A heart failure patient was discharged with new medications that included a beta-blocker and a diuretic. But unbeknownst to her clinicians, she had previously been prescribed the same medications and was now doubling up at home, said Neeraj Tayal, MD, who directs the division of general internal medicine and geriatrics at Ohio State University Wexner Medical Center in Columbus.
No harm resulted, as the pharmacist on the medical center's transitional care team called the patient a day after she was discharged, Dr. Tayal said. “She [the pharmacist] walked into my office yesterday with her head down and said, ‘Now I know why we make transition calls.’”
Clinicians at OSU Wexner have been making these discharge calls to high-risk patients since 2013, a practice that they have been gradually expanding. Now the medical center's goal, as part of its participation in Medicare's Comprehensive Primary Care Plus initiative, is to call 75% of patients—ideally those at medium to high risk of readmission—within two business days after discharge.
Experienced physicians frequently develop a clinical sense regarding which patients are more vulnerable to unnecessary hospitalizations, including 30-day readmissions. But the ongoing challenge for clinicians at OSU Wexner and elsewhere is how best to systematically flag those complex patients in advance and then better support them, all amid the swirl of a busy practice. Medicare built some financial support into the payment system in 2013, adding a transitional care management (TCM) code that physicians, nurse practitioners, and other clinicians can bill for telephone calls and other non-office visit care during the first month after discharge.
A recent observational analysis, looking at nearly 19 million Medicare claims for the first three years that TCM reimbursement was available, was encouraging in terms of cost-effectiveness. When patients got TCM services, their costs were on average lower for a subsequent 30-day period, $3,033 versus $3,358 for those who lacked the support, according to findings published in September in JAMA Internal Medicine. Mortality rates during that same stretch also were lower in the TCM group, with 1% of patients dying versus 1.6% of patients who got usual care.
“We're seeing that this intervention can be quite effective,” said Andrew Bindman, MD, FACP, a study author and professor of medicine, epidemiology, and biostatistics at the University of California, San Francisco. “But it seems to be probably underutilized.” By 2015, according to the same analysis, just 7% of discharge claims involved TCM services compared with 3.1% in 2013.
Comprehensive care gaps
When delving further into TCM claims, Dr. Bindman's analysis found that the vast majority who billed were primary care physicians. They also were concentrated in a small number of medical practices; just 10% of practices accounted for 68.3% of billed TCM services.
“This suggested to us that there's likely to be some restructuring of the workflow in a practice that is necessary to be able to build and provide this service,” Dr. Bindman said. “Once a practice does that, what you see is that practice does provide this service with a fair degree of regularity to its discharged patients.”
While the study didn't look at the reasons for the limited TCM uptake, Dr. Bindman and his coauthor questioned whether the additional payment, roughly $40 above a typical office visit, was sufficient to compensate for the staff time and workflow changes involved.
An accompanying JAMA Internal Medicine editorial asked how feasible it is for outpatient doctors to routinely reach patients within two days after discharge. Just 3% of primary care doctors report being involved in conversations about discharge, according to a 2007 JAMA review article cited by the accompanying editorial. No more than 20% of doctors said they were always notified about a patient's hospital discharge.
Although the review article was published a decade ago, Teryl Nuckols, MD, doesn't believe that the situation is dramatically better today. “My husband is a primary care doctor, and he will first learn that a patient has been hospitalized when they come for their postdischarge visit or, I should say, if they come for a postdischarge visit,” said Dr. Nuckols, an ACP Member and a hospitalist who directs the division of general internal medicine at Cedars-Sinai Medical Center in Los Angeles.
Dr. Nuckols, a coauthor of the JAMA Internal Medicine editorial, also isn't convinced that bumping up the TCM reimbursement a small amount, for example, $15 or so, would significantly boost physician buy-in. The broader problem is that the payment doesn't align with the financial incentives, she said. Hospitals are the ones that have an incentive to reduce 30-day readmissions among the patients they discharge every day. But the typical primary care doctor, who would have to invest in reworking practice processes, likely doesn't see more than a handful of discharged patients in a given week, she said.
“It's not surprising to me that this policy has not been taken up widely,” Dr. Nuckols said. It would be interesting, she added, if future research looked at whether TCM services are used more often in accountable care organizations (ACOs) and other reimbursement models that share financial risk. (Dr. Bindman's analysis did find that nearly one-fourth of those Medicare clinicians billing for TCM also participate in the Medicare Shared Savings ACO.)
While better support of high-risk and complex patients to avert hospitalizations makes sense in theory, it's been more difficult to consistently demonstrate benefits, said Richard W. Grant, MD, a research scientist in the division of research at Kaiser Permanente Northern California. “Unfortunately a lot of the studies have not been able to show dramatic cost savings or dramatic health benefits,” he said. “It's really puzzling.”
While Dr. Bindman's analysis of transitional care management showed favorable results for cost-effectiveness, the results of a recent randomized analysis of primary care intensive management in the Veterans Affairs system were equivocal. The study, published June 19 in Annals of Internal Medicine, randomly assigned 2,210 patients considered at high risk for future hospitalizations either to intensive management or usual care at five VA primary care clinic sites.
The patients were nearly all men with an average age of 63 years and an average of seven chronic conditions. Those assigned to intensive management were offered a range of services, including interdisciplinary team care, pharmaceutical support, and TCM, among others. Yet average total costs between the two groups were similar during the 12 months before and after the patients' randomization, and there was no difference in mortality.
