In this webinar our speakers described a new statistical method, named ‘generalized pairwise comparisons (gpc)’ for the assessment of treatment benefit/risk.
Erratum to the recording: Please note there is mistake in the recorded webinar on slides 45 and 46. The column labels for adverse events are reversed, and the toxicity was higher for FOLFIRINOX, the arm with higher efficacy. As shown in slide 47, FOLFIRINOX has a net benefit that is significantly favorable regardless of the choice of threshold until 6 months, because the higher toxicity is not sufficient to offset the benefit.
IN THIS WEBINAR:
- We started by reviewing commonly used metrics to assess treatment benefit, especially considering survival endpoints.
- They then described the Generalized Pairwise Comparisons GPC statistical method as a novel approach to the analysis of clinical trial data, providing examples of recent studies in which this method has been used and has shed new light in the assessment of treatment benefit.
- Finally, they described ongoing efforts to enhance this methodology and make it available to end-users, such as clinical trialists, physicians, and patients.
ABSTRACT:
The management of cancer patients in clinical practice relies to a great extent on the results of clinical trials. Although several statistical methods are currently used to test the effects of treatments in clinical trials, all of them suffer from limitations. Moreover, it is not standard practice to take into account multiple endpoints to formally assess treatment results under a single metric. Very often, a new treatment yields positive results for the primary endpoint, but not for all secondary endpoints. Also, benefits from new treatments typically have to be assessed against the potential harms from these treatments. Not incorporating the results of these other endpoints may lead to an incomplete view of the benefit/risk ratio. Despite the promise of precision medicine, treatment choices could be taken one step further if patient preferences were taken into account in decision-making using a formal statistical framework. This might be called a truly “personalized medicine”. We have proposed a new statistical method, named generalized pairwise comparisons (GPC), which can be useful in the search for personalized treatment choices.
SPEAKERS:
