Reducing HF Readmissions: Integrating Automated Technology into Post-Discharge Follow Up

Background To reduce 30-day HF readmissions and provide transitional support after hospital discharge, Intermountain Healthcare (IH) partnered with CipherHealth (CH) to create an automated HF follow up call program (HFCalls). We sought to determine if HFCalls coupled with nursing escalation algorithms improved readmissions, to identify patient themes requiring nurse interventions, and to determine if this approach was acceptable to patients. Methods Beginning November 2016, IH conducted a staged roll-out of HFCalls after a HF hospitalization. Three calls were scheduled at 48 hours, 7 - and 21 days post-discharge. HFCalls were used for patients identified by IHu0027s automated HF Identification and Risk Report and were designed to identify readmission risk factors by focusing on patient education specific to medications, activity, weight, diet, and symptoms (MAWDS). Patients who indicated issues to follow-up questions were routed automatically to nurses with IHu0027s “Health Answers” team, who used internally developed and vetted escalation algorithms for prompt resolution. A multidisciplinary team reviewed the escalation process and documentation of patient issues and actions on a biweekly basis. The program was evaluated over a 12-month period. Results A pilot period of 6 months, led to refinement of the process prior to a staged rollout with full implementation and standardized workflow across the IH system by November 2017. Of the patients (n=1,167) who received HFCalls during January 2017 - December 2017, 79% (n=922) answered the call and 57.2% (n=668) engaged by answering the first question. The readmission rate among those who engaged was 9.3% (n=62), compared to 13.8% (n=69) among the unengaged (p = 0.01). Themes of escalated calls varied by the 3 timed calls: call 1- issues with symptoms, medications, weighing and follow up; call 2- low salt diet/fluid restriction and, follow up; call 3- issues with activity and medications. Overall, 66% of engaged patients were satisfied with HFCalls. Conclusion Integrating automated HFCalls with nursing escalation algorithms after HF hospitalizations was associated with a lower hospital readmission rate in a manner acceptable to patients. These results show the importance of access to nursing support after discharge and reveal how patient themes may be used to guide HF patient education. The HFCalls technology may be applicable to other chronic disease states, but this requires further study.