A trend toward declining inpatient costs was seen in the intensive management group, although it wasn't statistically significant, said Jean Yoon, PhD, a health economist at VA Palo Alto Health Care System in California and lead author on the Annals study. Among the intensive management services used, patients randomized to intensive management frequently required social services and mental health services, Dr. Yoon said.
“We found a lot of unmet needs for care that ended up being met by these interdisciplinary care teams,” Dr. Yoon said. “Hopefully, if we were able to put patients on a better trajectory, over the long term they would have better outcomes and use health care more efficiently.”
One reason that it's been so difficult to demonstrate cost-effectiveness is that researchers and clinicians are still trying to figure out which patients are the best candidates for these intensive care interventions, Dr. Grant said. “If you just look at people based on medical diagnoses, you don't really know very much about them,” he said.
Dr. Grant helped to lead a recent study in which care managers at Kaiser Permanente Northern California were asked to review whether patients who had previously been referred to them proved to be good candidates. “They actually had to take care of these patients,” Dr. Grant said. “They could tell us whether it was worth their investment of time and effort.”
Of the 1,178 patient cases that were reviewed by the nurses and social workers, 62% were deemed to have been appropriate patients, according to the findings, published in September in the Journal of General Internal Medicine. The remainder either didn't need care management at all or were deemed to require more help than the program could provide.
Researchers subsequently interviewed the care managers and learned that the patient's level of motivation and social support frequently tipped the balance one way or the other, Dr. Grant said.
“If we're going to be collecting electronic data, why don't we collect a variable about social support and keep that updated,” Dr. Grant said. For instance, he suggested asking patients, “Is there anyone who helps you at home?”
Through the interviews, the researchers also learned that there's a sweet spot in terms of an optimal patient, Dr. Grant said. Care managers, he said, “love finding the son two towns over and getting him engaged in the mom's care.” If patients have a lot of support and self-motivation, they likely don't need help despite multiple medical problems. Conversely, an elderly lady with developing dementia who has outlived her family is difficult to assist, he said.
During routine office visits, primary care practices can begin to build a more nuanced picture of their complex and vulnerable patients, said Brent Williams, MD, an ACP Member and medical director of the Complex Care Management Program at Michigan Medicine. He described five broad domains that can help frame a patient's vulnerability: any substance abuse issues, overall ability to care for themselves (e.g., taking pills, keeping appointments), financial stability and living environment, functioning in work and life, and breadth of their social support.
Nurses or medical assistants should be permitted to use their judgment regarding what's worth asking at a particular visit and should not be confined to a set checklist of questions, Dr. Williams said. During each visit, that clinician then can ask one question that addresses one of the domains that has been flagged as a potential issue and save other questions for a later visit, he said.
To gain more insights into a patient's social support, he suggested asking: “What are the meaningful relationships in your life right now?” For a better sense of a patient's daily functioning, ask, “Just walk me through the highlights of your day yesterday.”
Bolstering practice support
An outpatient practice can make some headway with reaching vulnerable patients by asking each clinician to identify three to five patients apiece to better support with telephone calls and other assistance, Dr. Williams said. Above all, it's key that the patient develops an ongoing relationship with one individual in the practice, he said, even if it's a medical assistant who must relay clinical questions to someone else on staff.
Like Dr. Grant, Dr. Williams emphasized the Goldilocks element of identifying the optimal set of complex patients to help. For some patients, unfortunately, their needs are so overwhelming that even investing a lot more time might not necessarily help much, he said.
Practices can begin to boost discharge support by focusing on one or two diagnoses or complex conditions, such as calling all patients newly discharged on an anticoagulant within two business days, said Stuart Beatty, PharmD, a pharmacist who works with OSU Wexner's transitional care management program. Once that workflow process has been established, the practice could then gradually add other categories of patients, he said.
To boost the chance that outpatient doctors are kept in the discharge loop, perhaps they can educate patients to call the physician's office if they are hospitalized, Dr. Nuckols said. Another strategy is for primary care doctors to build stronger ties with local hospitalist groups, and make it clear that discharge communication matters, she said.
“They have some leverage in that they can make that an expectation,” Dr. Nuckols said. “You notify our doctors when our patients get in, or the referrals will dry up.”
Automating discharge referrals
When OSU Wexner launched its transitional care management program, nurses typically made the postdischarge calls. But it quickly became apparent that pharmacists were better suited in many cases, given that so many of the questions or issues involved medications, said Dr. Tayal, who launched and still oversees the program.
In the first several years, the program reached about 10% to 15% of discharged patients, whose outpatient primary care physicians had referred them as high risk, Dr. Tayal said. Typically, those doctors would make the referral to OSU Wexner's pharmacy team, when they felt that patients would benefit from a transitional call. But that approach created a bit of a referral bottleneck, as it relied on overstretched primary care doctors to refer patients in a timely manner, he said.
To achieve the new goal of reaching 75% of discharged patients, the approach was changed in 2017, Dr. Tayal said. Rather than relying on physician referral, the electronic health record automatically categorizes patients by their level of readmission risk at discharge. Generally speaking, the pharmacists call the high-risk patients and the nurses call the moderate-risk patients, Dr. Tayal said.
Dr. Tayal credits OSU Wexner's risk evaluation approach with avoiding the low physician uptake identified by Dr. Bindman's analysis. Plus, he added, it's more systematically consistent.
Otherwise, Dr. Tayal said, “I think you get a lot of disparities. You'll have some doctors that will just send every patient they were informed about [to transitional care], and then other doctors will be sending few to none.”