Online Program

Return to main conference page
Saturday, June 1
Data Visualization
Visualization Methods
Sat, Jun 1, 1:00 PM - 2:35 PM
Regency Ballroom C

Data visualization techniques for the analysis of eczema-affected specific regions of the body as predictors of food allergy risk (306372)


Henry T Bahnson, Benaroya Research Institute 
Helen Brough, King’s College, Evelina, Guy’s and St. Thomas’ Hospital, London 
George Du Toit, King’s College, Evelina, Guy’s and St. Thomas’ Hospital, London 
Gideon Lack, King's College London 
Colin O'Rourke, Benaroya Research Institute 
Suzana Radulovic, King’s College, Evelina, Guy’s and St. Thomas’ Hospital, London 
Katerina Salavoura, King’s College, Evelina, Guy’s and St. Thomas’ Hospital, London 
*Alyssa Ylescupidez, Benaroya Research Institute and the Immune Tolerance Network, Seattle 

Keywords: animated graphics, Graphics Interchange Format, geographic bubble plot, longitudinal, hierarchical cluster analysis

The dual-allergen exposure hypothesis postulates that epicutaneous allergen sensitization through a dysfunctional skin barrier causes pediatric food allergy (FA). To assess the role of a broken skin barrier in the development of FA, we analyzed data from longitudinal clinical assessments of the skin between 4 months and 6 years of life in multiple pediatric cohorts. In a subset of 640 participants (with healthy skin to severe eczema) skin assessments of 35 regions of the body were tracked for 6 years in order to investigate eczema phenotypes via a two-way hierarchical cluster analysis.

We classified regions into 5 areas of the body: Foot/Ankle, Leg, Trunk/Groin/Genitals, Arm/Hand/Wrist, Neck/Head. Participants had eczema on an average of nearly 4 of the 5 areas between 4 and 11 months of age and this number reduced to <1 by 60 months of age. Geographic bubble plots were overlaid on a diagram of the human body so that areas of the body affected by eczema could be visualized. Venn diagrams, parallel coordinate plots, and contour plots were also used to display these eczema patterns and their association with food allergy. Lastly, GIF (Graphics Interchange Format) technology was used to show longitudinal changes in eczema location, severity, and prevalence.

Using longitudinal skin assessments from well characterized pediatric cohorts, we examined the relationship between eczema and risk of food allergy and visualized these associations and patterns over time using innovative static and animated graphical techniques.