Use of pre-admission quantitative measures data for a physical therapy program to predict performance on the national physical therapy licensure exam.
A primary outcome of concern to the physical therapy (PT) program is the ability of graduates to pass the national physical therapy exam (NPTE) for licensure. For each admitted student who completes the program but is unable to pass the NPTE after two attempts, the chances of ever obtaining licensure drops precipitously. We wanted to identify any pre-admission predictive variables that would increase the probability that a graduate would be successful in obtaining licensure, and predict which applicants may have a substantial risk of not becoming licensed, even if they are successful in completing the PT program. The purpose of this retrospective cohort study was to identify the predictive strength of pre-admission (pGPA), and the three Graduate Record Exam (GRE) test scores for graduates’ ability to eventually pass the NPTE. We planned to incorporate those pGPA and/or GRE pre-admission cutoff scores data to establish or confirm minimum GRE or pGPA thresholds for applicants to the Doctor of Physical Therapy program.
Pre-admission GPA, GRE analytic (aGRE), GRE verbal (vGRE) and GRE quantitative (qGRE) scores were collected on eleven years of graduates from a single MPT program (n=320). The Federation of State Boards of Physical Therapy provided NPTE data on all graduates. Pre-admissions data on those who failed to complete the program were excluded. Ability of each pre-admission variable to predict NPTE passing within two attempts was assessed by calculating sensitivity, specificity, likelihood ratios (LR), and relative risk. ROC curves were developed to determine optimal cutoff scores of each variable, and to calculate the area under the curve (AUC) to assess the predictive value of pGPA and each GRE score.
36 (11%) of graduates failed to pass on the first attempt, and 15 (4.7%) required two or more attempts to pass the NPTE. The optimal cutoff score for qGRE was 29th percentile (Sensitivity = 75%, specificity = 87%, LR+ = 5.66, LR- = 0.28, AUC = 0.81). For vGRE, the optimal cutoff was also at the 29th percentile (Sensitivity = 71%, specificity = 67%, LR+ = 2.12, LR- = 0.44, AUC = 0.71). For the aGRE the optimal cutoff was 24th percentile (Sensitivity = 80%, specificity = 53%, LR+ = 1.71, LR- = 0.38, AUC = 0.67). For pGPA, the optimal cutoff was 3.31 (Sensitivity = 68%, specificity = 53%, LR+ = 1.45, LR- = 0.60, AUC = 0.69).
Conclusions/Relevance to the conference theme: Shaping the Future of Physical Therapy Education
The only pre-admission variable that had a LR+ of moderate effect was the qGRE using a cutoff score of 29th percentile. Scores above the cutoff for pGPA, aGRE and vGRE had a small or negligible effect on predicting the ability to pass the NPTE within two attempts. Given the imperative to establish evidence-based admissions criteria in order to minimize the risk of graduates failing to obtain licensure, the admissions committee reaffirmed the use of a minimum aGRE score of 3.5, and instituted the minimum threshold of 29th percentile for the qGRE score as a qualifying requirement for physical therapy program applicants.
Utzman RR, Riddle DL, Jewell DV. Use of demographic and quantitative admissions data to predict performance on the national physical therapy examination. Physical Therapy. 2007;87:1181-1193.
Shiyko MP, Pappas E. Validation of pre-admission requirements in a doctor of physical therapy program with a large representation of minority students. Journal of Physical Therapy Education. 2009;23(2): 29-36.