Enabling Data-Driven Healthcare with Pervasive Technology - - PowerPoint PPT Presentation
Enabling Data-Driven Healthcare with Pervasive Technology Temiloluwa Prioleau Assistant Professor of Computer Science Dartmouth College Webinar on Diabetes and Technology Management Center for Digital Health Interventions March 17, 2020
Temiloluwa Prioleau Assistant Professor of Computer Science Dartmouth College Webinar on Diabetes and Technology Management Center for Digital Health Interventions March 17, 2020
Pervasive technologies provide unparalleled opportunities for:
2
Location Blood Glucose Rehabilitation Heart Rhythm Activity
Diabetes
glucose metabolism
State-of-art technology
3
https://jdrf.org.uk/information-support/treatments- technologies/continuous-glucose-monitors/
High blood glucose Low blood glucose
https://jdrf.org.uk/news/diabetes-tech-whats-on-its-way/
Developing Digital Tools for Improved Self-Management of Diabetes
4
User Studies
patient needs
Pervasive Sensing
technology for continuous sensing
Data Analytics
that affect management
to inform treatment
5
Reflection Needs for Personal Health Data
User Studies
Adherence to CGM Use
Data Mining
This Talk Algorithms for Blood Glucose Prediction Sensing More than Physiology
Machine Learning
Behavioral Factors and Glucose Trends
Data Mining
Other Work
6
Reflection Needs for Personal Health Data
User Studies
Adherence to CGM Use
Data Mining
Data in Diabetes”, Int. Conf. on Pervasive Computing Technologies for Healthcare, May 2020. (forthcoming)
Continuous Glucose Monitor (CGM)
Limitation
7
Fig: Conventional visual is daily overlay plot [https://www.dexcom.com/clarity]
Participants:
8
Needs Assessment (Survey) Develop Visualization (PixelGrid_v1) Concept Validation (Phase 1) Improve Visualization (PixelGrid_v2) User Study (Phase 2)
Example Question:
1. What do you wish CGM reports included that could help and/or speed up analysis and interpretation of diabetes management?
Top responses:
and suggest specific behavior changes
carbohydrate intake
9
10
1-Week CGM Assessment
Low Wear Time (< 50%) Very High BG (> 250 mg/dL) High BG (180 - 250 mg/dL) In Range BG (70 - 180 mg/dL) Low BG (54 - 70 mg/dL) Very Low BG (< 54 mg/dL)
PixelGrid: m x n matrix-based plot
Sun Multi-Week Summary q Poor TIR = 26% q Moderate TIR = 50% q Good TIR = 24%
Insufficient Data Poor TIR (< 50%) Moderate TIR (50 - 70%) Good TIR (> 70%)
* TIR means Time-in-Range
Example Question:
11
Outcome: Feedback will be used to inform future work on insight- suggestive visualization of long-term diabetes data.
S18 (physician)
I do like how I can visualize
screenshot.
S12 (patient)
If there is no detectable patterns, you [the end- user] will just see random colors which can be frustrating to a patient.
Example responses:
events and support personalized goal setting
& smartphones) to contextualize adverse glycemic events
12
13
Reflection Needs for Personal Health Data
User Studies
Adherence to CGM Use
Data Mining
10/10/19
Management,” Int. Conf. on Pervasive Computing Technologies for Healthcare, May 2020. (forthcoming)
14
Background
hemoglobin A1C [American Diabetes Association] Problem
some patients with diabetes
CGM Use
https://hcp.eversensediabetes.com/why-eversense-cgm
Research Objective
affect wearing behavior of CGMs in diabetes management
Data-driven analysis of patient data (n=44)
categories
15
Data gap is a period of contiguous missing data in CGM recording Assumption for analysis:
blood glucose) always exists when wearing device is worn.
16
Fig: conventional daily overlay plot showing 1-week data
Data Gap Characterization:
1. Subjects with poorly controlled diabetes (A1C > 7) had on average longer data gaps (i.e. worse adherence to CGMs) than subjects with well- controlled diabetes (A1C < 7) 2. Longer duration of data gaps occurred in suboptimal BG categories and the longest data gaps occurred in the most severe categories (i.e. very low, very high)
17
Implement data-driven adherence analysis in embedded algorithms to better understand context associated with CGM nonadherence
18
Increased CGM Use Improved Glycemic Outcomes
19
Algorithms for Blood Glucose Prediction Sensing More than Physiology Machine Learning Behavioral Factors and Glucose Trends Data Mining
Other Work
Reflection Needs for Personal Health Data
User Studies
Adherence to CGM Use
Data Mining
This Talk
Diabetes Management”, Under Review, 2020.
Prediction," Under Review, 2020.
Contact: Temiloluwa Prioleau; E. tprioleau@dartmouth.edu; W: www.t-prioleau.com 20
Endocrine Team - Dartmouth Hitchcock Medical Center (DHMC) Research Assistants Support
https://www.tidepool.org/