Most patients offered proactive care coordination services declined — including the majority who had already reported concerns about their own care coordination — choosing instead to manage communication among their own physicians, according to a randomized clinical trial published in JAMA Network Open.
Proactive outreach offering care coordination prior to hospitalization did not reduce emergency department (ED) visits or hospitalizations compared with usual care among older Medicare patients with cardiovascular disease (CVD) or CVD risk factors and highly fragmented ambulatory care. Under usual care, patients became eligible for care management following hospitalization or physician referral.
The trial included 400 Medicare beneficiaries aged 65 years or older attributed to NewYork Quality Care (NYQC), a Medicare Shared Savings accountable care organization (ACO) in New York City that brings together NewYork-Presbyterian Hospital, ColumbiaDoctors, and the Weill Cornell Physician Organization. Patients had CVD or at least one CVD risk factor, at least four ambulatory visits in the prior year, and highly fragmented care, defined by a reversed Bice-Boxerman Index of 0.85 or higher.
Lisa M. Kern, MD, of Weill Cornell Medicine, and researchers compared two strategies for allocating patients to care coordination. In the intervention group, care managers proactively attempted to contact patients in the community, administered a telephone survey about perceived gaps in care coordination, and offered care management to those who reported concerns.
The primary outcome was a composite of all-cause ED visits or all-cause hospitalizations during follow-up.
The researchers found no statistically significant difference between groups. The rate of ED visits or hospitalizations was 0.25 events per 100 person-days alive in the proactive outreach group and 0.21 events per 100 person-days alive in the usual-care group.
Acceptability was markedly lower in the intervention group. Of 201 patients targeted for outreach, 148 were reached, 96 completed the survey, and 45 reported concerns about care coordination. Of those 45 patients, nine accepted care management. Including four additional patients who accepted care management following hospitalization or physician referral, 13 of 49 eligible patients (27%) in the intervention group accepted services, compared with 17 of 17 (100%) in the usual-care group.
Among patients in the intervention group who reported concerns but declined care management, the most common reason was that they were already coordinating communication among physicians themselves.
The researchers also conducted post hoc subgroup analyses. Among intervention-group patients eligible for care management, those who declined services had lower acute care event rates than those who accepted services. However, that finding should not be interpreted as evidence that declining care management reduced acute care use, because patients who accepted services may have had greater clinical or care-coordination needs. Event rates were similar when the analysis was limited to patients who accepted care management in either trial group.
The trial had several limitations. It was conducted within a single New York City ACO, which may limit generalizability. The sample was mostly White and female, and patients with dementia were excluded. Follow-up was longer in the intervention group than in the usual-care group, and loss to follow-up was also higher in the intervention group. The researchers reported that findings were similar in a time-to-event sensitivity analysis.
In an invited commentary, Evelyn T. Chang, MD, of the Veterans Health Administration Greater Los Angeles Healthcare System, described the findings as consistent with prior trials of care coordination and case management that have shown mixed or limited effects on acute care use. Dr. Chang noted that the study was notable because it tested one component of care coordination within an ACO using a randomized design — a rigorous approach rarely used in ACO research.
Dr. Chang noted that primary care clinicians face significant barriers to coordinating care across health systems, including delays in obtaining outside records, unintuitive health information exchange platforms, and limited reimbursement incentives in productivity-driven environments — burdens that have increasingly shifted toward patients and caregivers in practice. Dr. Chang suggested that health systems may need to move toward more targeted or stepped care-coordination models, reserving higher-intensity services for patients with factors that may limit their ability to coordinate care independently, such as health illiteracy, cognitive impairment, lack of caregiver support, active mental health symptoms, uncontrolled substance use, or difficulty with instrumental activities of daily living.
"Lack of effective communication across physicians is a health system problem," the researchers wrote. The results, they added, raise questions about whether offering care management alone is sufficient when many patients choose to coordinate communication across physicians themselves.
Disclosures can be found in the published research and commentary.
Source: JAMA Network Open
