Funding Source
National Library of Medicine (NLM), R01LM012815
Project Period
9/7/2017 - 8/31/2022
Principal Investigator
Paul J. Barr, PhD (Geisel School of Medicine at Dartmouth College); Saeed Hassanpour, PhD (Geisel School of Medicine at Dartmouth College) (Multiple PIs)
Other Project Staff
William Haslett, PhD; Craig Ganoe, MS; Jesse Schoonmaker, MD, MPH; W. Moraa Onsando, MD, MPH; Kyra Bonasia, PhD; James Finora; Lisa Oh; Martha Bruce, PhD; Glyn Elwyn, MBBCh, PhD; James O’Malley, PhD; Sheri Piper (Patient Partner); Roger Arend (Patient Partner); Amar Das, MD, PhD;
Project Summary
Forty to eighty percent of clinic visit information is forgotten by patients immediately post visit, a significant barrier to self-management leading to poor health outcomes. Visit summaries can improve recall, yet patient uptake is limited and clinicians report significant burden in creating summaries for laypeople. Patients are beginning to audio record visits and clinics are now offering this service. When patients receive a clinic recording, 71% listen and 68% share it with a caregiver, resulting in improved understanding and self- management. Yet, unstructured recordings are difficult to navigate. Personal health libraries (PHLs) may help patients organize health information; yet current PHLs do not facilitate clinic-recordings. The objective of this project is to develop a PHL that integrates clinic audio-recordings (Audio-PHL), using data science methods to link medical terms from the recording to trustworthy patient resources, which can be retrieved, organized, edited and shared by patients. The specific aims are: Aim 1 Identify health information seeking needs and strategies of older adults with multimorbidity and caregivers; Aim 2 Develop an Audio-PHL using data science methods to securely analyze clinic visit recordings and make this information accessible and understandable for patients; and Aim 3 Demonstrate the usability and use of an Audio-PHL in older adults with multimorbidity and caregivers. Applicants hypothesize: (1) The Audio-PHL will surpass acceptable usability metrics in older adults and caregivers and (2) natural language processing (NLP) methods developed for the Audio-PHL will accurately identify key visit information (e.g. medication) and connect it to credible patient resources. The development of the Audio-PHL follows a user centered design model. In Aim 1, the applicants will use participatory design activities with 48 end-users to inform Audio-PHL design. In Aim 2, the Audio-PHL will be created in iterative cycles informed by findings from Aim 1. In Aim 3, extensive usability evaluation will be conducted in human computer interaction (HCI) laboratory settings to ensure Audio-PHL surpasses acceptable usability metrics. Field testing of the Audio-PHL will follow via a patient-randomized pilot trial with older adults with multimorbidity from primary care. Participants (N=70) will receive an Audio-PHL (intervention) or PHL (control) with no recordings. Usability metrics and satisfaction will be assessed at one-month. Preliminary data on the impact of an Audio-PHL on patient ability to seek, find and use health information with high confidence, patient activation and caregiver confidence will also be gathered. The research is innovative because it will provide patients and caregivers secure access to a PHL based on clinic-recordings that uses data science methods to organize visit information and connect it to trusted resources. The results are expected to have a major positive impact because they will provide proof-of-principle for the use of an Audio-PHL that utilizes the benefits of clinic recordings through the novel application of data science methods, to improve health outcomes for older adults with multimorbidity through greater knowledge and confidence in their ability to self-manage.
Public Health Relevance
Between 40 – 80% of clinic visit information is forgotten immediately by patients, a significant impediment to self-management for older adults with multimorbidity that leads to poor health outcomes. The proposed research will integrate audio-recordings of clinic visits into a Personal Health Library (Audio-PHL), using data science methods to link medical terms from the recording to trustworthy patient resources, which can be retrieved, organized, edited and shared by patients. It is expected that the Audio-PHL will be easy to use and highly utilized, making patients and caregivers more knowledgeable and confident of their health care needs, resulting in greater self-management capabilities.