Predicting First-Time National Physical Therapy Examination Performance for Graduates of an Entry-Level Physical Therapist Education Program.

Abstract

The National Physical Therapy Examination (NPTE) is a standardized examination designed to assess competence after graduation from an entry-level physical therapist education program. Previous studies have identified applicant and student variables that are related to NPTE performance, with applicant variables reflecting performance before admission and student variables reflecting performance after admission. However, there are very few articles describing how these variables can be combined to predict NPTE performance. The purpose of this study was to develop, evaluate, and describe models to predict first-time NPTE scores and NPTE outcomes (pass vs fail), based on various applicant and student variables related to academic performance. Pre- and postadmission data and NPTE scores were recorded for 185 individuals who graduated from an entry-level physical therapist education program. Multiple linear regression was used to develop a model to predict NPTE scores, and binary logistic regression was used to develop a model to predict NPTE outcomes (pass vs fail). A model including undergraduate prerequisite grade point average, grade point average in basic science courses taken during the program, and comprehensive examination scores combined to explain 30.9% of the variance in NPTE scores and accurately predicted NPTE outcomes (pass vs fail) 81.1% of the time. In general, our findings support the notion that prediction of NPTE performance should be based on a combination of applicant and student variables. The models described in this article could be used to identify students who may be likely to struggle on the NPTE, making it possible to provide additional support to these students. Various applicant and student variables related to academic performance can be combined to predict NPTE performance. The results of this study provide a framework for programs interested in applying models to predict NPTE performance.