1 Angela Vinken 19th October 2016Successfully Integrating EHR and EDC Data: Putting the ‘Real’ in Real-World Data Angela Vinken 19th October 2016

2 Discussion Topics How companies can successfully navigate the EHR and secondary data landscape What data sources are available and what benefit will they bring to your study and your strategy? Designing and implement studies that reap the benefits of combining EDC and EHR data to answer key questions Transforming the theory of real-world hybrid studies into reality. 2

3 Strong Design is Key to Study SuccessOBJECTIVES OPERATIONAL APPROACH Regulatory mandate? Information for payers? Assess standard of care? Publications? Achieve a leadership position? Why are we conducting this study? Optimal Study Design Optimal Operational Model Optimal Study Budget Sites easily engaged Minimised Scope and Budget Adjustments Positive Site Experience + = Effectiveness? Safety? Adherence? Quality of Life? Health resource utilization? Prospective, retrospective, or a combination? What type of Information will be collected? What data sources are available? Real-world practitioners? KOLs? Major institutions, clinical, managed care, etc.? Secondary data sources? 3

4 What is a Hybrid Study? Why Do They Matter?EDC is the “gold standard” EHR adoption/ diffusion is progressing rapidly Secondary data sources are becoming more common and available EDC data can be very thorough, but require monitoring & querying EHRs are the source data, but are not always complete Secondary data sources promise rich data, but are not always the best fit Hybrid Study EDC EHR Claims Data Secondary Data PROs 4

5 Generate Evidence Through Global Data SourcesOPTUM®, Anolinx™, IMS™, Truven® APAC: ASPEN, Korea, Singapore, Taiwan, China, Japan UK: CPRD, NW e-Health, SHIP, SAIL, Farr Institute US: VA, Explorys/IBM® Watson Healthcare, Geisinger, PACER, Kaiser, Aetna®, OHDSI. EU: NORDICs, Italy, Netherlands, IMI EMIF (emerging), EU-ADR network  South America – Brazil Public and Private National health insurance databases Understand which data sources to use and when Be able to aggregate data from multiple sources Efficiently analyze data and communicate results 5 5

6 Data Availability and Linkage are KeyFull EHR All clinical data in EHR system and can be linked Separate data collection (through ed links/call center) Samples to be sent to “central” lab/storage EHR / EDC Combination Some clinical data in EHR system. For critical fields, physician enters data into EDC Full EDC EHR data not available– physician needs to enter all data into EDC Clinical PRO Bio- Bank 6

7 Linking Data Assets offers the MOST -Comprehensive View of the Patient ExperienceEHR Clinical Data Healthcare Utilization & Cost Data from Claims Other Secondary Data Physician-Reported and Patient-Reported Measures 7

8 Advantages of a Hybrid ApproachResearch Challenges Hybrid Solutions Site-based evidence collection is relatively expensive EHR/claims data are not collected for research purposes and can lack consistency or particular data of interest Typical EDC data collection can be burdensome to the site Site-based patient recruitment is slower and less predictable EHR/claims data can be much less expensive Site-based evidence collection can target key data points where consistency or access to specific data is critical Accessing data via EHR can significantly reduce site burden and increase site satisfaction Improved recruitment predictability and speed by selecting appropriate sites with relevant patients EHR/Claims combined with global site-based data can capture the best of both approaches 8

9 Retrospective EHR Data Extraction Continuous EHR Data ExtractionCASE STUDY: Hybrid Strategy to Build a Cost- Effective Registry Solution National Registers (where available) Background Data Retrospective EHR Data Extraction Comprehensive Clinical Registry Historic Patient Data Site EHR Database Continuous EHR Data Extraction Participating Site Structured/Routine Clinical Data CRF Critical Data Points Biomarker Testing Pt Biomarker Data PRO Data Pt Reported Data Patient 9

10 What Advantages will a EDC/EHR Hybrid Study Offer?Reduced Cost Reduced monitoring and oversight Reduced site payments Less programming and data-management cost Faster recruitment via improved site/patient selection High(er) Quality Data Integrated “source data” from EHR Greater completeness, rigorous edit checks, and additional cleaning for critical data Increased Site Satisfaction Less data-entry burden for site Fewer queries and less data review/cleaning Easier patient recruitment via improved site/patient selection Strategic Benefit Ability to answer different – and more – questions within one study Ability to better understand practice patterns, site-level insights 10

11 What are the challenges around accessing and integrating EHR and claims data? How can these be overcome?

12 Integrating EHR, Claims, and EDC Data Can Be ChallengingAccess to data varies by region Access is much more difficult ex-US: data-privacy regulations, system variability Claims, EHR, and EDC data are very different Need to understand what benefit each brings to the study, and how best to utilize them Technology solutions still developing Data volume, variability, and refresh-frequency can impact integration

13 Claims Offer Breadth…EHR Offers DepthIntegrating EHR, Claims, and EDC Data Can Be Challenging Claims Offer Breadth…EHR Offers Depth Relevant information Claims EHR Large population denominator  Longitudinal data, inpatient and outpatient follow-up Partial Sequence of treatment regimens All physician and facility services (e.g., testing), filled retail/specialty pharmacy drugs; total reimbursed amounts Conformity to practice guidelines Partial GAP Disease progression and severity Date of disease onset Laboratory results Height, weight, BMI, blood pressure, smoking status and history GAP Survival Severity of adverse events Clinical characteristics (e.g., affected joint count, reported symptoms, tumor histology, stage at diagnosis) Patient-reported outcomes (general: SF, EQ5D; disease specific Hamilton Depression Rating Scale, Beck Depression Inventory. FLIE, CAT, PASI) Socioeconomic status data Health risk assessment (smoking status, alcohol use, exercise, behavior change readiness)

