discussions about who should be the custodian of data, and the role of patients in generating, validating, and sharing their own data. Audience members were able to give comments and ask questions of the panelists. The following sections summarize the discussion session.
Jethwani spoke about the need for intuitive data collection, perhaps integrating data collection into the flow of daily life. He suggested this might be the role for wireless and mobile health, since phones are ubiquitous and part of the daily flow of work and life. Clifford added that people frequently look to mobile technology as the solution, but that while the number of health and fitness applications has risen dramatically, most people rarely use these applications beyond the first download. He stated that the mobile application development community is not going to solve the problem alone. Instead, he argued, the panelists had highlighted that technology can only solve the problems of health care insofar as it is integrated intelligently into work flows, life flows, and clinical flows. Jethwani agreed, adding that this is why mHealth developers need to come together with the larger team in care plan design.
One participant asked if there are any data on how patients would like to connect with providers using technology. Clifford responded that the individuals who use PatientsLikeMe are not representative of the entire population (e.g., have high health literacy, are extremely engaged in their health care, have already embraced technology) and so may not reflect the needs and demands of all patients. He stated that their members want health care providers to have a dialogue with them, to speak to them intelligently, and to allow them to be participants in their care, and that they largely do not care whether it is virtually or face-to-face.
Evidence and Research
One participant asked if patients will start to decide the research questions, noting that patients are starting to aggregate themselves and identify researchers who are interested in studying their conditions. Clifford responded that there are several emerging models that run counter to traditional research models, and that this might be useful for some research questions. He noted one project that is trying to build a new health commons that includes portable legal consent—the ability for any person to give global consent to allow the data that they gather about themselves (from electronic health records and other data sources) to get aggregated into an open database that can then be used by researchers internationally. He referred to another part of this project in which a group of individuals can agree to a research question, make their data available on a website, and offer a monetary award to whichever researcher can come up with a more apt model to answer that question than the one that exists.
Muntz added that more should be done to promote clinical research. He asserted that health care providers need to be more connected to the clinical trial enterprise because treatments can significantly impact the course of a clinical trial. He noted that in his previous organization, physicians were often not aware of the clinical trials happening within that same site, so they instituted a system that queried a database of all the people enrolled in a clinical trial and would alert clinicians when that patient came in for care. Muntz added that this needs to happen across boundaries, such as in health information exchanges, to promote clinical research.
One participant asked whether smaller, local programs can serve as pilots that lead to more research. Britton noted that North Carolina is holding a summit to discuss starting chronic disease consortiums that allow for centralized remote monitoring (to keep costs down), to keep standardization in place, and to serve as a repository of data. This would require just one institutional review board and have the ability to bring all that data together to have greater numbers to take to the federal government.
Clifford disagreed that the RCT should be continued as a gold standard, stating that such research is less valuable than the tremendous amount of data collected by sensors, and yields real-world information. A participant responded that the randomized controlled trial is “here to stay,” but should only be one part of the research armamentarium, and that newer research techniques should be used as well.
Adoption of Technology
One participant talked about linking small programs that yield positive findings with clinical translational science programs that might also
be able to address how to improve adoption of newer processes based on these findings.
Another participant noted that adoption of previous health care tools (e.g., stethoscopes, X-rays) was not necessarily based on a tremendous amount of evidence, yet telehealth is held to a higher standard. He asked how this related to the development of federal policy because the government has a history of being unable to change quickly in response to rapid innovation. Kaushal stated that in most cases, technological innovations will outpace regulatory or policy innovation. He agreed that many of the connected technologies discussed on the panel have less of a value proposition in a fee-for-service world, and so payment models are key. Kaushal further added that he is less worried about culture change for clinical process innovation, because, he argued, once the right incentives are in place (e.g., providers can get paid for the use of a technology), patterns will change. He noted that another key lever is to get regulatory clarity around how innovators can get some of these technologies through the FDA’s clearance process. Clifford stated that FDA regulation does not make sense for the use of data and technology in clinical decision making.
Muntz questioned what compels people to act and collaborate. He stated that in the past, the markets were not efficient and many people developed software that would not talk to other software. He argued that in many ways, the government is the last resort when a market does not do things efficiently. Muntz maintained that stimulus money helped to computerize health care records in a way that could not have happened without government intervention. He suggested that current regulations are forcing interactions in a way that creates more of a commodity-like use of the data.
One participant noted that while some patients are highly educated and look to their physicians for advice, others merely want to be told what to do. She asked how payment systems will enable providers to go into the patient’s home, reconcile medications, and get the patients to their appointments, given that the less engaged patients are always more expensive. Britton argued that from a policy perspective, one needs to act locally (e.g., state Medicaid programs) and that federal policy changes will be difficult without randomized controlled trials. Britton added that from a practice perspective, hospitals are often willing to act because they want to change the way they do things so they can take care of patients and wrap services around patients who are the most vulnerable.
One participant asked what we need to do to make affordable broadband connectivity available to everyone. Kaushal referred to Muntz’s earlier remarks regarding the role of government, and suggested that this is where the FCC has a role to play. However, Kaushal stated that there are other areas where there is no market failure; he asserted that connectivity will become ubiquitous as the cost of implementing technology and building and managing networks decreases.