Data Mining Applications In Healthcare

1
Data Mining Applications In Healthcare
TEPR 2004 May 21, 2004 V. Juggy Jagannathan VP
of Research juggy_at_medquist.com
2
Introduction
Goals of todays presentation

• Provide an overview of the technologies that are
  relevant to the development and deployment of
  data mining solutions in healthcare

Allow participants to evaluate where the
technology is useful
3
(No Transcript)
4
.
Topic Outline

• Data mining
• Uses
• Algorithms
• Technology
• Applications in healthcare

5
.
Data Mining Uses

• Descriptive

Understand and characterize
Clustering Summarization Association
Rules Sequence Discovery

• Predictive

Extrapolate and forecast
Classification Regression Time-Series
6
Data Mining Algorithms

• Clustering
• Hierarchical
• Partitioned
• Genetic
• Association
• Apriori Algorithm
• If.Then rules

• Classification
• Statistical
• K-nearest neighbors
• Decision trees
• ID3
• C4.5
• Neural Networks (Self Organizing Maps)

7
Technology
Technology solutions
Data Mining Infrastructure Technologies

• Database Technologies
• On-Line Analytical Processing (OLAP)
• Visualization Technologies
• Data scrubbing technologies
• Natural Language Processing (NLP)

8
Database Technologies

• Database
• OLAP
• Visualization
• Scrubbing
• NLP

• Data warehouse vs. Data mart
• Relational technologies
• Oracle
• Microsoft
• XML-databases
• Raining Data

9
On-Line Analytical Processing

• Database
• OLAP
• Visualization
• Scrubbing
• NLP

• Analyze multi-dimensional data
• N-dimensional data cubes
• Operations
• Roll-up
• Drill-down
• Slice and dice
• Pivot

10
Visualization

• Database
• OLAP
• Visualization
• Scrubbing
• NLP

• 2D/3D Charts
• Topographic displays
• Cluster displays
• Histograms
• Scatter plots
• Advanced visualization (genomic data patterns)
• http//www.ncbi.nlm.nih.gov/Tools/

11

• Database
• OLAP
• Visualization
• Scrubbing
• NLP

• Data cleansing
• Filling in missing data
• In healthcare, there is a strong need for
  de-identification to protect privacy

12
De-Identification of Medical Records

• Names
• all elements of a street address, city, county,
  precinct, zip code, their equivalent
• geocodes, except for the initial three digits of
  a zip code for areas that contain over 20,000
  people
• all elements of dates (except year) for dates
  directly related to the individual, (e.g., birth
  date, admission/discharge dates, date of death)
  and all ages over 89
• and all elements of dates (including year)
  indicative of such age, except that such ages and
  elements may be aggregated into a single category
  of age 90 or older
• telephone numbers
• fax numbers
• e-mail addresses

• social security numbers
• medical record numbers
• health plan beneficiary numbers
• account numbers
• certificate/license numbers
• license plate numbers, vehicle identifiers and
  serial numbers
• device identifiers and serial numbers
• URL addresses
• Internet Protocol (IP) address numbers
• biometric identifiers, including finger and voice
  prints
• full face photographic images and comparable
  images
• any other unique identifying number except as
  created by IHS to re-identify information.

Source Policy and Procedures for
De-Identification of Protected Health Information
and Subsequent Re-Identification 45 CFR
164.514(a)-(c) posted by IHS (Indian Health
Services)
13
Natural Language Processing

• Database
• OLAP
• Visualization
• Scrubbing
• NLP

• NLP Uses
• translation, summarization, information
  extraction, document retrieval or categorization
• NLP Approaches
• Clustering, Classification, Linguistic analysis,
  knowledge-based analysis

• NLP Companies in health care
• A-Life
• Language and Computing

14
Applications in Healthcare

• Safety and quality
• Clinical Research
• Financial
• Public Health

15
To err is Human IOM Report

• Safety and Quality
• Clinical Research
• Financial
• Public Health

• Characterization
• JCAHO Core Measures
• CMS Quality measures starter set
• Improves patient care reactive response
• Prediction
• Identifying cases that can result in bad clinical
  outcomes and raising appropriate alarms
• Impacts patient care proactive response

16
Quality Measures Initial Set
Source http//www.cms.hhs.gov/quality/hospital/o
verview.pdf
17
Safety and Quality

• University of Mississippi Medical Center
• Data Warehouse Technologies to understand
  Medication Errors Funded by AHRQ
• Anonymous report data collection
• Data mining technologies
• Use of Neural networks and associative rule
  inference

18
Clinical Research Clinical Trials

• Safety and Quality
• Clinical Research
• Financial
• Public Health

• Pharmacy and medical claims data
• Drug efficacy and clinical trials for example
  how effective is a particular drug regimen
• Protein structure analysis
• Genomic data mining
• Diagnostic Imaging data research

19
The bottom line on cost

• Safety and Quality
• Clinical Research
• Financial
• Public Health

• General Utilization review does the care
  provided meet accepted clinical and cost
  guidelines
• Drug Utilization review
• Outlier analysis exceptions to treatment
  analyzing treatments which cost more than the
  normal or less than normal.

20
Data mining in public health

• Safety and Quality
• Clinical Research
• Financial
• Public Health

• Syndromatic surveillance
• Bio-terrorism detection
• Communicable disease reporting (Centers for
  Disease Control (CDC))
• DAWN (Drug Awareness and Warning Network)
• Federal Drug Agency (FDA) reporting of adverse
  drug events.

Example effort AEGIS
21
Conclusion

• Descriptive
• Predictive

• Classification
• Clustering
• Association rules

• Database
• OLAP
• Visualization
• Scrubbing
• NLP

• Data mining
• Uses
• Algorithms
• Technology
• Applications in healthcare

• Safety and Quality
• Clinical Research
• Financial
• Public Health

22
Conclusion
Technology solutions
juggy_at_medquist.com

• uestions?