From Biomedical Literature to Electronic Health Record Health Grid for Research and Clinical Decisions - PowerPoint PPT Presentation

Loading...

PPT – From Biomedical Literature to Electronic Health Record Health Grid for Research and Clinical Decisions PowerPoint presentation | free to view - id: 7ec79-ZDc1Z



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

From Biomedical Literature to Electronic Health Record Health Grid for Research and Clinical Decisions

Description:

Health Grid for Research and Clinical Decisions. Graduate Institute of. Medical Informatics ... Graduate Institute of Medical Informatics. National Health Insurance ... – PowerPoint PPT presentation

Number of Views:97
Avg rating:3.0/5.0
Slides: 25
Provided by: Ash13
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: From Biomedical Literature to Electronic Health Record Health Grid for Research and Clinical Decisions


1
From Biomedical Literature to Electronic Health
RecordHealth Grid for Research and Clinical
Decisions
  • Graduate Institute of
  • Medical Informatics
  • Taipei Medical University, Taipei, Taiwan
  • Yu-Chuan (Jack) Li, M.D., Ph.D.
  • Arbiter Lin, M.D., M.S.

2
Biomedical Data for Research and Clinical use
  • Scale, complexity and timeliness
  • Massive data and heavy computation
  • Data ownership problem
  • Privacy problem
  • Competition and collaboration among hospitals and
    research institutes
  • Grid may be ideal here!

3
Range of Applications
Genome
Literature text mining Microarray data mining
Transcriptome
Proteome
Metablome
Disease
Literature text mining NHI data mining EHR
data mining
Treatment
4
Scale and Complexity
  • Human Genome 18,000 defined in Gene Ontology,
    total 25,000?
  • Human Proteome not well defined, total
    50,00075,000?
  • Disease 11,000 defined in ICD-9-CM
  • Treatment 20,000 defined in NHIs (National
    Health Insurance) medication procedures

5
National Health Insurance
  • Bureau of Nation Health Insurance
  • National Health Insurance for all people in
    Taiwan since 1995
  • NHI Smart Card issued to all 23 million people in
    Taiwan since 2004.01

6
NHI Smart Card
7
Scale and Complexity (cont.)
  • National Health Insurance DB 5TB_at_500GB/yr
  • 600 hospitals and 17,000 clinics connected in
    real-time to the NHI for the health smart card
    authentication
  • But the bandwidth is mostly 512KB/64KB in ADSL

8
Size of Medical Data in Taiwan
  • Outpatient 300 million visits / yr
  • Inpatient 2.8 million-days / yr
  • 1.5 billion prescription / yr
  • 900TB image data per year
  • 30TB text/coded data per year
  • Growing exponentially in the next 5 years while
    Electronic Health Record (EHR) matures

9
Standardized EHR Project
10
MIEC Project
  • National Medical Information Exchange Center
  • prototype in 1997
  • Hospitals treat health and medical data as their
    own property
  • Not willing to share with other hospitals
  • Concern about privacy, legal and business issues

11
To Share, or Not to Share
  • Medical data are sensitive and proprietary
  • De-identification is not enough
  • Practice patterns, medication consumption
    patterns, outcome variations, case-mix indexetc.
    are still sensitive information
  • Share only the results of aggregated computation,
    not individual hospital
  • Privacy enhancing technologies
  • Multiparty private computation

12
Scenario Carpal Tunnel Syndrome
  • A physician of rehabilitation may want to know
  • Percentage of different treatment on CTS in whole
    Taiwan
  • Surgical Operation
  • Rehabilitation
  • Acupuncture
  • Outcome of each treatment options for a patient
    with specific age/sex

13
Scenario A 58 year-old female
  • Lab Data cholesterol 480mg/dl
  • A doctor may want to know
  • Treatment options at these age/sex/lab
  • Medication usage etc.
  • Outcome of each treatment options in Taiwan
  • The percentage of people who eventually get
    Coronary Artery Disease

14
A Health Grid Can Work
  • For physicians
  • Weighing treatment options for individual patient
  • For patients
  • Know our options and risks
  • For health policy maker
  • Public health policy making

15
Networking Environment
16
Biomedical Literature Mining for Gene and Disease
Relationship
17
Range of Applications
Genome
Literature text mining Microarray data mining
Transcriptome
Proteome
Metablome
Disease
Literature text mining NHI data mining EHR
data mining
Treatment
18
Probabilistic Relationship Among Genes and MeSH
terms
  • Medical Literature 13 million citations
    collected in Medline
  • Medical Terms 341,000 defined in MeSH (Medical
    Subject Headings)
  • 18,000 gene names

19
MeSH term Breast Neoplasms
20
(No Transcript)
21
Complex Joint Probability Computation
22
Future Applications
  • Text mining on literature and free-text data from
    EHR
  • Data mining on coded/numerical data from EHR or
    NHI DB or Gene/Protein chips
  • Support medical decision making and public health
    policy making

23
Conclusion
  • Use Grid technology to collaborate hospitals and
    academic institutes
  • Build a testbed and demo site of a Health Grid in
    Taiwan
  • Increase international collaboration working
    with EGEE
  • EU project - Supporting and structuring
    HealthGrid Activities Research in Europe (SHARE)

24
Welcome to ISGC 2004 in Taiwan!
Q A
  • Thanks you!
About PowerShow.com