Biometrical methods for analysis of adverse events in benefit assessment
This method improves how side effects of drugs are analyzed in health technology assessments by using survival time techniques instead of simple counts. It accounts for differences in how long patients are followed and when they stop treatment, which helps avoid misleading conclusions about a drug's safety and benefit.
At a glance
Use when
Analyzing adverse events in HTA submissions where follow-up times differ between groups or when censoring due to treatment discontinuation, switching, or noncompliance is present
Avoid when
When only aggregated count data without timing information are available, or when events are extremely rare and time-to-event modeling is unstable
Inputs
Adverse event data with timing and censoring information (e.g., time to event, treatment discontinuation, switching, noncompliance), treatment group assignments, follow-up durations
Outputs
Time-to-event estimates for adverse events (e.g., hazard ratios, survival curves), adjusted safety profiles accounting for unequal follow-up and censoring
How it works
The method addresses limitations of standard contingency table analyses and incidence rates in assessing adverse events, particularly when follow-up times are unequal or censored due to treatment discontinuation or switching. It advocates for the use of survival time methods (e.g., time-to-event analysis) to properly account for time dependencies and variable observation periods, ensuring more valid inference in drug benefit assessments, especially in regulatory and reimbursement contexts such as in Germany.
- HTA domains
- Clinical Effectiveness, Safety
- Categories
- AppraisalHeterogeneity
- Assumptions
- Adverse events are time-dependent; censoring is non-informative; appropriate follow-up time data are available; events can be reliably recorded over time
- Strengths
- Accounts for variable follow-up times and censoring due to treatment discontinuation or switching,Reduces bias compared to simple proportions or crude incidence rates,Aligns safety analysis with rigorous time-to-event methodology used in efficacy assessment
- Limitations
- Requires detailed longitudinal data, which may not always be available in reimbursement dossiers,More complex to implement and interpret than simple frequency tables,Assumes non-informative censoring, which may not hold in all clinical scenarios
- Also known as
- Survival time methods for adverse event analysis, Time-to-event analysis for safety in HTA, Censoring-aware adverse event analysis
Questions this answers
- › How can adverse events be analyzed more accurately when patients have different follow-up times?
- › Why are simple proportions misleading in safety assessments of drugs?
- › What statistical methods should be used to account for censored data in adverse event reporting?
- › How do safety analyses in benefit assessment differ from those in drug approval?
- › What are the consequences of using inadequate methods for adverse event analysis in reimbursement dossiers?
- › How can survival analysis improve the credibility of safety evidence in HTA?
References & sources
Similar by meaning
- Estimands framework for adverse-event analysis with varying follow-up (benefit assessment)
- SAVVY (Survival analysis for AdVerse events with VarYing follow-up times)
- Fractional polynomial network meta-analysis of survival data
- NICE DSU Technical Support Document 14
- Change-point survival models for relative treatment effect extrapolation
Beta record. Generated from the primary source via AI extraction and independent audit, pending final human review.