Camille Mojica Rey Contributing Editor Clinical OMICs
Incorporating Consumer Devices and Disease-Specific Monitoring, Wearables are Changing Healthcare
There is a tectonic shift occurring in modern healthcare that is being driven by technology. In the past, data flowed from provider to patient. A routine office visit would include the generation of measurements, such as blood pressure and weight, at sporadic points in time. Physicians would use these data to make decisions and patients would passively receive care. Now doctors have a new tool in the data arsenal as they incorporate patient-generated, continuous data in decision-making thanks to medical wearables, implantables, and hand-held devices that track their patients 24/7. It’s a future Joseph Kvedar, M.D. describes with his co-authors, Carol Colman and Gina Cella, in 2015’s The Internet of Healthy Things and 2017’s The New Mobile Age: How Technology Will Extend the Healthspan and Optimize the Lifespan. It’s a future in which real-time biometric data is automatically captured and used to learn more about the impact of lifestyle on chronic diseases and wellness.
The ultimate goal is to change our behavior to improve our health. But, we’ve only just begun to realize this vision, Kvedar said. For the most part, we are playing catch-up with the technology. “The reason we haven’t been able to harness the digital phenotype for health outcomes is partly because the data streams are not normalized yet and partly because we as a profession haven’t figured out how best to use the data,” Kvedar explained. “There is a big gap in the amount of machine learning and analytics that has been applied to medical images and genomic data that hasn’t been applied to wearable data.”
Kvedar is in charge of connected health programs at Partners Healthcare, a Boston-based non-profit hospital and physicians network that includes Bringham and Women’s Hospital and Massachusetts General Hospital. He is a pioneer in the field of connected health and first started developing ways to deliver care remotely in 1994. At Partners HealthCare, patient-generated data is integrated with electronic medical records, and numerous mobile health and virtual care programs have been implemented for its 1.5 million patients. But with large patient populations continuously spitting out data, the sheer amount of data generated is daunting to physicians. Progress will occur once doctors get over the misconception that any one clinician has to look at, for example, every blood pressure reading taken. “If we have software analyzing that data and presenting the physician with information rather than streams of data, that would be most helpful. We haven’t gotten there, yet,” Kvedar said.
Evidation is one of the private companies working on ways to reach that goal. Luca Foschini, Evidation’s co-founder and chief data scientist, said measuring in healthcare is usually relegated to sporadic points in time and, usually, when something is wrong. What is needed by just about everyone in the healthcare space—from physicians to drug makers—is continuous streams of data from everyday living. But, Foschini agreed with Kvedar that healthcare providers are not used to multi-channel streams that provide data nonstop. “That’s the new beast here,” he said, adding that engineers and even brokers on Wall Street know more about dealing with these kinds of data than doctors do. “What we are trying to develop are digital biomarkers and interpretation of biomarkers.” That’s because those continuous streams of data reveal biomarkers when the biomarkers are shown to be reliably predictive of health outcomes.
Making the Most of Technology We Have
As some researchers work to improve our use of data, others are taking advantage of available technology. The smartphone—a device market research firm Statista estimates is used by 224.3 million people in the U.S.—is one example of a technology that has not been fully utilized for the improvement of health outcomes. Emmanuel Agu is a professor of computer science at Worcester Polytechnic Institute in western Massachusetts and director of its Health Delivery Institute. He and his team have developed a smartphone app that tracks the progress of wounds as they heal and allows remote assessment of the healing. “A lot of people who have non-healing wounds have to go to the hospital every few days to have the wound checked. Going for the appointment takes half of the person’s day and, in between visits, they wonder if it is getting any better,” Agu said.
According to U.S. Health and Human Services statistics, there are an estimated 5.7 million people with chronic wounds that cost $20 billion to treat annually. These wounds can take nine months to a year to heal, as opposed to the three or four weeks it might take a healthy person’s wound to heal. Many of the patients with these wounds suffer from chronic diseases, like obesity and diabetes. Sometimes, limited mobility means these patients can only get to the hospital for wound care by ambulance, costing an estimated $200 million per year. If complications are missed, amputations might be indicated.
