30.01.2015 Views

Managing Director, Analytical Applications

Managing Director, Analytical Applications

Managing Director, Analytical Applications

SHOW MORE
SHOW LESS

Transform your PDFs into Flipbooks and boost your revenue!

Leverage SEO-optimized Flipbooks, powerful backlinks, and multimedia content to professionally showcase your products and significantly increase your reach.

TDWI<br />

Dan Housman<br />

<strong>Managing</strong> <strong>Director</strong>, <strong>Analytical</strong> <strong>Applications</strong><br />

July 26, 2010<br />

Copyright © 2010 Recombinant Data Corp. All rights reserved.<br />

1


About Recombinant<br />

Best of breed software/analytics company focused 100%<br />

on secondary uses of clinical and pre-clinical data<br />

Core Competencies<br />

• Healthcare data warehousing & integration services<br />

• Translational research & quality reporting solutions<br />

• Data strategy, governance & compliance consulting<br />

• Open Source implementations & extensions<br />

Core Values<br />

• Pragmatism<br />

• Effective Communication<br />

• Trust<br />

2


Translational Medicine Defined<br />

Non-profit research<br />

Pharma R&D<br />

Clinical<br />

Trials<br />

Adoption of<br />

evidence based medicine<br />

3


High growth high expertise<br />

• Recombinant Customers<br />

• 11 – Year Ago<br />

• 27 – Today<br />

• Recombinant Employees<br />

• 21 – Year Ago<br />

• 42 – Today<br />

• 50 projected by end of 2010<br />

• Healthcare requires expertise<br />

• Matvey Palchuk, MD, MS<br />

<strong>Director</strong> of Medical Informatics<br />

• D. Doug Liu, PhD<br />

Chief Scientist<br />

• David Fram, PhD<br />

Research <strong>Director</strong><br />

4


2010 Business Drivers<br />

• Research/Quality convergence<br />

• Quality activities require research approaches<br />

• Cross-over of tools (e.g. GPP, i2b2 for quality)<br />

• Registries<br />

• Bio-banking / genomics<br />

• Lower costs for genetic research drive clinical data<br />

• Big studies/big data sets<br />

• ARRA driving new data demands<br />

• HIE, Interoperability, accountable care, shared decision<br />

making, permissions<br />

• Public data use<br />

• Profiles, pre-competitive sharing, benchmarking<br />

5


Target vision - incremental approach<br />

Clinical<br />

Data<br />

6


Research Study Process<br />

7


HIPAA Wall<br />

Patient Care HIPAA/IRB Wall Clinical Research<br />

Comparative<br />

Effectiveness<br />

Research<br />

Cohort<br />

Identification<br />

Population<br />

Research<br />

Virtual<br />

Trials<br />

Claims<br />

Trial Analysis<br />

EHRs<br />

Clinical Data<br />

Warehouse<br />

De-identification<br />

Research<br />

Mart<br />

EMPI<br />

eIRB/Protocols<br />

Quality<br />

Reporting<br />

PHI<br />

Honest Broker<br />

Identity<br />

Management<br />

Marts<br />

Patient<br />

Recruitment<br />

Consent<br />

CTMS<br />

„Omics<br />

Repository<br />

Trial EDC<br />

Clinical LIMS Tissue Consent Tissue Bank High Throughput „Omics LIMS<br />

8


What is the benefit to each side<br />

Research To Quality<br />

Benefits<br />

Research<br />

Access rich<br />

research registry<br />

and study data<br />

Translate<br />

findings to<br />

improve care<br />

model<br />

Bridge genomics<br />

personalized<br />

medicine<br />

Leverage<br />

Informatics<br />

expertise and<br />

tools<br />

Quality to Research<br />

Increased data<br />

availability<br />

Increased data<br />

quality/source<br />

knowledge from<br />

use in PTO<br />

Access to<br />

patient cohorts<br />

Quality study<br />

collaboration<br />

Quality<br />

Translational Approach<br />

