by Laurie Loscocco
Modern health care generates more data and more information than ever before. Comparing one person's data to the universe of data now available will help health providers of the future recognize disease patterns, confirm diagnoses and consult with other experts worldwide. Each patient will have access to his or her medical record and genetic code information 24/7 anywhere in the world.
In traditional medicine, a healthcare provider's insight into a patient is somewhat like a snapshot. In P4 Medicine, it's a documentary being continuously edited.
At its core, P4 Medicine is about data, but not just the data derived from an office visit or a hospital admission. Mobile devices that track and transmit blood glucose levels, cell phones that follow someone's physical activity and other surrogate markers, such as implantable fluid monitors, provide richer sources of information.
Genetic and genomic information is becoming increasingly available, and use of the electronic medical record adds another layer of data. Additionally, patients and their families are now able to contribute directly to their medical records.
The amount and types of data being produced have enormous potential in the field of P4 Medicine. They also present challenges.
"We are dealing with unprecedented amounts of data. The question is how to extract that data into usable information," says Philip Payne, PhD, chair of the Department of Biomedical Informatics and executive director of the Center for IT Innovation at Ohio State.
"A patient's healthcare record may be coming at us in 15 different ways, in different languages. The challenge is to aggregate, integrate and create relationships," says William Dalton, MD, PhD, of the Moffitt Cancer Center and Research Institute in Tampa and board member of the Personalized Medicine Coalition (PMC) in Washington, D.C. "Ultimately, you want a system that provides real-time response, is dynamic and learns as you use it."
Informatics is a very rich source of new knowledge, says Daniel Masys, affiliate professor of Biomedical and Health Informatics at the University of Washington, but it also has the potential to result in information overload.
"Omics-based medicine yields many millions of data points at once, but only a small number of those are informative in terms of making medical decisions," Masys says. When looking at genomic information in particular, "We have to apply strict mathematical methods to ensure that any patterns you see are not due to chance alone."
Clay Marsh, MD, executive director of the Center for Personalized Health Care at Ohio State and a member of the PMC board says, "We need to be able to separate the signal from the noise." Many of the technology pieces are in place to do so, says Marsh, who is also senior associate vice president for Health Sciences Research at Ohio State and vice dean for Research in Ohio State's College of Medicine. "It's a matter of putting it all together."
Peter Embi, MD, vice chair of Ohio State's Department of Biomedical Informatics and chief research information officer for Ohio State's Medical Center, agrees. "There is a lot of infrastructure already built around this. Now it's a matter of repurposing it."
The approaches to some of the issues include changing workflow to better collect and store data or using the right computational tools when looking at pattern recognition.
In P4 Medicine, patients themselves are providing an increasing amount of data. Embi is working with Ohio State's Division of Rheumatology and Immunology to reorganize workflow to more systematically collect and store data. Part of the restructuring involves a change in patient engagement, he says. For example, rather than spend an idle 15 minutes in a waiting room, patients could contribute to their own databases, noting any changes or problems since their last visit. The Medical Center's OSUMyChart online consumer patient portal allows patients to log in, enter and review information and communicate with their health providers via the IHIS electronic medical record system, which went live systemwide in October.
Opportunities to capture information are continually evolving. Social media and retail marketing present novel ways to present a clearer picture of someone's health.
A customer loyalty card may be able to tell whether a patient filled his or her prescription at a specific pharmacy, or what kind of foods they are buying at the grocery store, Embi says.
"This is not about being 'Big Brother', " Embi cautions. "It's about partnerships. We need to be very careful with privacy issues."
Indeed, Dalton says, "Data governance is extremely important. You've got to protect the patient at all times."
At the same time, Payne says it's vital for practitioners to be able to "close the loop" to ascertain whether there are obstacles preventing orders from being carried out or the patient is having problems getting a prescription filled.
