To clarify the complexity arising from pitfalls and assumption violations in disproportionality analysis (DPA) and propose methods for improvement.
Approach:
Case Studies Review: Revisits two case studies from Fusaroli et al. to illustrate the interconnectedness of pitfalls in DPA and their implications.
Regression-Based Strategies: Highlights the use of regression-based strategies, particularly the regression-adjusted gamma-Poisson shrinker (RGPS), to refine associations in DPA.
Key Findings:
RGPS adjusts for co-reported products and covariates, improving the accuracy of disproportionality measures.
Masking can conceal true drug-event associations, which RGPS can help unmask by including relevant non-target drugs.
Confounding factors can distort observed drug-event associations, and RGPS addresses this by conditioning on co-reported drugs.
Interpretation:
RGPS provides a more nuanced understanding of drug-event associations by accounting for confounding and masking, enhancing the reliability of DPA results.
Limitations:
The analysis relies on the availability of comprehensive reporting data.
Potential biases in reporting, such as notoriety bias and solicited reports, may still affect results.
Conclusion:
The commentary highlights the importance of advanced analytical methods in addressing common pitfalls in DPA, advocating for the use of RGPS to improve signal detection.
Dana-Farber Cancer Institute's Dr. Jennifer Ligibel shared updated results from the BWEL study showing a weight loss intervention led to significantly better physical function, global physical and mental health, and symptom in patients with early breast cancer.