9


AMCs are Building Toward Personalized Medicine Vision<br />

10


Trust<br />

11


Our solution Map<br />

12


Data Trust Conceptual Model<br />

13


Open source initiatives<br />

14


Open source initiatives<br />

• i2b2<br />

• Pentaho ETL/BI<br />

• Hudson<br />

• Indivo PHR<br />

• REDCap Electronic Data Capture<br />

• Gene Pattern<br />

• caTissue<br />

• Profiles Research Networking Software<br />

• Lucene search<br />

15


Open source data management<br />

• Sometimes free<br />

• Shared ideas<br />

• Interoperability<br />

• Science friendly (share)<br />

• Evidence based directive<br />

• Metrics/Benchmarks<br />

• Anyone can contribute<br />

• Grants<br />

• Innovation<br />

• Dissemination<br />

• Distrust by IT/CIOs<br />

• Support model<br />

• Marketing/sales<br />

• Solution providers<br />

• Commercial FUD<br />

• Completeness<br />

• Predictability<br />

• Financial limits<br />

What is in the best interest of patients<br />

16


Data Access Problem<br />

Unmanaged<br />

Data<br />

PHI<br />

Controls<br />

System<br />

Load<br />

Unapproved<br />

Research<br />

Bottleneck<br />

Repeat<br />

Queries<br />

Project<br />

Distractions<br />

17


The Solution: Managed self-service<br />

18


Full Research Workflow<br />

Step 1 Step 2<br />

Step 6<br />

Data Collection<br />

Institutional<br />

Data Sources<br />

EMR<br />

Claims<br />

Clinical<br />

Data<br />

Warehouse<br />

(PHI)<br />

Integration<br />

ETL<br />

(data extracts<br />

and HL7 feeds)<br />

Step 3<br />

Create<br />

“Limited<br />

Data Set”<br />

(destroy<br />

identifiers)<br />

Operational<br />

Use – QR/QI<br />

H I P A A W A L L<br />

Step 4<br />

i2b2 research<br />

mart<br />

(Limited Data<br />

Set)<br />

Step 5<br />

i2b2 queries<br />

One-time Data Use<br />

Agreement<br />

eIRB protocol<br />

request<br />

Step 7<br />

IRB approval<br />

EMPI<br />

IRB-Approved<br />

PHI is retrieved<br />

Step 8<br />

Schedule<br />

Etc.<br />

Step 9<br />

i2b2 research<br />

mart with<br />

patient<br />

identifiers<br />

Step 10<br />

SUBMIT<br />

Re-issue query<br />

to data mart<br />

with PHI<br />

Data Center<br />

(Honest Broker and server support)<br />

19


i2b2 History<br />

1999<br />

RPDR initially<br />

established at<br />

Partners<br />

2004<br />

i2b2 grant<br />

to “port<br />

RPDR” as<br />

an open<br />

source<br />

framework<br />

2008<br />

Strong<br />

adoption<br />

2010<br />

i2b2<br />

renewal<br />

2000-2004<br />

RPDR and<br />

tool<br />

evolution<br />

at Partners<br />

2006<br />

First beta<br />

release of<br />

i2b2<br />

2009<br />

De-facto<br />

CTSA<br />

standard<br />

CRDW<br />

* RPDR (Research Patient Data Registry) was the precursor to<br />

i2b2<br />

20


i2b2 Academic User Group / Community<br />

United States<br />

• Beth Israel Deaconess Hospital, Boston, MA<br />

• Boston University School of Medicine, Boston, MA<br />

• Brigham and Women's Hospital, Boston, MA<br />

• Case Western Reserve Hospital<br />

• Children's Hospital of Boston, MA<br />

• Children's Hospital of Cincinnati, OH<br />

• Children's Hospital of Denver, CO<br />

• Children's Hospital of Philadelphia, PA<br />

• Cleveland Clinic, Cleveland, OH<br />

• Duke Medical College<br />

• GE Research, Niskayuna, NY<br />

• Group Health Cooperative<br />

• Harvard Medical School, Boston, MA<br />

• Kaiser Permanente, Oakland, CA<br />

• Kimmel Cancer Center, Philadelphia, PA<br />

• Maine Medical Center, Portland, ME<br />

• Marshfield Clinic, Wisconsin<br />

• Massachusetts General Hospital, Boston, MA<br />

• Morehouse School of Medicine, Atlanta, GA<br />

