Identifying Patient Readmissions: Are Our Data Sources Misleading?

Mi Puente (My Bridge) Care Transitions Program for Hispanic/Latino Adults with Multimorbidity: Results of a Randomized Controlled Trial

BACKGROUND

Multimorbidity frequently co-occurs with behavioral health concerns and leads to increased healthcare costs and reduced quality and quantity of life. Unplanned readmissions are a primary driver of high healthcare costs.

OBJECTIVE

We tested the effectiveness of a culturally appropriate care transitions program for Latino adults with multiple cardiometabolic conditions and behavioral health concerns in reducing hospital utilization and improving patient-reported outcomes.

DESIGN

Randomized, controlled, single-blind parallel-groups.

PARTICIPANTS

Hispanic/Latino adults (N=536; 75% of those screened and eligible; M=62.3 years (SD=13.9); 48% women; 73% born in Mexico) with multiple chronic cardiometabolic conditions and at least one behavioral health concern (e.g., depression symptoms, alcohol misuse) hospitalized at a hospital that serves a large, mostly Hispanic/Latino, low-income population.

INTERVENTIONS

Usual care (UC) involved best-practice discharge processes (e.g., discharge instructions, assistance with appointments). Mi Puente ("My Bridge"; MP) was a culturally appropriate program of UC plus inpatient and telephone encounters with a behavioral health nurse and community mentor team who addressed participants' social, medical, and behavioral health needs.

MAIN MEASURES

The primary outcome was 30- and 180-day readmissions (inpatient, emergency, and observation visits). Patient-reported outcomes (quality of life, patient activation) and healthcare use were also examined.

KEY RESULTS

In intention-to-treat models, the MP group evidenced a higher rate of recurrent hospitalization (15.9%) versus UC (9.4%) (OR=1.91 (95% CI 1.09, 3.33)), and a greater number of recurrent hospitalizations (M=0.20 (SD=0.49) MP versus 0.12 (SD=0.45) UC; P=0.02) at 30 days. Similar trends were observed at 180 days. Both groups showed improved patient-reported outcomes, with no advantage in the Mi Puente group. Results were similar in per protocol analyses.

CONCLUSIONS

In this at-risk population, the MP group experienced increased hospital utilization and did not demonstrate an advantage in improved patient-reported outcomes, relative to UC. Possible reasons for these unexpected findings are discussed.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT02723019. Registered on 30 March 2016.

Disease Management in Skilled Nursing Facilities Improves Outcomes for Patients With a Primary Diagnosis of Heart Failure

OBJECTIVE

Skilled nursing facilities (SNFs) are common destinations after hospitalization for patients with heart failure (HF). Our objective was to determine if patients in SNFs with a primary hospital discharge diagnosis of HF benefit from an HF disease management program (HF-DMP).

DESIGN

This is a subgroup analysis of multisite, physician and practice blocked, cluster-randomized controlled trial of HF-DMP vs usual care for patients in SNF with an HF diagnosis. The HF-DMP standardized SNF HF care using HF practice guidelines and performance measures and was delivered by an HF nurse advocate.

SETTING AND PARTICIPANTS

Patients with a primary hospital discharge diagnosis of HF discharged to SNF.

METHODS

Composite outcome of all-cause hospitalization, emergency department visits, and mortality were evaluated at 30 and 60 days post SNF admission. Linear mixed models accounted for patient clustering at the physician level.

RESULTS

Of 671 individuals enrolled in the main study, 125 had a primary hospital discharge diagnosis of HF (50 HF-DMP; 75 usual care). Mean age was 79 ± 10 years, 53% women, and mean ejection fraction 46% ± 15%. At 60 days post SNF admission, the rate of the composite outcome was lower in the HF-DMP group (30%) compared with usual care (52%) (P = .02). The rate of the composite outcome at 30 days for the HF-DMP group was 18% vs 31% in the usual care group (P = .11).

CONCLUSIONS AND IMPLICATIONS

Patients with a primary hospital discharge diagnosis of HF who received HF-DMP while cared for in an SNF had lower rates of the composite outcome at 60 days. Standardized HF management during SNF stays may be important for patients with a primary discharge diagnosis of HF.

A Randomized Controlled Trial of Heart Failure Disease Management vs Usual Care in Skilled Nursing Facilities

OBJECTIVE

Patients discharged from the hospital to a skilled nursing facility (SNF) are not typically part of a heart failure disease management program (HF-DMP). The objective of this study is to determine if an HF-DMP in SNF improves outcomes for patients with HF.

DESIGN

Cluster-randomized controlled trial.

PARTICIPANTS

The trial was conducted in 47 SNFs, and 671 patients were enrolled (329 HF-DMP; 342 to usual care).

METHODS

The HF-DMP included documentation of ejection fraction, symptoms, weights, diet, medication optimization, education, and 7-day visit post SNF discharge. The composite outcome was all-cause hospitalization, emergency department visits, or mortality at 60 days. Secondary outcomes included the composite endpoint at 30 days, change in the Kansas City Cardiomyopathy Questionnaire and the Self-care of HF Index at 60 days. Rehospitalization and mortality rates were calculated as an exploratory outcome.

RESULTS

Mean age of the patients was 79 ± 10 years, 58% were women, and the mean ejection fraction was 51% ± 16%. At 30 and 60 days post SNF admission, the composite endpoint was not significant between DMP (29%) and usual care (32%) at 30 days and 60 days (43% vs 47%, respectively). The Kansas City Cardiomyopathy Questionnaire significantly improved in the HF-DMP vs usual care for the Physical Limitation (11.3 ± 2.9 vs 20.8 ± 3.6; P = .039) and Social Limitation subscales (6.0 ± 3.1 vs 17.9 ± 3.8; P = .016). Self-care of HF Index was not significant. The total number of events (composite endpoint) totaled 517 (231 in HF-DMP and 286 in usual care). Differences in the 60-day hospitalization rate [mean HF-DMP rate 0.43 (SE 0.03) vs usual care 0.54 (SE 0.05), P = .04] and mortality rate (HF-DMP 5.2% vs usual care 10.8%, P < .001) were significant.

CONCLUSIONS AND IMPLICATIONS

The composite endpoint was high for patients with HF in SNF regardless of group. Rehospitalization and mortality rates were reduced by the HF-DMP. HF-DMPs in SNFs may be beneficial to the outcomes of patients with HF. SNFs should consider structured HF-DMPs for their patients.