Upcoming SlideShare
×

# Vojtech huser-data-warehouse-evaluation-2010-04-idr-snapshot014c

622 views
534 views

Published on

1 Like
Statistics
Notes
• Full Name
Comment goes here.

Are you sure you want to Yes No
• Be the first to comment

Views
Total views
622
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
11
0
Likes
1
Embeds 0
No embeds

No notes for slide
• open Tset
• Completeness. Modeling all relevant performance factors to provide a holistic measurement of the concept. Concision. A calculation that is as simple and straightfoward as possible, making it understandable and logical to users. Measurability. Using direct performance data rather than relying too heavily on proxies or subjective measures. And from a practical perspective, if you can’t reliably gather valid data, the exercise is futile. Independence. The components of the measure need to be independent so that variation in one component doesn’t directly drive another.
• Completeness. Modeling all relevant performance factors to provide a holistic measurement of the concept. Concision. A calculation that is as simple and straightfoward as possible, making it understandable and logical to users. Measurability. Using direct performance data rather than relying too heavily on proxies or subjective measures. And from a practical perspective, if you can’t reliably gather valid data, the exercise is futile. Independence. The components of the measure need to be independent so that variation in one component doesn’t directly drive another.
• Completeness. Modeling all relevant performance factors to provide a holistic measurement of the concept. Concision. A calculation that is as simple and straightfoward as possible, making it understandable and logical to users. Measurability. Using direct performance data rather than relying too heavily on proxies or subjective measures. And from a practical perspective, if you can’t reliably gather valid data, the exercise is futile. Independence. The components of the measure need to be independent so that variation in one component doesn’t directly drive another.
• Completeness. Modeling all relevant performance factors to provide a holistic measurement of the concept. Concision. A calculation that is as simple and straightfoward as possible, making it understandable and logical to users. Measurability. Using direct performance data rather than relying too heavily on proxies or subjective measures. And from a practical perspective, if you can’t reliably gather valid data, the exercise is futile. Independence. The components of the measure need to be independent so that variation in one component doesn’t directly drive another.
• Completeness. Modeling all relevant performance factors to provide a holistic measurement of the concept. Concision. A calculation that is as simple and straightfoward as possible, making it understandable and logical to users. Measurability. Using direct performance data rather than relying too heavily on proxies or subjective measures. And from a practical perspective, if you can’t reliably gather valid data, the exercise is futile. Independence. The components of the measure need to be independent so that variation in one component doesn’t directly drive another.
• ### Vojtech huser-data-warehouse-evaluation-2010-04-idr-snapshot014c

