Outcomes

The primary outcome was 30-day hospital readmission during the study period. Prespecified secondary outcomes were 30-day postdischarge missed outpatient physician appointments and emergency department (ED) visits. Outcomes were determined by the number of participants with 1 or more hospital admissions, ED visits (including ED observation stays but excluding ED visits resulting in hospitalization), or missed appointments during the 30 days after discharge from the index hospitalization. Automated email notifications of Partners Healthcare admissions and ED visits were generated by a hospital readmissions database and sent to study staff within 1 hour of occurrence. Hospital admissions were captured at Partners Healthcare–affiliated hospitals in Massachusetts (3 academic hospitals; 6 community hospitals) and New Hampshire (1 community hospital). Missed appointments were captured similarly by a hospital database and tabulated quarterly. Since the EMR was unable to capture encounters outside Partners Healthcare, participants were asked about clinical encounters during a 30-day postdischarge questionnaire. Owing to the nature of the intervention, patients and CHW staff were unable to be blinded to the study treatment arms. Study staff collecting EMR and patient-reported outcomes and performing data analysis were blinded. Clinical outcomes were adjudicated independently of the trial.