Related Experiment Videos
Problems in the use of large data sets to assess effectiveness.
International Journal of Technology Assessment in Health Care
|January 1, 1990
Summary
Large data sets offer valuable insights for outcome assessment but present challenges. These retrospective analyses require further clinical trials to confirm findings due to potential biases and omitted details.
Area of Science:
- Health Informatics
- Clinical Research Methodology
- Biostatistics
Background:
- Large datasets are increasingly utilized for outcome assessment.
- However, their application carries inherent risks and limitations.
- Retrospective and unblinded nature of database evaluations poses challenges.
Purpose of the Study:
- To highlight the challenges and risks associated with using large datasets for outcome assessment.
- To emphasize the limitations of database evaluations in establishing causality.
- To underscore the necessity of follow-up clinical trials.
Main Methods:
- The study reviews the characteristics of large dataset evaluations.
- It identifies common issues such as multiple analyses and endpoints.
- It discusses the difficulty in ascertaining analytical procedures and choices made.
Main Results:
- Database evaluations are retrospective and unblinded.
- They often result from multiple analyses of multiple endpoints.
- Critical details regarding analytic procedures are frequently omitted, hindering reproducibility.
Conclusions:
- Large datasets can generate hypotheses and suggest potential outcomes.
- However, database analyses alone are insufficient for proving outcomes.
- Clinical trials are essential to validate findings derived from large data analyses.