A decade ago, New England Baptist Hospital (NEBH) changed its top leadership with a new CEO and tightened its focus by winnowing to an orthopedics-only specialty hospital. Now, to keep the hospital on track and high performing, NEBH is counting on analysis of patient-reported outcomes (PROs) and other critical health care data to guide clinical leadership, says Scott Tromanhauser, MD, MBA, MHCDS, Chief Medical Officer and Chief of Spine Surgery.
“We are just starting to see the fruits of our labor,” Tromanhauser says of one recent analytics effort, started a year ago, to gather long-term patient feedback around total hip replacements, total knee replacements, and spine surgeries to gauge outcomes and patient experience. The post-op score assigned to each PRO helps determine how NEBH is doing as an institution as well as how surgeons and their teams are performing. The more PROs collected, the richer the data to be analyzed, he says.
Nearly 60% of respondents to a recent NEJM Catalyst Insights Council survey on the topic of analytics and outcomes rank “guiding clinical leadership” as the top goal of their organization’s analytics efforts. Guiding business leadership ranks second with a score of 53%.
NEBH’s collected PROs also will be in instrumental in deciding which patients shouldn’t have surgery. “One of our busiest total joint replacement surgeons pointed out to leadership that nationally, 20% of people who undergo total knee replacement are unhappy with the outcome,” Tromanhauser says, adding total knee replacement is the number one surgery reimbursed by the Centers for Medicare and Medicaid Services. “What if we could predict who would have a bad outcome, based on past PROs, and offer them a nonoperative approach?”
By coupling the PROs with other key inputs, NEBH aims to enhance shared decision-making with a predictive modeling tool, currently under development with Cyft, a Boston-area software company that uses machine learning and natural language processing as part of its data analysis, that can show providers, patients, and their caregivers the risks of procedures. If the risks are deemed too high or the benefits too low, alternatives can be discussed. “Even the best surgeon would do better with predictive modeling,” Tromanhauser says.
While a majority of survey respondents, 69%, pinpoint the goal of analytics as driving improvement and driving accountability, only half call their organization “effective” when it comes to supporting care decisions or individual patients — the kind of care Tromanhauser has described.
Ankita Sagar, MD, MPH, FACP, Director for Ambulatory Quality at Northwell Health’s Department of Internal Medicine in New York, says that along with guiding clinical leadership, she sees tremendous opportunity for analytics to guide population health efforts, which Council members rank third in effective use of health care data. “We are trying to invest resources and effort in understanding how we can leverage analytics for patients at risk for a certain illness and optimize their health care for their risk factors,” she says.
For instance, in the NYC Tristate area, Northwell Health serves a vast population at risk for falls, heart disease, and cancers. In years past, primary care physicians seeing patients who had been to multiple other physicians had limited visibility into their medical history. Now, through a clinical dashboard, physicians have access to metrics that gauge a patient’s risk of falls, heart disease, and completed/overdue cancer screenings. They can document and share this information system-wide. For example, if Sagar sees an elderly patient who reports two falls since his last visit, but no falls had been documented, “that’s a significant change from baseline.” That simple act of transparency into the patient’s complete risk assessment, Sagar says, allows providers to take appropriate action for the patient at the time of care and has improved health care for patients.
To Gary Wainer, MD, DO, Director of Quality at Northwestern Medicine in Chicago, analytics has to reach across institutions to easily fill in the gaps in a patient’s history for true success in population health and for providers to be able to learn from the data. (Only 28% of Insights Council members say their organization is effective at learning from health care data.)
Northwestern Medicine has a single electronic medical record system, but identifying gaps in care, which is a key part of population health, is difficult without a comprehensive record that spans health systems as well as other organizations, such as retail clinics, that administer treatment and vaccines. “When patients see doctors outside the system, it is not always an easy process to obtain that information,” Wainer says. “Even in circumstances where data is available, the data allowed to be shared varies by organization. We may or may not be able to find a blood sugar done elsewhere; we just know that the patient said they had it done. We don’t see it in our record, and we need to deliver it.”