We used a cross-sectional design to describe knowledge in managing musculoskeletal conditions among physical therapist students and licensed physical therapists in the uniformed services. The study was approved by the Institutional Review Boards at Wilford Hall Medical Center in San Antonio, TX and at Texas Tech University in Lubbock, TX before subject recruitment and data collection began. All subjects provided informed consent prior to participation.
Based on an a priori sample size estimation, a total of 26 first-professional physical therapy programs accredited by the Commission for Accreditation of Physical Therapy Education were randomly selected for participation. Educational programs are rapidly transitioning to doctoral programs, with approximately 80% of programs having completed the transition to the Doctor of Physical Therapy degree or in the transitioning process at the time of the study . Therefore, randomization was blocked by the degree to be conferred upon graduation: master's (n = 13) vs. doctoral (n = 13) to permit comparisons based on degree status. Program directors were contacted initially by email to inform them of their program's selection, describe the study procedures, and invite the program's participation. Students in these programs were in the terminal phase of their program's curriculum, defined as having completed substantial portions of the didactic curriculum and clinical affiliations. All licensed physical therapists in the four uniformed health services (U.S. Air Force, U.S. Army, U.S. Navy, and U.S. Public Health Service) with at least one year of clinical experience were also invited to participate.
Participants completed the identical examination originally developed by Freedman and Bernstein to assess knowledge in musculoskeletal medicine among physician interns,  and more recently administered to medical students, residents, and a variety of physician specialists . The examination consists of 25 open-ended questions that were selected based on commonly encountered musculoskeletal diagnoses encountered in the primary care setting (ie, fractures and dislocations, low back pain, sciatica, and arthritis) and consideration of orthopaedic emergencies that warrant immediate referral to an orthopaedic surgeon or the emergency department (ie, compartment syndrome, hip dislocation, etc.) . Additional details related to the development and validation of the examination are reported elsewhere [7, 27].
The examination was administered in a web-based format using Web Surveyor, version 3.6 (Web Surveyor Corporation, Herndon, VA). No time limit was imposed to be consistent with previous methodology [7, 12]. Subject confidentiality was strictly maintained through assignment of a unique computer-generated code. Administration of the examination was preceded by a brief demographic survey that queried patients as to their educational background, board-certification status (Orthopaedic and/or Sports Clinical Specialist designation via the American Board of Physical Therapy Specialties), experience in different practice settings, and familiarity with the studies by Freedman and Bernstein [7, 27]. Data from any therapists familiar with the studies by Freedman and Bernstein [7, 27] were excluded from analysis because the examination questions and answer key were published verbatim in these manuscripts.
Educational programs were requested to have participants complete the study in a group setting with a proctor present (eg, a computer lab) when possible. This would insure that participants did not use any outside resources (ie, textbooks, information available on the internet, personal communication, etc.) to assist them in answering the questions. To maximize participation, however, programs were alternatively given the option to have participants complete the study on their own if a computer lab or similar arrangement was unavailable, or if a participant was not available at the designated time. Licensed physical therapists were also asked to complete the study in a proctored setting. All participants were queried at the end of the study as to whether they used any outside resources to assist them in the completion of the examination. The results of the demographic survey and content of the examination were stored in a secure, password-protected centralized database for subsequent analysis.
A total of 6 judges, blinded to the demographic survey results and whether the participant was a physical therapist student or licensed physical therapist scored blocks of 4–6 questions, resulting in each question being scored by two raters. Judges were physical therapist faculty with considerable experience in providing direct access care for patients with musculoskeletal conditions. Each rater was also trained in the scoring procedures by one of the investigators. An overall score and passing rate were determined using identical procedures as those described by Freedman and Bernstein, however a brief review is provided here. Each question was assigned a maximum possible of 1 point. Partial credit was assigned based on the criteria for partial credit outlined in the answer key . Scores were not penalized for incorrect spelling. Sums of individual scores represented the overall score, which was multiplied by 4 to obtain a percentage score. Inter-rater reliability of the overall score was examined using the intraclass correlation coefficient (ICC), equation 3,1 . The ICC and associated 95% confidence interval was 0.91 (0.89, 0.92). Given a sufficiently high reliability coefficient, only data from the first rater were used in the analysis. Using the results from a single rater is also consistent with the procedures utilized by Freedman and Bernstein . Participants were judged to have passed the examination if their score exceeded the previously established threshold of 73.1% .
Descriptive statistics, including frequency counts for categorical variables and measures of central tendency and dispersion for continuous variables were calculated to summarize the data using SPSS for Windows 11.0.1 (Chicago, IL). Independent sample t-tests were used to directly compare differences in knowledge between educational programs conferring the doctoral versus master's degree and between licensed physical therapists who were board-certified and those who were not. Differences in the passing rates among the physical therapist subgroups were examined using the Pearson chi-square statistic. The alpha-level was established a priori to be 0.05 utilizing a two-tailed test.