The wound-tracking app combines machine learning algorithms and computational photography, allowing patients or caregivers to photograph the wounds. The app compares the size, color, and other features of the wound over time—features that become digital biomarkers and ones familiar to physicians as being reliable indicators of wound healing. Once these biomarkers are analyzed, the app generates a healing score that lets the patient know if the wound is getting better; is unchanged; or is worsening. It also generates one of three messages: stay the course; consult a wound specialist; or seek immediate care. In a pilot study, Agu and his colleagues used the app to analyze 114 images over the course of a year that were taken at a University of Massachusetts clinic. In February, the team was awarded a $1.6 million grant from the National Institutes of Health to continue the development of the app, expand the use to other kinds of wounds—such as pressure sores—and to address some of the app’s shortcomings, such as inadequate lighting in photos. “Wounds, wound management, and amputations have a huge cost, both financially and physically, for the people who suffer from them, as well as for their families. Wound management is a problem that imaging technology can help with,” Agu said.
In another use of wearable or remote monitoring technology, researchers at the University of Texas Southwestern (UTSW) Medical Center, are testing whether physical activity monitors (PAMs) can accurately measure functional status in patients with cancer receiving chemotherapy and whether patients will comply with use instructions.
“You see young people carrying Fitbits, but most patients with cancer may be older and technology illiterate. We didn’t know if they would be able to do all of this,” said Arjun Gupta, M.D., who was chief resident in internal medicine at UTSW when the study was done. (Gupta is now an oncology fellow at Johns Hopkins University’s Sidney Kimmel Comprehensive Cancer Center.) To determine the feasibility of using PAMs with patients receiving chemotherapy, UTSW doctors gave PAMs to 24 patients with different kinds of cancer who had Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 to 2, meaning they were not bed-bound more than 50 percent of the time. The team reported in the May issue of Clinical Cancer Informatics that PAM-derived measures of physical activity correlated with clinician-assessed ECOG scores. Only one patient was eliminated from the study. The rest used the PAMs as instructed, wearing the monitor more than 50 percent of the time over 10 weeks and uploading data to a smartphone weekly. Gupta said minimum steps taken per day was the most sensitive biomarker. “If a patient took less than 800 steps, it meant they were not getting out of bed much and not doing very well,” Gupta said. This information is critical to caring for patients. Those with poor PS, as determined by clinical judgment and patient questionnaires, have poorer survival, reduced response rates, and worse quality of life. Unfortunately, those who want more chemotherapy despite their PS may report feeling fine, while those who want less may report feeling worse than they do. Gupta said PAMs have the potential to give oncologists a more realistic clinical picture of their patient’s PS.
Team Approach to Innovation
Harnessing the power of wearables to improve health outcomes for chronic diseases—particularly in underserved communities—is a complex problem that is most likely to be solved by large, multi-disciplinary teams of experts. At least that’s the idea behind the National Science Foundation–funded Engineering Research Center (ERC) on Precise Advanced Technologies and Health Systems for Underserved Populations (PATHS-UP) led by researchers at Texas A&M University (TAMU). The PATHS-UP team, which includes everything from engineers and medical researchers to behavioral psychologists and fashion designers, also includes investigators from UCLA, Rice, Florida International University, and an evaluation team from Illinois University. The team’s research spans device development, deployment, and patient acceptance.
Gerard Coté, Ph.D., is a TAMU professor of biomedical engineering and director of PATHS-UP. Coté’s own work focuses on developing hand-held and wearable point-of-care technologies and systems using optics, electronics, microfluidics, paper fluidics, nanoparticles, and assays. These technologies are used to detect and diagnose chronic diseases (diabetes, cardiovascular, cancer), blood toxicants (BPA, PCBs), and infectious disease (malaria). The PATHS-UP research is focused on developing technology—lab-in-a-palm and lab-on-a-wrist—that will be used in underserved communities to monitor diabetes and cardiovascular disease. “The research also focuses on participatory design and stakeholder engagement in which we get the communities and stakeholders involved so that we are not just throwing technologies at them,” Coté said.