• Ohio State University Medical Center, Columbus, OH<br />

• Oregon Health & Science University, Portland, OR<br />

• Renaissance Computing Institute, Chapel Hill, NC<br />

• Tufts New England Medical Center, Boston, MA<br />

• University of California, Davis, CA<br />

• University of California, Irvine, CA<br />

United States (continued)<br />

• University of California, San Francisco, CA<br />

• University of Cincinnati, Cincinnati, OH<br />

• University of Massachusetts Medical School, Worcester,<br />

MA<br />

• University of Michigan Medical Center, Ann Arbor, MI<br />

• University of Pennsylvania School of Medicine, Phila, PA<br />

• University of Rochester Medical Center, Rochester, NY<br />

• University of Texas Health Sciences Center SW, Dallas,<br />

TX<br />

• University of Texas Health Sciences Center, Houston, TX<br />

• University of Texas Health Sciences Center, San Antonio,<br />

TX<br />

• University of Washington, Seattle, WA<br />

• Utah Health Science Center, Salt Lake City, UT<br />

• Veterans Administration Boston and Utah<br />

• Weil Medical College of Cornell, NYC, NY<br />

International<br />

• Georges Pompidous Hospital, Paris, France<br />

• University of Goettingen, Goettingen, Germany<br />

• University of Pavia, Pavia, Italy<br />

• University of Seoul, Seoul, Korea<br />

Pharmaceutical<br />

• Johnson and Johnson<br />

21


i2b2 Hive<br />

22


Star Schema: i2b2<br />

• Entity Attribute Value (EAV) Data Model<br />

23


Grid projects: SHRINE and caBIG<br />

24


Research systems<br />

25


J&J gets it<br />

27


Best Practices Award<br />

• Recombinant Solution<br />

Wins 2010 Bio-IT World<br />

Best Practices Award in<br />

Knowledge Management<br />

• Implementation at<br />

Centocor R&D, Inc., a<br />

subsidiary of Johnson &<br />

Johnson<br />

• tranSMART designed to<br />

share drug research and<br />

development data across<br />

Johnson & Johnson<br />

28


Infrastructure Solution<br />

29


tranSMART primary components<br />

Search<br />

Data Analysis<br />

30


Subset Definitions<br />

31


Set Value for Query<br />

32


Box Plot<br />

33


Grid View<br />

34


Integration of Omics and i2b2<br />

1.Select<br />

Cohort(s)<br />

2. Initialize<br />

<strong>Analytical</strong><br />

Pipeline<br />

3. Interactive<br />

Analysis<br />

Wizard<br />

4. Compile<br />

Results<br />

Across<br />

datasets<br />

5. Present<br />

results to<br />

scientist<br />

Workflow<br />

Engine<br />

i2b2<br />

Clinical<br />

Data set<br />

Reference Data<br />

(Gene Annotation)<br />

Gene<br />

Expression<br />

SNP<br />

Chip<br />

35<br />

35


Heat Map<br />

36


GenISIS Architecture: MVP<br />

37


Profiles Research Networking Software<br />

Open Source Initiative<br />

38


Profiles Research Networking Software 1<br />

• Research networking &<br />

expertise mining tool<br />

• Developed as part of Harvard<br />

Catalyst, The Harvard Clinical<br />

and Translational Science<br />

Center<br />

• Deployed at UCSF as an alpha<br />

site<br />

• Now being released as open<br />

source software<br />

• Recombinant is an authorized<br />

support provider for Profiles<br />

Research Networking Software<br />

• Working with Griffin Weber to<br />

implement two beta sites<br />

1<br />

Profiles Research Networking Software was developed under the supervision of Griffin M Weber, MD, PhD, with support from<br />

Grant Number 1 UL1 RR025758-01 to Harvard Catalyst, The Harvard Clinical and Translational Science Center from the National<br />

Center for Research Resources and support from Harvard University and its affiliated academic healthcare centers<br />

