Use of Smartphone Technology and Machine Learning to Improve Access to Care

Friday, May 13, 2016: 2:09 PM
Room 308 (Baltimore Convention Center)
C. Lajonchere1,2 and B. Vaughan1, (1)Cognoa, Inc., Palo Alto, CA, (2)Biomedical Engineering, University of Southern California, Los Angeles, CA
Background: The use of cell phones and other patient-centered technology is making an impact on the way people access and engage in their healthcare. Increased use of mobile platforms for healthcare delivery can also be effective in providing improved access for low income and minority populations who face greater healthcare disparities.  This is particularly relevant to early screening of young children who may wait as long as 18 months to receive an early diagnosis despite the early warning signs of a developmental delay. Cognoa’s evidence-based mobile tools leverage the use of big data and machine learning to give parents and providers an estimate of a child’s risk of a developmental delay based on information that parents collect from the privacy of their homes.  Providing clinicians with a child’s results using a web-based user interface can help parents and providers flag concerning behaviors sooner and can help high-volume clinics with screening and triage.  

Objectives: To determine whether Cognoa’s mobile evidence-based risk assessment could be integrated into a high volume developmental clinic and evaluate parent and clinician satisfaction with the Cognoa interface and video-based platform.  

Methods: In preparation for a multi-site clinical study, the Center surveyed over 400 families on their wait list to determine smartphone or tablet use.  Of those that owned a smartphone/tablet, 356 were invited to complete the Cognoa flow prior to their clinic visit.  Completion rates for Cognoa were compared to the standard clinic completion rate and qualitative ratings of parent and clinician satisfaction were obtained.  

Results: Ninety-six percent (96%) of respondents surveyed owned a smartphone or tablet. Fifty-five percent (55%) had android phones, 35% iOS, and the remainder used a different kind of smartphone. The average completion rate was 50%, which was similar to the Center’s average completion rate (p>.05). The majority of respondents (72%) were identified as having elevated risk of a developmental delay on the Cognoa questionnaire.  Seventy-eight percent (78%) of those who received elevated risk on the questionnaire completed the video evaluation. Ninety-four percent (94%) of parents surveyed indicated that they were very happy with the experience and personalized results.  Clinicians and staff found the clinician dashboard easy to use and a select number of high risk children were fast tracked by Center staff upon review of the video assessment.  Sensitivity of the tool in this small pilot was .85.

Conclusions: Results demonstrate that the use of smartphone-based screening can be easily integrated in a clinic population. The use of mobile technology to engage parents is extremely promising and highlights the potential to develop high quality, rapid approaches to improve early detection and reach a more diverse population.