14 Integrating EHR, Claims, and EDC Data Can Be ChallengingChallenge: Are the Required Outcome-measures Data in the EHR? Solution: In-depth Examination of EHR to Inform Protocol & Study Design Stage Pre- Study Design and finalize protocol and study operations Analysis type Feasibility analysis based on EHR and/or administrative claims Key questions Do the inclusion/exclusion criteria in the protocol allow enough eligible patients to conduct the trial quickly? What is the demographic profile of patients? What are current patterns of treatment and outcomes? Does the proposed design align with the current treatment practices? Frequency and completeness of data capture in the EHR Benefits Identify unrealistic protocol criteria and refine protocol objectives Plan more sites when patient counts are low Identify top backup sites if enrollment falls behind Decrease attrition rates Identify which data need to be captured via DataLabs® EDC vs. EHR Optum’s existing data streams however may not only be used as a means of collecting outcomes of interest, prior to initiating a study, the data may be used to inform the protocol Potential for improving recruitment rates and ultimately the success rate of PCTs

15 Key Considerations with EHR-linked Approach: Data Availability Overall & By Specific Sites/IDNsComorbidities CAD MI Stroke CHF Depression Diabetes Hypertension Osteoporosis Osteopenia Written Rx (Written Rx Date, Class-Level Med) Non-biologic DMARDs Biologic DMARDs Corticosteroids Labs Diagnosis & Prognosis ESR (Erythrocyte sed. rate) RF (Rheumatoid Factor) Anti-CCP (Anti-cyclic citrullinated peptide Ab) Cardiometabolic Risk HDL LDL Liver & Renal Function Albumin ALT/AST eGFR Vitals/Observations SBP/DBP BMI Smoking status Exercise levels Pain Age Gender Race Household Income Education status Payer Type Demographics Specialty (Rheum., IM, FP, or Mid-Level) Patient volume (average # of patients seen Outpatient Visits/Provider Attributes Visits & LOS Diagnoses Inpatient Radiography Ultrasonography, MRI Bone Density TEE Procedures Data Assessment % Provider groups who capture all of the measures listed in each of the data categories Demographics: XX% Comorbidities: XX% Patient Visits: XX% Written Rx: XX% Labs: XX% Observations: XX% Procedures: XX% Partner Evaluation Provider Feedback 15

16 Study Data Analysis to Optimize Collection StrategyConcept Secondary Data Source Primary Data Source EHR Database Claims Database Physician-Reported Patient-Reported Inclusion/exclusion criteria  Study drug initiation Medical history Patient demography Clinical characteristics Survey/PRO measures Treatment compliance and outcomes Rating scale Healthcare utilization and costs Safety assessment 16

17 What are the critical success steps TO integratE EHR and EDC data within a real-world study?

18 Leveraging Real-World Data AssetsStep 3: Determine Real-World Study Types Step 4: Develop Optimal Operational Design EHR and Secondary Data Assets, PROs, Site-based EDC Database Extraction Traditional Pragmatic/NIS Study “Hybrid” Study Step 2: Assess Real-World Data Options Step 1: Define Goals and Data Needs Think Strategically Throughout the Process

19 Study-Design Decision TreeDefine Goals & Objectives Effectiveness Efficacy Safety Internal Decision Support Determine What Data Are Needed Clinical Economic Humanistic Identify Best Source of Data Patient- / Caregiver-reported Physician-reported Claims Other secondary data EHR Determine Best Way to Collect Data PRIMARY: Survey, eCRF SECONDARY: Claims, EHR, other existing data With our access to data and Parexel’s exceptional site-based observational research capabilities we are able to truly follow the scientific method and allow the research question to drive the study design What is the best source of data for each of the critical elements of the study What is the best way to collect it 19

20 Planning is Key: Finding the Right Data, Access, & TechnologyDefine Clearly identify the objectives – what questions are you hoping to answer? Be realistic and practical Engage all stakeholders to ensure all perspectives and needs are met Decide Thoroughly assess available assets – which fit best? Assess data availability early – what can you truly access? Identify structural / technology platforms and limitations Understand regional, monitoring, and other operational requirements DECIDE: Is a hybrid approach best for this study? Develop Secure access to all data sources Develop technology and operational platforms Establish a streamlined management team and model

21 Hybrid Approaches Offer the Potential to Resolve Critical Evidence-generation IssuesResearch Challenges Site-based evidence collection is expensive EMR/claims data are intended for non-research purposes and can lack consistency or particular data of interest Patient recruitment timelines are challenging to predict Hybrid Solutions EHR/claims data can be much less expensive (at scale) Site-based evidence collection can target key data points where consistency/existence is critical to the study Improved predictability/speed by selecting sites with high volumes of patients By carefully managing which data elements are collected via EHR vs. EDC, hybrid studies can provide a better approach to research

22 Questions & Comments?

23 Thank You