ECHO: A System for Fighting Childhood Obesity with Screening, Intervention and Tracking.
Childhood obesity is a serious national public health epidemic in the United States threatening the health of millions of
children/youth. Over the past three decades, according to many national studies, obesity rates in the US have more than doubled
among children ages 2 to 5 and more than tripled among those ages 6 to 11. Roughly, nine millions kids over the age of 6
are considered obese and the health consequences are already apparent. More and more children are being diagnosed with diseases
previously considered adult illnesses. Developing a mass, effective tracking and intervention system to fight childhood obesity
is a national health imperative.
State of Illinois, like other parts of the county, has seen obesity prevalence also continuously increase during the past decades.
Fifty nine percent of Illinois adults are overweight or obese (23.2% obese). The obesity rate among Illinois adults doubled between 1990 and 2002.
Fourteen percent of low-income children between two and five years of age in Illinois are overweight.
(Source:
Trust for America's Health.)
With the support of the Illinois Department of Public Health, we have begun an effort to
collect and store the Child Health Exam Certificate (CHEC) records in a statewide database.
Currently, this database serves mainly the group-data storage and analysis
functions, and no individual tracking and intervention functions are included.
The purpose of our project is to develop and pilot
a state-wide, state-of-the-art student personal health record tracking system called
ECHO (Electronic Children Health Observatory)
that can screen and track childhood obesity, as well as provide intervention if needed, for the State of Illinois.
There are three specific aims in this study:
- To complete the development of the major system architecture of ECHO and its application elements;
- To pilot the system in different application settings and to evaluate the efficiency and utility of the system;
- To modify the system based on the feedback, develop instructions for the future implementation of the system
and estimate the cost of the implementation.
Test application and database
A prototype of database of children has been developed.
Any real application, of course has to
deal with either Parental authorization for ANY sharing outside school or comply with
Health Insurance Portability and Accountability Act
(HIPAA) and
Family Educational Rights and Privacy Act
(FERPA)
de-identification procedures.
For more information on HIPAA
and FERPA,
see a handbook that "provides practical guidance to help school district personnel navigate the myriad and sometimes conflicting legal
considerations connected to managing student health information.":
N. Schwab, M. Rubin, J.A. Maire, M. Gelfman, M.D. Bergren, D. Mazyck and B Hine
Protecting and disclosing student health information: Guidelines for developing school district policies
and procedures.
Kent, OH: ASHA. 2005.
CHEC form processing
System is required for the automated extraction of handwritten information from
Department of Human Services (DHS) paper forms that parents fill out with child's health history.
The form is not standardized for automated data extraction and inconsistency of labeling may exist.
Additionaly distortion is introduced via faxing and scanning.
We have also prototyped ECHO system by collecting the forms for recognition and extraction of handwritten information as well
as classification of binary answers with high confidence.
An example CHEC form is shown in Figure 1. The two images in Figure 1 illustrate document warping due to faxing and scanning, as it
poses challenges to automated processing. Once information has been extracted from the form, it is then analyzed using various
classification schemes (e.g. Body Mass Index) to provide better understanding of the health records.
Figure 1. (a) An example of Child Health Exam Certificate (CHEC) form:
Document with fax related artifact in the middle.
(b) Detail of scanned document and its incomplete lines due to scanning process.
A system for automated processing has been developed will following features:
(a) Scanning and storing forms, (b) registration of scanned forms, (c) extraction of binary entries,
(d) extraction of handwritten entries, (e) alternative PDF document processing
Registration of Scanned Forms
Figure 2. Registration tool. (Left) Original CHEC form and faxed and scanned copy (right).
Registration is either Manual by selecting control points or Automated by image feature matching.
Figure 3. Original form registered with the scanned one is shown as subtracted image.
Extraction of Binary Entries
Figure 4. Comparison of the checked answers (left) and distorted ones (right). Analysis of sub-areas.