UCSF

Background

Hospitals are implementing discharge support programs to reduce readmissions, and these programs have had mixed success.

Objective

To examine whether a peridischarge, nurse-led intervention decreased emergency department (ED) visits or readmissions among ethnically and linguistically diverse older patients admitted to a safety-net hospital.

Design

Randomized, controlled trial using computer-generated randomization with 1:1 allocation, stratified by language. (Clinical Trials.gov: NCT01221532).

Setting

Publicly funded urban hospital in Northern California.

Patients

Hospitalized adults aged 55 years or older with anticipated discharge to the community who spoke English, Spanish, or Chinese (Mandarin or Cantonese).

Intervention

Usual care versus in-hospital, one-on-one, self-management education given by a dedicated language-concordant registered nurse combined with a telephone follow-up after discharge from a nurse practitioner.

Measurements

Staff blinded to the study groups determined ED visits or readmissions to any facility at 30, 90, and 180 days after initial hospital discharge using administrative data from several hospitals.

Results

There were 700 low-income, ethnically and linguistically diverse patients with a mean age of 66.2 years (SD, 9.0). The primary outcome of ED visits or readmissions did not differ between the intervention and usual care groups (hazard ratio, 1.26 [95% CI, 0.89 to 1.78] at 30 days, 1.21 [CI, 0.91 to 1.62] at 90 days, and 1.11 [CI, 0.86 to 1.43] at 180 days).

Limitations

This study was done at a single acute-care hospital. There were fewer outcomes than expected, which may have caused the study to be underpowered.

Conclusion

A nurse-led, in-hospital discharge support intervention did not show a reduction in readmissions or ED visits among diverse, low-income older adults at a safety-net hospital. Although wide CIs preclude firm conclusions, the intervention may have increased ED visits. Alternative readmission prevention strategies should be tested in this population.

Primary funding source

Gordon and Betty Moore Foundation.