Omni-HealthData™ Analytics Beyond the EMR Big Data

Omni-HealthData Analytics Beyond the EMR Big Data Governance, Enabling Technologies, and Cutting Through the Hype Shawn S. Sutherland, CPHIMS Shawn [email protected] Manager, Patient and Member Outcomes www.OmniHealthData.com

Hi, my name is Shawn, and . 2

Where to begin 3

People 1st (and technology for lunch) 4

Define clear roles for the different stakeholders IT Business Clinicians 5

Data – must be treated as a valued asset 6

Inclusive Process – Nothing falls through the cracks 7

Upstream vs downstream management of data 8

Big Data, really? 9

Avoid the hype – There are many choices for one decision and all can go different directions 10

Cut through the HYPE – What is Big Data? How we used to talk about Big Data in Healthcare D ata R ich I nformation P oor 11

Cut through the HYPE – What is Big Data? Modern definition 12

Healthcare Variety Example #1 – extra easy Patient Gender(22) Omni-Patient 01 M F T I N 2 3 4 99 MALE 01 02 FEMALE 03 05 UNKNOWN A B C D 09 TRANSGENDER N/A OTHER NULL Validation Harmonization Standardization Patient Gender (6) A F M N O U Ambiguous Female Male Not applicable Other Unknown Marital Status (16) Marital Status (37) 01 02 MRD DIV OTHR NULL 03 04 05 LSPMRD LSP DIV 06 07 99 ALD OTR SGL ,02S A01 D01 M00 DPR UNK MS 001 MS 002 W04 D02 I01 WDW MS 003 S01 R01 2013 M1 MS 004 2320 M2 Omni-Codes Other Reference Medical Terminology A B C D E G I M N O P R Separated Unmarried Common law Divorced Legally Separated Living together Interlocutory Married Annulled Other Domestic partner Registered domestic partner S T U W Single Unreported Unknown Widowed 13

Healthcare Variety Example #2 – Hospital, Unit, Bed – kind of easy? Hospital IBH of New York IB and Jersey Medical Center IBHealth Hospital of New York IBJ Med Center The IBHealth Medical Center IBJ Medical Center IBHealth Medical Center IBJ NY CMS Hospital Name TIN Hospital Name IBJ NY and Jersey . Unit Four North Four-North FourNorth 4 North 4-North 4 North 04North 04 4N Bed 4N 420 4N-420 4 N 420 0420 RM 420 420 00420 N420 0420 14

Healthcare Variety Example #3 – Ok, ok, that’s hard! Terminology Sets (partial list) SNOMED CT ICD-9-CM ICD-10-CM/PCS CPT-4 Medical necessity Age/gender edits HCPCS APC MS-DRG, AP-DRG LOINC DSM IV Medications: RXNorm, FDB, NDC, NDF-RT, Medi-Span, Multum HLI Medical Specialty Subsets PQRS Subsets HLI Medical Specialty Subsets Nursing: NIC, NOC, NANDA HL7 CDT UCUM UNI HRG CCI CVX, MVX Rev codes Multiple languages Mappings (partial list) SNOMED CT to ICD-9-CM SNOMED CT to CPT SNOMED CT to MeSH SNOMED CT to ICD-10-CM/PCS ICD-9-CM to ICD-10-CM/PCS ICD-10-CM/PCS to ICD-9-CM ICD-9-CM to SNOMED CT ICD-10-CM to SNOMED CT CPT to SNOMED CT DSM IV to SNOMED CT RxNorm to NDF-RT CPT to CVX 15

Big Data Environment Architecture Source System Integration Tools Clean Write Back to Source Systems Source System Integrity Tools Real Time ETL Function call requests to source systems to leverage their analy tic functions for data enrichment. Batch Load ESB Master Data Management 360 Object View Remediation Portal Domain Design Mastered Domains Reference Domains Workflows Analysts Hypothesis Testing Vulnerability Identification MDM workflows utilized to coordinate Analysts efforts. 3rd Party Taxonomy Local Knowledge Base Big Data Integration Flume Spark MLLIB SparkR Kafka Topics HDFS Data Quality Profile Clean Standardize Validate Enrich Knowledge Based System Data Stewards SME rules Perform Remediation Parquet Use of DQ, R, MLLIB capabilities for relational, temporal, geospatial, classification and statistical analytics. Data Store Interface Architect Data Scientist Data to feed a Knowledge Based System (KBS) for classification analysis based on 3rd party and local SME knowledge Enterprise Architect Intelligence Tools Rstat (Hadoop Version) Predictive Prescriptive WebFOCUS Magnify Business Users / Customers WebFOCUS Scorecards Reports Dashboards Hadoop Enabled BI Developer 16

MDM platform leveraging Big Data for staging and consumption Pre-built data models for mastered and transactional domains Pre-built processing, quality, mastering, and remediation rules 360 Degree View on Members and Providers through Data and Analytics Reference data Code sets: HLI Analytic Analytic Data Data Staging Staging Environment Integrate, Cleanse, Correlate, Steward Consumption: HEDIS Grouper, CCD, views, etc. Subject Oriented Data On Ramp External data Member Administrative Clinical (CCD, HL7, etc.) Facility Provider info Data Lake Internal data Member (e.g., Initiate) Claims Eligibility Downstream apps Provider relations Claims adjudication Analytics Data warehouses Data marts External Provider & member portals Reimbursement 17

Healthcare is collaborative by nature 18

So, read THIS and join THEM 19

Omni-HealthData Analytics Beyond the EMR Big Data Governance, Enabling Technologies, and Cutting Through the Hype Shawn S. Sutherland, CPHIMS Shawn [email protected] Manager, Patient and Member Outcomes www.OmniHealthData.com