From: Real-world evidence for coverage determination of treatments for rare diseases
Challenge | Details | Suggestions |
---|---|---|
Selection bias, confounding, and use of multiple clinical sites | • These challenges lead to limited generalizability for informing decisions • Rare diseases often have small, heterogeneous, and limited patient populations • Typical bias mitigation strategies are often limited in rare diseases because of feasibility constraints | • Consider how patients will be recruited into the study • Acknowledge and account for the main source of bias • Determine the statistical methods for adjusting for bias |
Historical/ external controls | • Historical cohorts are often used as the control arm in rare disease trials • Many entities grant approval or coverage based on these comparisons • These are problematic for addressing questions of comparative/relative effectiveness | • Be clear on the intended use of the RWE • Decide on study design considerations • Consider how patients will be recruited into the study |
Outcome selection and surrogate endpoints | • Rare diseases may lack standards for outcome measurements • MCID can be difficult to determine in rare diseases because effect sizes may not be precise and values in literature may be sparse • Surrogate endpoints can be useful but can be difficult to validate in rare diseases | • Identify appropriate outcomes and consider measurement practicality • Assess the overall feasibility of the study • Start with the research question of interest when selecting outcomes |
Length of study | • RWE is often considered as a means of filling evidence gaps, perhaps due to a need for long-term evidence • Length of study for rare diseases that are often chronic and/or have slow symptom onset may be limited by feasibility, cost, and risk of withdrawal | • Consider the overall length of the RWE study • Identify appropriate outcomes and consider measurement practicality |
Data quality | • RWE data are collected from registries, health insurance claims, EHRs, and other forms of data, all of which have potential for missing or incomplete information • RWE data is often collected for one reason and re-purposed for another • Incompleteness of data may necessitate linking between sources • Data quality issues can result in information bias | • Determine how data quality will be monitored |
Practical issues | • There are currently no global standards for use of RWE by payers • Roles and incentives of various stakeholders can be difficult to align • MEA and CED schemes may be difficult to implement for rare diseases due to significant methodological and implementation challenges | • Assess the overall feasibility of the study • Be clear on the intended use of the RWE • Clearly define stakeholder roles, investments, and involvement |
Generalizability & reproducibility | • Practical issues within jurisdictions are compounded when considered between jurisdictions • Differing global requirements may prohibit or limit use from country to country, which can result in additional required studies and further costs if studies are not generalizable | • Consider the points listed above in the context of each country of interest |