Control of internal validity threats in a modified adaptive platform design using Halili Physical Therapy Statistical Analysis Tool (HPTSAT)

Abstract

Some of the challenges in evaluating the effectiveness of physical therapy practice stem from the common lack of interventional standardization, as well as problems with both the availability and routine use of specific, valid outcome tools. But even if these issues are dealt with, there are still significant validity threats when trying to understand the effectiveness of physical therapy interventions. Among the most critical internal validity threats: repeated testing effect, study sample uniformity, and increase in type I or type II error. The purpose of this analysis is to illustrate how these internal validity threats were controlled using the Halili Physical Therapy Statistical Analysis Tool (HPTSAT). The original concept behind the HPTSAT tool was the adaptive platform design used in the PREVAIL I and II protocols(1,2,3). However, this concept has been significantly expanded upon in the HPTSAT design to allow for the simultaneous analysis of hundreds of treatment combinations while controlling for the above-mentioned critical internal validity threats. HPTSAT allows for concurrent computerized analysis modeled and improved upon the adaptive platform design used in the PREVAIL I and II protocols. This analysis is possible because the tool isolates the average rate of change (ARC5) instead of average change after treatment. This paper provides the mathematical basis for the algorithm used in the tool to control for several internal validity threats including repeated testing effect, study sample uniformity, and increase in type I and II error.