NOVEL INDIRECT COMPARISON METHODOLOGY FOR ESTIMATING TIME-DEPENDENT RESPONSE TO ANTIMUSCARINICS FOR THE TREATMENT OF OAB
Author(s)
Snedecor SJ, Sudharshan L
Pharmerit International, Bethesda, MD, USA
Presentation Documents
Background: Common indirect treatment comparison (ITC) methodology in overactive bladder involves combining absolute reduction in urge urinary incontinence (UUI) episodes at study endpoint (e.g., week 12) to estimate the overall treatment effect. Trials with differing endpoints must assume equivalence to be included in the network. Further, analyses of endpoint data are not sufficient to predict efficacy at intermediate time points (e.g. 4 or 6 weeks). We developed and tested an alternate methodology to utilize available intermediate time points into an ITC of published studies of fesoterodine and tolterodine. Methodology: Study-level mean UUI reduction over time can be represented as the percent reduction from baseline, which can be modeled as a monotonically-increasing function with a theoretical maximum of 100%. This function is expressed with two parameters: %red = bi*time/(ci + time), where bi is the maximum possible reduction for treatment i, and ci is the time required to reach half the maximum reduction. The inverse %red is a linear function of 1/time that can be used within a Bayesian ITC framework to generate a placebo-adjusted indirect comparison of efficacy. Conclusions: The endpoint results obtained from the alternate methodology were comparable to those obtained from an endpoint ITC. This novel methodology has the additional advantage of utilizing all available time point data within a single analysis, which can then be used to generate efficacy estimates at intermediate time points, which may be utilized within economic models. Limitations include unavailability of uncertainty estimates of the %red variable and difficulty of estimating combinations of parameters within functional constraints. Finally, our alternate methodology may be used for any longitudinal data exhibiting a monotonic increase or decrease and may be expanded to include a network with multiple treatments.
Conference/Value in Health Info
2014-11, ISPOR Europe 2014, Amsterdam, The Netherlands
Value in Health, Vol. 17, No. 7 (November 2014)
Code
PRM236
Topic
Methodological & Statistical Research
Topic Subcategory
Confounding, Selection Bias Correction, Causal Inference
Disease
Urinary/Kidney Disorders