39


Professional Open Source Model<br />

Community<br />

Awareness<br />

Licensing<br />

Implementation<br />

End Users<br />

Innovation<br />

Support<br />

40


Profiles Reference Data Service<br />

ClinicalTrials.gov<br />

Other New<br />

Sources<br />

Grants<br />

Thomson Web<br />

of Knowledge<br />

PubMed<br />

Profiles Reference Data Service<br />

Disambiguation<br />

Search on Demand<br />

rProfiles<br />

Tagging<br />

Reference Update<br />

Client A<br />

Client B<br />

Client C<br />

Client D<br />

Profiles Research<br />

Networking Software<br />

Deployment<br />

Profiles Research<br />

Networking Software<br />

Deployment<br />

Profiles Research<br />

Networking Software<br />

Deployment<br />

Profiles Research<br />

Networking Software<br />

Deployment<br />

41


Active & Passive Networking<br />

42


Keyword Networks<br />

43


Co-Author Networks<br />

44


Co-Author Networks: Map View<br />

45


Co-Author Networks: Network View<br />

46


Similar People Networks<br />

47


Quality Reporting Overview<br />

48


Quality Reporting Architecture<br />

49


Avoid Reinventing the Wheel<br />

50


Recombinant Goal<br />

51


Analytic Engine Process<br />

Encounters<br />

Calculated<br />

Facts<br />

Roll-Up<br />

Relationships<br />

PQRI<br />

Calculation<br />

Engine<br />

PQRI<br />

Data Trust<br />

HEDIS<br />

Rule Meta Data<br />

HEDIS<br />

Custom<br />

PQRI HEDIS Custom<br />

Custom<br />

Terminologies<br />

Rule Agent<br />

CPT<br />

LOINC<br />

ICD-9<br />

RxNorm<br />

eMeasure<br />

(HL7)<br />

XML<br />

52


hours<br />

Operations<br />

Percent<br />

Operations<br />

• Denominators (panels, locations)<br />

• Practice snapshots<br />

400<br />

Visit to note cycle time<br />

95<br />

Percentage of notes initialized within N hours<br />

• EHR adoption<br />

• Visits/utilization<br />

350<br />

300<br />

250<br />

200<br />

150<br />

100<br />

50<br />

0<br />

2003-11<br />

2004-3<br />

Avg. Time initialization (hrs)<br />

2004-7<br />

2004-11<br />

2005-3<br />

2005-7<br />

2005-11<br />

2006-3<br />

2006-7<br />

2006-11<br />

2007-3<br />

2007-7<br />

Avg. Time to finalization (hrs)<br />

90<br />

85<br />

80<br />

75<br />

70<br />

65<br />

60<br />

55<br />

50<br />

45<br />

2003-11<br />

2004-3<br />

2004-7<br />

2004-11<br />

2005-3<br />

2005-7<br />

2005-11<br />

2006-3<br />

2006-7<br />

2006-11<br />

2007-3<br />

2007-7<br />

% initialized 72 hr from visit % initialized 120 hr from visit<br />

% created total<br />

Chronic Disease Mgmt.<br />

Preventative Medicine<br />

Patient Satisfaction<br />

In Patient/Hospital<br />

Ad-hoc queries<br />

Report Library<br />

53


Operations<br />

Chronic Disease Management<br />

• Diabetes<br />

• Asthma<br />

• CAD<br />

• CHF<br />

Chronic Disease Mgmt.<br />

Preventative Medicine<br />

Patient Satisfaction<br />

In Patient/Hospital<br />

Ad-hoc queries<br />

Report Library<br />

54


Data Needed for Typical Report<br />

55


Patient Satisfaction<br />

56


Bending the curve<br />

2010<br />

2013<br />

Payment<br />

Reform<br />

57


What’s coming – ACO analytics<br />

SPORT trial – Jim Weinstein et. al<br />

58


Thank You<br />

Recombinant Data Corp.<br />

255 Washington Street, Suite 235, Newton, MA 02458<br />

Tel: (617) 243-3700 Fax: (617) 243-3788<br />

www.recomdata.com<br />

59

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!