Safety surveillance in real world data: leveraging transactional - - PowerPoint PPT Presentation

Worldwide Safety Strategy

Safety surveillance in real world data: leveraging transactional insurance claims databases and electronic medical records

Andrew Bate Andrew Bate Senior Director, Analytics Team Lead, Epidemiology, Worldwide Safety & Regulatory

International Society of Pharmacovigilance Annual Conference Cancun, Mexico November 1, 2012

Worldwide Safety Strategy

Overview

• Real world data for analysis

– Examples of databases

• Recent and ongoing global initiatives on

safety surveillance research safety surveillance research

• Differing data access models for surveillance

– Centralized v distributed/federated

Worldwide Safety Strategy

Some selected observational databases

Database Country Characteristic Population Size

THIN UK GP primary care database 10.5 M1 Danish National Health Service Register Database Denmark Healthcare registry of care 5.5 M2 Premier US Clinical data from the hospitals 130 M+ patient discharges3 Normative Health Transactional Normative Health Information (NHI) Database US Transactional claims records of a commercial health insurer 60 M+4 Health Insurance Review and Assessment Service (HIRA) Korea Insurance Claims from near universal national system 48 M5

1 Blak et al Generalisability of The Health Improvement Network (THIN) database: demographics, chronic

disease prevalence and mortality rates. Informatics in Primary Care 2011;19:251–5

2 Furu K. et. al. The Nordic Countries as a Cohort for Pharmacoepidemiological Research. Basic & Clinical

Pharmacology &Toxicology 2009; 106: 86-94

3 Fisher BT et al. In-hospital databases In Pharmacoepidemiology 5th Edn 2011 pp 244-258 4 Seeger J, Daniel GW. Commercial Insurance Databases. In Pharmacoepidemiology 5th Edn 2011 pp 189-208 5 Kimura T et al. Pharmacovigilnace systems and databases in Korea, Japan and Taiwan.

Pharmacoepidemiology and Drug Safety. 2011; 20: 1237–1245

Worldwide Safety Strategy

International and National Initiatives addressing database surveillance

• FDA Sentinel Initiative
• Observational Medicines Outcomes Partnership (OMOP)
• Innovative Medicines Initiative (IMI) project: PROTECT

– European Community's Seventh Framework Programme European Community s Seventh Framework Programme (FP7/2007-2013) for the Innovative Medicine Initiative

• European Commission Seventh Framework Programme

(FP-7) of the Research Directorate: EU_ADR

• CIOMS VIII “Practical Aspects of Signal Detection in

Pharmacovigilance”

• Many other international, national and regional initiatives

Worldwide Safety Strategy

Sentinel Initiative

• Use large claims databases and EHRs for

analysis of drug outcomes

• Link in “distributed network”
• FDAAA call for access to: 25M patient
• FDAAA call for access to: 25M patient

records by 2010

– 100M patient records by 2012

Source: Janet Woodcock ”CDER Priorities for 2009” Accessed FDA website 14/1/2009

Worldwide Safety Strategy

Novel Use of Claims & EMRs for signal detection/refinement

How to best utilise the wealth of Real World Data and does its value change depending on purpose?

6

Signal Generation

• Any Medical Event
• Designated Medical Events

Signal Refinement Signal Evaluation

Rapid Detect the unexpected Less persuasive Time Consuming Test the anticipated Convincing Product Approval & Launch

Worldwide Safety Strategy

OMOP

• The Observational Medical Outcomes

Partnership (OMOP) is a public-private partnership designed to help improve the monitoring of drugs for safety. The partnership will conduct a two-year initiative to research methods that are feasible and useful to analyze existing healthcare databases to identify and evaluate safety and benefit issues of drugs already on the market.

Worldwide Safety Strategy

OMOP Data Community

OMOP Extended Consortium

OMOP Research Core Humana HSRC Partners HealthCare Regenstrief Research Lab

http://omop.fnih.org

Distributed Network SDI Health Centralized data GE Thomson Reuters VA

Worldwide Safety Strategy

Common data model role in OMOP Analysis process

Source 1 Source 2 Source 3 Transformation to OMOP common data model OMOP Analysis results Analysis method

http://omop.fnih.org

Worldwide Safety Strategy

Database model heat map

Database model is that

• f OMOP
• f OMOP

CDM

Shows how well different variables convert into a Common Data Model

Worldwide Safety Strategy

Performance characteristics of methods on THIN in OMOP CDM

Measure Threshold Sensitivity Specificity PRR PRR 95% LBCI >1 0.67 0.68 USCCS OR >1 and LBCI >1 0 78 0 59 USCCS OR >1 and LBCI >1 (a=0.05) 0.78 0.59 HDPS RR >1 and LBCI >1 (a=0.05) 0.50 0.76

Comparison against the OMOP reference set of established drug-event combinations1

1 Stang et al (2010). "Advancing the science for active surveillance: rationale

and design for the Observational Medical Outcomes Partnership." Annals of Internal Medicine 153(9): 600-606.

Worldwide Safety Strategy

Challenging Issues

• Optimal data set(s)s, or combinations thereof for

surveillance for specific medicinal products?

• Implications of different Data access approaches
• Under what circumstances do algorithmic methods

perform best and with what level of effectiveness?

• What performance characteristics and what processes
• What performance characteristics and what processes

are needed to effectively use novel surveillance strategies?

Worldwide Safety Strategy

Conclusions

• Multiple rich heterogeneous and intricately constructed

‘real world’ data sets of Electronic Medical Records and Transactional Claims databases

– Surveillance approaches are now being innovatively applied to such data S l i iti ti d t hi d i ti l – Several initiatives and partnerships doing essential foundational work in the field – Challenging how to determine how to best utilise this wealth of

data, and how to best incorporate such analyses into overall safety strategies

• Analysis of real world data is only one potential

component of an overall continual assessment

• f risk benefit