Experts agree that electronic medical recordkeeping should be standardized to make the data more fluid and transferrable. "It's a research tool," says Edward Abrahams, PhD, president of the Personalized Medicine Coalition. "The more data you have, and the better data you have, the more effective you'll be."
Current policy efforts focus on improving health information technology to allow easier data sharing, more balanced privacy controls, rapid learning and feedback models and predictive models for physician practices, Abrahams says.
The human genome will also become a widespread and rich source of data.
The falling price of gene sequencing, coupled with high-capacity screening technology, means that in the next five years, it's possible that everyone will have their genotype done, Marsh says.
However, he notes, gene sequencing tells only part of the story. "The interplay between a genome and its environment is as important as the genome itself."
Payne says lifespan data "are the most important data we can get, and familial information is crucial." Ohio State is working with Nationwide Children's Hospital in Columbus on a project to collect data from pre-conception through early childhood.
In an effort to delve deeper into factors that contribute to perturbations in health and to try to prevent them, Ohio State is embarking on another project to spot heart failure patients who are most at risk for readmission. A team, including experts in cardiology, computer science, public health and complex systems, will look for "outliers" among heart patients.
The initial phase will involve obtaining micro narratives from 300 patients drawn from the 60,000 people who are insured through the University's health plan and receive care at Ohio State hospitals and ambulatory sites.
Ann Pendleton-Jullian, professor in Ohio State's Knowlton School of Architecture, specializes in complex systems approaches to problems. She said these micro narrative "probes" bring information to the system without altering it.
She conducted a similar type of project for the Department of Defense to determine where radicalization for terrorism occurs related to Afghanistan. Workers with recording pens were sent to Afghan camps and neighborhoods where they interviewed people about topics related primarily to justice.
Using those micro narratives and software that transformed the data into a three-dimensional model, researchers could visualize where the outliers – and the most instability – were. The research showed that most radicalization occurred not in camps but in London suburbs.
At Ohio State, teams began collecting micro-narratives this summer. "We're trying to get more than clinical data," Pendleton-Jullian says. "We're looking at genetic factors, other health factors and behavioral factors that affect the risk of readmission. We're also looking at how someone's emotional community can affect wellness. We're looking at things in their entirety."
Embi says several "non-clinical" factors may contribute to a patient requiring re-hospitalization. "Maybe we didn't fully understand their home situation. Were they able to fill their prescriptions? Did they have cab fare to make a follow-up visit with their primary care doctor?"
He and Payne say researchers will mine the data to determine what key characteristics and decisions have put patients at high risk, so that practitioners can intervene.
A precursor to the micro-narrative project has been the use of indwelling biosensors to predict admissions. These fluid-monitoring devices sent information to physicians and provided early warning of heart failure problems, which reduced admissions by 39 percent, according to Marsh.
Such "hot-spotting" projects are essential to controlling the cost of health care, Marsh says. For example, Medicare and Medicaid databases indicate that 1 percent of that population base accounts for 23 percent of healthcare spending; 5 percent accounts for 50 percent of the cost, he adds.
In one famous story among healthcare advocates documented in an article by Ateal Gawande in The New Yorker magazine, a physician in Camden, New Jersey, examined data from local hospitals and discovered that a single building sent more people to the hospital with serious falls than any other in the city, resulting in almost $3 million in healthcare bills. Looking further, he found that two city blocks incurred the highest healthcare costs. One contained a large nursing home, and the other, a low-income housing tower. Public health officials, in turn, focused a number of health improvement efforts on this neighborhood.
"Hot-spotting" such as this, and other methods to interrogate data, will create a higher level of care, Marsh believes.
While it's probably not possible to customize care for each individual from the outset, data analytics can help stratify people into different risk groups, he says. "Right now, a high percentage of the drugs we give to people don't work. Our goals are creating much more precise care and more effective treatments."
