Team:  Sara Murray, Michelle Mourad, Jinoos Yazdany, Maria Otto

Background 

Multiple studies have shown that recording discharge instructions for patients improves their understanding of their disease and improves outcomes such as hospital readmission.  Likewise, in clinic appointments, patients forget as much as 80% of what their doctor tells them.  Many physicians spend a great deal more time explaining things to patients in both the inpatient / outpatient settings, but this is often not captured for patients in a permanent way.  Voice recording offers a unique opportunity to improve patient care, satisfaction, outcomes such as readmission and understanding of complicated directions from outpatient visits.  While it has been tested and implemented at other institutions in the inpatient setting with favorable results, it is not yet standard of care for hospital discharges. In addition, since most care is administered in the ambulatory setting for chronic disease, it has potential to be utilized in that setting with equal benefit to patients and providers.

Proposal 

We propose implementation of technology that enables voice recordings of patient instructions and complicated conversations in both the inpatient and ambulatory settings at UCSF.  We will partner with a company (such as building upon our existing institutional relationship with Cipher using their Echo platform) to trial implementation of technology that has already been developed. This platform will enable patients to have an application on their phone (loaded with the assistance of patient-care assistants or medical assistants), and providers would be able to record their discharge or final clinic conversations for later reference.  Recording conversations would be optional and editable for providers.

These conversations could be integrated into Epic and MyChart.  Long-term, we would like to be able to present the same information in multiple ways to patients, including not only the audio instructions but also speech-to-text translation that they could read within the interface.   This could also potentially replace current patient instructions in the medical record if the provider wishes (to avoid increasing provider work-load).

We propose initially implementing this technology as a RCT that involves several ambulatory clinics and inpatient medicine teams. Providers would be randomized over a 3-6 month period, with half utilizing this method and half using the current standard-of-care (typed patient instructions included in the AVS).  We would then collect feedback and measure qualitative outcomes (patient and provider satisfaction) to be used for iterative improvement of the technology.  We would also measure longitudinal outcomes (readmissions, adherence to treatment and appropriate monitoring) to assess the long-term impact in our patient populations.