The PATHS-UP researchers have two main goals. First, they want to engineer new technologies that can overcome the barriers usually faced by point-of-care devices, which can be unreliable, difficult to deploy, and expensive. The second goal is to recruit and educate a diverse group of scientists and engineers who focus on four areas of research:
- Implantable devices for continuous monitoring of biomarkers, and microfluidic cartridges and assays for finger prick monitoring to determine biomarkers that do not have to be monitored continuously;
- Development of affordable hand-held imaging and sensing devices for use with the cartridges and assays;
- Wearable devices for monitoring the biochemical markers with the implant and physical parameters noninvasively; and
- Development of the models and algorithms for smart data analysis, sending the right information to the right person, and for modifying behavior.
Coté said it is still very early in the process of turning wearables, hand-helds, and implantables into useful, medically relevant devices.
“Like many emerging technologies, the early units are market–driven and hence focused on fitness rather than health. I believe this will change as we get more devices out there that are true medical devices that have the required sensitivity, specificity, and robustness to pass regulatory hurdles and focus on real medical conditions that can truly impact health,” Coté said.
He also emphasized that there are three kinds of data that are patient-generated: chemical/biochemical (blood glucose levels), physical (blood pressure), and contextual (resting, walking, running). The ideal device would gather all three kinds of data and do so redundantly so that back-up systems are available in case of failures. And, its use would begin with the gathering of baseline data. In that way, Coté said, healthcare providers could truly leverage the continuous data being generated by an individual.
Mark Wolff, chief health analytics strategist at SAS, a maker of data analytics software, agreed with this assessement. “As we collect and analyze growing volumes of ever-more granular data about the diagnosis, treatment, outcomes and value of care, we are realizing the promise of true biomedical informatics. Such a wealth of data—and the ubiquity of connected devices—will ultimately drive the adoption of more intelligent clinical decision support systems.”
As Kvedar predicted, connected care in the new mobile era is truly becoming a reality.
Devices at The Forefront
If you don’t count smartphones and fitness trackers, the first true wearable medical devices are continuous glucose monitors (CGMs). Currently, Medtronic (Guardian Sensor 3) and Abbott (FreeStyle Libre) both have CGMs on the market. Medtronic’s MiniMed system includes their CGM and an insulin pump, making a hybrid personalized closed-loop system. Abbott has partnered with Bigfoot Biomedical to create a similar insulin pump/CGM system. CGMs rely on sensors that are worn on the back of the arm or abdomen for seven to 10 days. The devices mean that those who use them can avoid the painful finger pricks normally required to measure blood glucose. Abbott also makes Confirm RX, a smartphone compatible insertable cardiac monitor used to diagnose patients with cardiac symptoms.
There are many lessons to be learned from the makers of these front-line devices. Ali Dianaty, vice president of research and development for Medtronic’s Diabetes Group, said the focus was on clinical outcomes during the development of the CGM sensor. Patient comfort came second. “Today, the sensor is small, flexible, and painless to insert. Patients wearing it don’t feel it,” Dianaty said. “These are the little things that drive choice and how compliant patients are.”
Likewise, Avi Fischer, M.D., Abbot’s cardiac rhythm management medical director, said he and his colleagues have learned the value of easy-engagement of patients in their own healthcare. The company conducted research among different age groups and cultures when creating the device and its smartphone app, which has now been translated into 30 to 40 languages. “The feedback we have received from patients has been almost uniformly positive. There is no use in having a device out there that is not user-friendly.”
This article was originally published in the September/October 2018 issue of Clinical OMICs. For more content like this and details on how to get a free subscription, go to www.clinicalomics.com.