Evaluating Proactive Care Coordination Among Medicare Beneficiaries

Key Takeaways

• In a study involving Medicare beneficiaries, proactive care coordination did not reduce hospitalization events.
• Acceptability was lower with proactive outreach, and many participants who declined said they were already coordinating care themselves.

Among older Medicare beneficiaries with cardiovascular disease or risk factors and highly fragmented ambulatory care, proactive outreach for care coordination before hospitalization did not lower emergency department visits or hospitalizations compared with usual care management after hospitalization or physician referral.

A randomized clinical trial in a New York accountable care organization randomized 400 participants, including 202 to the intervention and 198 to control. Eligible participants were community-dwelling Medicare beneficiaries aged 65 years or older who had cardiovascular disease or at least 1 cardiovascular risk factor. They also needed highly fragmented ambulatory care, defined by a reversed Bice-Boxerman Index of 0.85 or higher, and at least 4 ambulatory visits in the attribution year. Investigators used a parallel 1:1 design in a type 1 hybrid effectiveness-implementation trial with pragmatic features. Participants had a mean age of 75.8 years, 71.8% were female, and their median prior use was 14 ambulatory visits to 8 physicians.

In the intervention arm, care managers used proactive telephone outreach, asked about perceived care coordination, and immediately offered care management when responses suggested a problem. The comparator was usual care, with care management triggered after hospitalization or physician referral. The primary outcome was the combined rate of emergency department visits or hospitalizations during follow-up, with 104 ED visits, 62 hospitalizations, and 166 total events recorded.

13 of 49 eligible intervention participants accepted care management, or 26.5%, compared with 17 of 17 eligible control participants, or 100%. The most common reason for declining in the intervention arm was that participants said they were coordinating their own care. Acceptability was lower in the intervention group, while appropriateness and fidelity were 100% in both groups.

No adverse events were expected or observed during the trial. Median follow-up was longer in the intervention group than in control, at 205 vs 143 days, and loss to follow-up was higher, at 28 vs 15 participants. The investigators noted that the single accountable care organization setting in New York City may limit generalizability. Fewer than half of participants were racial or ethnic minorities or had low income levels.