1. 1. A Methodology for Quantitative Measurement of Quality and Comprehensiveness of a Research Data Repository (IDR Snapshot) Vojtech Huser MD PhD
2. 2. Acknowledgement <ul><li>Wisconsin Institute for discovery </li></ul><ul><ul><li>Private branch: Morgridge Institute </li></ul></ul><ul><li>CTSA grant (NIH (NCRR) 1UL1RR025011) </li></ul><ul><ul><li>U of Wisconsin-Madison + Marshfield Clinic </li></ul></ul><ul><li>Marshfield Clinic Research foundation </li></ul>marshfieldclinic.org/birc
3. 3. Agenda <ul><li>Introduction </li></ul><ul><li>Methods </li></ul><ul><ul><li>Design requirements </li></ul></ul><ul><ul><li>Measure components </li></ul></ul><ul><li>Results </li></ul><ul><ul><li>Marshfield Clinic </li></ul></ul><ul><li>Discussion/Conclusion </li></ul>marshfieldclinic.org/birc
4. 4. <ul><li>Clinical Informatics </li></ul><ul><ul><li>Subfield of Biomedical Informatics </li></ul></ul>Background marshfieldclinic.org/birc Background Hersh BMC Medical Informatics and Decision Making (2009) 9:24 doi:10.1186/1472-6947-9-24
5. 5. <ul><li>Clinical Informatics </li></ul><ul><ul><li>Subfield of Biomedical Informatics </li></ul></ul>Background marshfieldclinic.org/birc Background Hersh BMC Medical Informatics and Decision Making (2009) 9:24 doi:10.1186/1472-6947-9-24
6. 6. Introduction <ul><li>Advances </li></ul><ul><ul><li>Federated research data warehouse </li></ul></ul><ul><ul><li>Future data growth </li></ul></ul><ul><li>Conduct of research </li></ul><ul><ul><li>RCT vs. cohort and case-control studies </li></ul></ul><ul><li>IDR = Integrated Data Repository </li></ul><ul><li>Choice of multiple institutions </li></ul><ul><li>Quick evaluation of an IDR </li></ul>marshfieldclinic.org/birc Frueh FW. Back to the future: why randomized controlled trials cannot be the answer to pharmacogenomics and personalized medicine. Pharmacogenomics 2009;10(7):1077-81. TRIAD
7. 7. SHRINE: Query federation example <ul><li>Weber at al. (2009) </li></ul><ul><ul><li>Application of Information Technology: The Shared Health Research Information Network (SHRINE): A Prototype Federated Query Tool for Clinical Data Repositories </li></ul></ul><ul><ul><ul><li>JAMIA 2009;16:624-63 </li></ul></ul></ul><ul><li>Fusaro et al. </li></ul><ul><ul><li>Electronic Medical Record Analysis Using Cloud Computing </li></ul></ul><ul><ul><ul><li>AMIA CRI summit, 2010 </li></ul></ul></ul>marshfieldclinic.org/birc
8. 8. Motivation <ul><li>Motivation </li></ul><ul><ul><li>10 institutions with IDRs; budget allows only 3; How do you choose? </li></ul></ul><ul><ul><li>Tracking improvement within an organization </li></ul></ul><ul><li>Assumptions (limitations) </li></ul><ul><ul><li>Goal: lifetime and complete EHR (+genetics, +behavior, +environment) </li></ul></ul><ul><ul><li>Improvement limits </li></ul></ul><ul><ul><li>General measures (rather then project specific) </li></ul></ul><ul><ul><li>Simple evaluation, pragmatic approach </li></ul></ul><ul><ul><li>Dichotomous approach to data sources (not profiling them) </li></ul></ul><ul><ul><li>Early stage (community interest solicitation) </li></ul></ul>linkedin.com/in/vojtechhuser
9. 9. Measure <ul><li>Measures </li></ul><ul><ul><li>Qualitative description </li></ul></ul><ul><ul><li>Quantitative: Simple counts, Composite measure </li></ul></ul><ul><li>Composite scores: APGAR, FICO, Google PageRank </li></ul><ul><ul><li>Score: A short-hand way to communicate about a complex concept </li></ul></ul><ul><li>Data warehouse quality </li></ul><ul><ul><li>Size (1M vs. 10M), quality (billing data vs. structured EHR data) </li></ul></ul><ul><li>Possible requirements </li></ul><ul><ul><li>Design issues </li></ul></ul><ul><ul><ul><li>Completeness, Concision (simple), Measurability (objective, reliable), Independence </li></ul></ul></ul><ul><li>Our requirements </li></ul><ul><ul><li>Intuitive to interpret </li></ul></ul><ul><ul><li>Facilitates improvement </li></ul></ul><ul><ul><li>Fairness </li></ul></ul>linkedin.com/in/vojtechhuser
10. 10. Data Structures (VDW) linkedin.com/in/vojtechhuser
11. 11. Data Structures (i2b2) linkedin.com/in/vojtechhuser
12. 12. Data Structures (HealthFlow) linkedin.com/in/vojtechhuser healthcareworkflow.wordpress.com
13. 13. Data structures (data schemas) <ul><li>VDW </li></ul><ul><ul><li>Several tables </li></ul></ul><ul><ul><li>Data domain specific structures </li></ul></ul><ul><ul><li>Optimized for queries, users </li></ul></ul><ul><li>i2b2, HealthFlow </li></ul><ul><ul><li>One event table (+ attributes) </li></ul></ul><ul><ul><li>Generic data structures </li></ul></ul><ul><ul><li>Flexibility, extensibility </li></ul></ul>
14. 14. Event table (IDR Snapshot v1) code.google.com/p/IDRSnapshot (Event-DOB) + ‘3000-01-01’
15. 15. Results <ul><li>Events table (ev_measure01) </li></ul><ul><li>Generation took 2.5 hours </li></ul><ul><li>Size: 43 GB (includes additional info) </li></ul>code.google.com/p/IDRSnapshot