In our information-infused age, health care organizations can use big data to improve quality of care and outcomes for patients, especially those with chronic conditions. Equally important, they can harness the power of big data to prevent diseases and improve wellness initiatives.
Growing use of electronic health records plays a big role in building up health care data sets. As of last year, nearly six in 10 hospitals had adopted at least a basic EHR system, according to the Office of the National Coordinator for Health Information Technology, and 93 percent had a certified EHR system in place. The National Ambulatory Medical Care Survey also showed that in 2013, close to 80 percent of office-based physicians used some type of EHR system. These systems play host to structured and unstructured data — including a mountain of physician notes — for millions of patients across the country, all of it ripe for use in analytics to help improve and personalize treatments as well as in predictive modeling to better understand where risks lie.
Doing the Right Thing
A recent conversation I had with Blackford Middleton, M.D., chief informatics officer at Vanderbilt University Medical Center, made it plain that leading-edge institutions want to build clinical decision support systems that exploit big data — an organization's high-volume, high-variety, high-velocity information assets. The goal, as the good doctor put it: "to ensure the right things happen for the right patient at the right time."
Vanderbilt University Medical Center is taking on some big data–infused efforts to improve the quality of care it renders under the new value-based health care services model. At the same time, it wants to use big data to reduce unnecessary services.
"We don't need to do thousands of hospital and ambulatory care tests that don't impact care outcomes. Do more of the right ones and fewer of the wrong ones and you will lower costs and increase the value of care provided," Middleton says. Those same results can be achieved when health care systems analyze big data to help a patient avoid a single emergency clinical visit, for example, or an inappropriate medication that has adverse effects.
A big part of getting to this end, though, is moving to a continuous care model that extends the doctor-patient connection beyond scheduled visits. Good continuous care is dependent on more regular infusions of patient data than the updates generated during visits — it requires engaging the patient to be a more active participant in his or her own health care.
It also means investing in communications and other technologies so patients can respond more easily to between-appointment recommendations such as changes in treatment plans. And, it requires coming up with solutions to ensure that the data obtained can be put to use effectively. Such extensions of the connection between the health care team and the patient should translate into successful outcomes and mitigate excessive care costs.
The Road to Better Health Care
Here are five ways your health care organization can work with patients, and with their data, to build a better care model:
Make the patient a responsible member of the health care team. In some respects, it's easier than ever to do so, given the range of smartphone apps, fitness wristbands, smart watches and other tools to record exercise, food intake, weight and blood pressure and transmit it to health care providers. Maybe, suggests Middleton, health care providers should take a closer look at ways to use these tools to make such data contributions into a gamelike experience "to maximize the health care productivity function for each patient."
For example, perhaps patients with high blood pressure conditions can accumulate points for every reading they transmit, and once they reach a certain number, exchange them for a coupon for a healthy smoothie at their next doctor visit. Understanding how various measures consistently trend over time, after all, can be important for doctors treating patients with chronic conditions such as diabetes, hypertension, obesity and even depression.
Think ahead to a holistic approach to data acquisition and analysis. In addition to driving a continuous care model with the help of patients' self-reported information on structured data points, other data — perhaps in unstructured format — can be included in the patient-care analytics mix. As Middleton points out, social, cultural and environmental factors play a big part in people's health and can provide important context for health care management. Commentary on social networks, employee wellness program information and community health statistics — all are venues that might provide useful data related to these parts of patients' lives.
As Middleton says, "There are a lot of issues to be wrestled with," such as getting patients to authorize access by providers to their private information. "But, I think it will come. I think patients will be willing to trade off what's viewed as privacy against what they may view as bringing convenience or value to them."
There is an opportunity, as well, to combine a better clinical understanding of patients with an already robust financial understanding of those same individuals, based on the data stored within health care organizations' administrative systems. Incorporating patient receivables data into health care analytics processes, for example, can help a health care system to understand if it has been successful at effectively engaging patients in their own health care.
Remove the noise. While there is value to be gained from adopting electronic medical records, health care organizations also find that the current generation often floods them with too much undifferentiated information. It's a work in progress for many hospitals to synthesize and distill from that flood the essential signals that pertain to the care of a particular condition for a particular patient. And, as data come in from more sources, perhaps in different formats, the synthesis and distilling challenges are poised to grow.
Vanderbilt is taking steps to meet aspects of the challenge of highlighting important EHR data with projects like Pharmacogenomic Resource for Enhanced Decisions in Care and Treatment, or PREDICT. PREDICT works so that automatic point-of-care decision support is launched when a certain drug is prescribed for individuals whose EHR data indicate that they are at risk for variant genotypes. "Personalization becomes very real," Middleton says.
Let the data inform task automation. As more information becomes part of the health care equation, health care professionals can spend more time doing what they love: engaging in the hands-on care part of the equation while letting data analytics drive task completion.
For example, it is possible to have data from EHRs, sensors and wearable medical devices feed into the health care provider's patient engagement system and trigger just-in-time patient engagement communications specific to that individual. That way, a health care worker doesn't have to, for example, manually monitor incoming data to direct a response to a spike in a patient's glucose levels. Rather, business rules can be created to assess the level of urgency and take the appropriate next step, such as scheduling a screening appointment if the rise in glucose levels was minimal.
Better facilitate feedback from providers to patients.Health care communications infrastructures have unfortunately lagged behind those found in many other sectors. The ability to acquire and analyze more data to drive best-evidence and best-experience recommendations in a continuous care model can't be underestimated.
But the promise of technologies such as wearable devices and EHRs is negated if providers can't deliver their findings — or any information, advice or encouragement — to patients using the communications pathways a patient prefers, whether that's via Web portals or personalized, automated emails, voice mails or text messages. According to the TeleVox Healthy World Research Initiative, for example, close to 50 percent of patients prefer email for communications about patient care between visits, followed by text messaging at 31 percent.
Sadly, when patients disengage, health care organizations risk recurrence of events that might otherwise have been avoided. Says Middleton, "Patients' preferences and utilities impact to a great degree how we see them engage in care or pursue healthy outcomes and behavior."
We know there will be obstacles to overcome in adopting new ways of working with patients and data to turn the vision of continuous care into a reality. But we also know that we can't fail — not if we are to create a more healthful populace without further driving up health care costs.
Scott Zimmerman is a published authority on using technology to engage and activate patients. He is the president of TeleVox Software and spearheads TeleVox's Healthy World initiative, which uses ethnographic research to uncover, understand and interpret patient and provider points of view.