Vanderbilt University has used genotyping to alter prescribing patterns in heart patients who are candidates for stents. Since September 2010, more than 1,200 patients have been genotyped, with data stored in their medical records. About 300 have a minor DNA variation that affects the way they metabolize the anticoagulant drug Clopidigrel. That information has triggered several dozen "decision support alerts" that resulted in physicians changing prescriptions for these patients, Masys says.
"This presents an entirely new source of information that will assist decision-making."
Game Theory in P4 Medicine
Can Facebook friends build a better antibody? In the growing world of social media, it can.
In a study published in the journal Nature, 57,000 online players of the game FoldIt proved better than a sophisticated computer at folding proteins. The object of the "game," developed by a University of Washington biochemist, was to determine how a linear chain of amino acids curls into a three-dimensional shape to arrive at an optimum energy state. The 57,000 players, who competed and collaborated with one another to solve the puzzle, were listed as co-authors.
"FoldIt is one example of what we're trying to realize with the fourth 'P' of P4 Medicine – participation," says Clay Marsh, MD, executive director of the Center for Personalized Health Care at Ohio State and vice dean for Research at The Ohio State University College of Medicine. Marsh sees a wide use of such crowd-sourcing in the future to solve complex problems, scientific and otherwise.
The FoldIt players were so successful that the game's creator, David Baker, PhD, has given the players opportunities to create entirely new proteins. One puzzle asked players to create a more stable version of a protein involved in creating antibody-like compounds.
FoldIt is one of several endeavors in the emerging field of distributed thinking.
The network Social Chocolate also uses games to solve problems and improve lives. Founder Jane McGonigal created the game SuperBetter to help heal herself after a traumatic brain injury in 2009. It worked, and since then she has allowed people all over the world to create their own versions of the game to help them recover from other ailments and promote wellness. They can invite friends and relatives to play with them as a modern-day support network.
Ohio State became one of the first academic institu- tions to link with Social Chocolate, as researchers launched clinical trials of SuperBetter in September to determine ideal versions for healthcare providers to play with their patients and for patients to play at home with their families.
In addition to solving vexing problems, gaming and social networking "create a connection, which is critical in health care," Marsh says. Facebook apps designed to help people stop smoking "provide a worldwide web of support."
A paper written by Marsh and five other experts in the field of P4 Medicine, which is currently awaiting publication, says research has shown that people with richer social networks are generally healthier and have better outcomes when facing a variety of health interventions.
They write that games activate brain centers that allow participants to get into "flow," a mental state in which a person is fully immersed in energized focus. The positive feelings generated during gaming have a positive effect on the person psychologically.
Marsh believes social media "communities" can have a positive influence on wellness. "The question is how to engage people to be interested in their health."
As physicians, Marsh says, "We give people a lot of information, but we don't make it easy for them to translate that information to their lives. We have to reframe the discussion." He cited iTunes Genius, which recognizes user preferences, as an example of ways to engage people.
"So, in the case of medicine, it could be something like, 'If you liked the Komen Walk, maybe you'd like Pelotonia, too,' " Marsh says, referring to the cycling fundraiser held each year between Columbus and Athens, Ohio.
- For P4 Medicine to reach its full potential, everything in health care must adapt, from the patient-physician relationship to insurance coverage, medical education, the design of medical facilities and the very treatment plans developed for each patient.
- Ohio State's Medical Center and a few other academic hospital systems in the country are reorganizing in ways that are leaps beyond the guidelines set in federal healthcare legislation. These pioneers are entirely re-envisioning health and health care in America.
- The ability to ascertain the likelihood of disease far in advance of a diagnosis means that treatments will more frequently be preventive rather than medical – more about lifestyle changes than prescriptions and surgeries. This places more responsibility on the individual to manage his or her own health.
From Left: Edward Abrahams, PhD; William Dalton, MD, PhD; Peter Embi, MD; Daniel Masys, PhD; Philip Payne, PhD; Ann Pendelton-Jullian, PhD