Blood Glucose Portal
Xinformatics Blue Team
Overview
• Use Case - sumitra
• Modeling - fred
• Design - hithika
• Back-End - scott
• Front-End - evan
• Demo - evan
• Summary - evan/sumitra
Use Case Outline
Use Case Name : Collaborative Health Care Tracker
Goal : Patient and physician develop plan and track progress
towards healthy living style with the help of a collaborative
health information system
Primary Actors :
• Patient
• Physician
Secondary Actors:
• Physician's record system
• Patient's personal health records
• Nutritionist's record system etc
Basic Flow
1. Doctor uses system to retrieve health records of the patient
and nutritionist's reports
2. Performs tests and inserts results into the record system
3. If results indicate pre-diabetes symptoms (for example),
doctor sets goals for the patients and annotates
recommendations
4. Patient takes periodic measurements of weight and blood
glucose, enters this into his personal record and compares
this with limits imposed by the doctor
5. Patient returns for a check-up after a month having
succeeded in meeting goals encoded in system. Symptoms
no longer persist.
Model
conceptual model • The patient:
o contributes their own data
(PHR)
o contributes to their PHR as well
as the nutrition record
• The doctor:
o updates the EMR
o prescribes a treatment plan
o diagnoses the patient
• The EMR:
o is updated by the doctor
o is the central repository for the
patient's medical information
Model
logical model
• This details the primary as well as the foreign keys and their relationships in
this portal
• EMR is dependent on all other factors in the portal
Design
• Blood glucose information is displayed on a plot with time on
the X-axis and blood glucose in mg/dL on the Y-axis
• The graph is composed of two components - Unhealthy
Areas and Data Point Collection
• The unhealthy areas are colored in a light red to differentiate
them from the ‘normal’ area as defined by the doctor
• Data points have two attributes: a direction attribute and a
goodness attribute
• Direction is determined by the data point’s value relative to
the boundaries
Design (contd)
• The goodness of an attribute is defined by its relationship to
the previous data point and as a function of the direction
• Intuitively, if the data point is closer to the center of the good
blood glucose region, then the data point will turn green
• If it moves in the opposite direction, it turns red and neutral
points and will always appear green
• The averaging feature help smooth out anomalous data
• The use of a simple triangle as a symbol to indicate desired
direction also follows these same principles of making
information easily identifiable, easily readable
• These are some of the areas where information uncertainty,
semiotics, cognition, and architectures were taken into
consideration
Back-End
• Java HTTP servlets running on Tomcat.
• Distributed nature makes traditional databases
more difficult to use and introduce uncertainty in
schema.
• Data kept on disk in NetCDF.
• Transfer between Front-End and Back-End done in
JSON.
NectCDF and Why.
• Standard for array-based scientific data.
• Open Standard.
• Binary.
• Self-Describing and Machine-Independent.
o Reduce Information uncertainty.
o Allow for data to be portable.
• Add annotations to individual data points.
Front-End Presentation
• Optimizing two dimensions:
o Understanding of information content (max)
o Time spent interpreting information (min)
• Data points
o Triangles indicate direction (square = neutral)
o Color indicates direction of movement in addition to
change in Y-axis
• Smoothing via averaging
o Reduces noise in the data and presents a more long-term
view of how the information changes
• Annotations
o Doctors and patients can annotate data points
o Doctors identify healthy boundaries annotated on the
graph by light red regions
Demonstration
Summary
• In health sciences, reducing information uncertainty should
be a high priority goal
• Growing obesity and diabetes epidemic are candidates
where the application of informatics could make large
improvements
o Information needs to be readily available and understood
o Online tools can provide useful feedback at higher
frequency than traditional offline methods
• Informatics over the web introduces additional challenges
o Security and authenticity of information
o Different interpretations of standards by browsers can
introduce information uncertainty
Questions?