Table 4. Considerations in the generation of high-quality real world data.
| Stage of the
RWE journey |
Considerations |
|---|---|
| Planning | • Understand the needs of local healthcare decision-makers
• Collaborate with external experts for advice on study designs and access to real world data • Generate a comprehensive RWE study (and publications) plan aligned to company strategy and local evidence needs • Be aware of local limitations: constraints in budget, time, drug exposure; logistics and study delivery; availability/willingness of patients and investigators to participate in the study • Identify a priori the potential sources of bias or confounding factors and identify measures to minimize them • Determine if data are required from a single country or multiple countries • Define research question a priori following FINER and PICO criteria 33 |
| Generation | • Select the most appropriate study design and data source to address the research question,
considering the strengths and limitations of each: – primary versus secondary data collection – prospective, retrospective or hybrid approach – randomization – descriptive versus analytic – cohort, case-control or cross-sectional study • Evaluate the benefits, risks and consequences to healthcare decision-makers of the selected study design and data source |
| • Clearly define the primary and secondary objectives or endpoints
• Evaluate the potential for missing data | |
| • Train internal and external study teams and investigators
• Assess availability of suitable data extraction, management and analytics resources • Follow the FAIR data principles when appropriate (e.g. when machines are used to find and use reusable data) 26 • Evaluate site monitoring, source data verification and quality aspects in primary data collection studies | |
| Communication | • Report methodology and data sources
• Ensure transparency in publications strategy • Commit to publication, regardless of the results • Follow best practice guidelines and recommendations (e.g. STROBE, MOOSE, RECORD) 27– 29, 34, 35 |
FAIR, Findability, Accessibility, Interoperability and Reusability; FINER, Feasible to answer, Interesting, Novel, Ethical, Relevant; MOOSE, Meta-analysis Of Observational Studies in Epidemiology; PICO, Patients, Intervention, Comparators, Outcomes; RECORD, REporting of studies Conducted using Observational Routinely-collected health Data; RWE, real world evidence; STROBE, STrengthening the Reporting of OBservational studies in Epidemiology.