The Cathedral and The Data Bazaar - Enabling Lean Data

We can truly free the data only if we can show a bazaar like value for it! Bazaars flourish in short, small value loops. Cathedrals take time!

Read the full talk in detail at: http://bit.ly/lean-data-talk

Transcript

1. The Cathedral And The Data Bazaar Enabling Lean Data @vaibhavb
2. Who is Vaibhav? DIRECTOR OF PRODUCT MANAGEMENT, OPTUM AUTHOR, O’REILLY MEDIA TWITTER – @vaibhavb BLOG – http://healthblog.vitraag.com ME – http://about.me/vaibhavb Hobbies: Interactive Furniture, Mixology, Cycling, Mountaineering Associations: United States Cycling Federation (USCF) Referee
3. Agenda •  The Cathedral model for Healthcare Data •  The Bazaar - Lean Data •  Clinical Data Connectivity –  Purpose –  Access –  Security –  Sharing
5. Problem with current model(s) •  Central Monolithic design with too many interdependencies •  Hard to see value when doing data aggregation •  At scale & adoption issues for data sets released –  Open Data –  Meaningful Use Mandated data •  Design by committee
6. ONC “Learning Health System” Roadmap •  10 years! •  Over 50 use cases •  A standards advisory issued every year
7. Picture
9. Bazaar – Lean Data •  A bare bones simple platform •  Participants can create and absorb value independently •  Data is judged and consumed by value!
10. Lean Data – A Case Study Value Based Consumers Paper Charts EMRs Labs Clinical Data Aggregation
11. Purpose First! •  Drive Measures –  ABA: Adult BMI Assessment –  CDC-LDL: Comprehensive Diabetes Care & LDL screening –  ..
12. Pillars of Access •  Transport •  Format •  Semantics
13. Data Transport •  Secure FTP •  XDS.B •  Direct •  Proprietary REST API •  FHIR REST API
14. Format •  Proprietary CSV Files •  HL7 V2 (ADT, ORU etc.) •  C32 •  CCDA •  FHIR JSON
15. Semantics •  Problems –  ICD 9/ 10 –  SNOMED-CT •  Procedures –  CPT •  Labs –  LOINC •  For example BMI values require SNOMED-CT codes
16. Security •  Self Signed X509 Certificates –  SFTP –  SSL Mutual Authentication (XDS.B/ REST API) •  Direct Certificates & Identity Proofing –  Direct •  O-Auth2 –  FHIR 16
17. Sharing – Data Access •  Data Access –  Simple •  Member Demographics (First Name, Last Name etc.) shared with Data Provider •  Data Shared only on a full match! –  Standards Based •  IHE PIX/PDQ 17
18. Sharing – Data Use •  Data Use established per use-case / measure •  Eventual standards based – XACML •  Data shared with HEDIS engines through network share in flat files – Lean! 18
19. Data Value == Data Freedom •  When you are freeing the data make sure there is immediate value! •  Choose Lean vs. end all be all •  Establish valid crawl, walk run approaches for –  Access –  Security –  Sharing
20. Thank You! Questions? TWITTER – @vaibhavb BLOG – http://healthblog.vitraag.com ME – http://about.me/vaibhavb