Adaptation and factor structure of three psychometric instruments measuring behavioral aspects of medication prescribing in physician assistants.

OBJECTIVES: (1) Present the factor structure of two psychometric instruments for self-efficacy and one for outcome expectations of medication prescribing; (2) evaluate the reliability of the scales, and (3) present preliminary evidence of validity. METHODS: Physician assistants (PA) and PA students completed a survey evaluating three psychometric instruments: (1) Self-Efficacy in Prescribing (SEP), (2) Self-Efficacy in Prescribing-Geriatric (SEPG), and (3) Outcomes Expectations of Prescribing Errors (OEP). Students also evaluated 3 hypothetical prescriptions, two of which contained a prescribing error. Students were instructed to identify (1) if an error occurred and (2) what type of error. The data were analyzed using parallel analysis with a varimax rotation, Cronbach's α, Pearson and Spearman correlations. RESULTS: One hundred eighty five (n = 185) respondents completed the survey (response rate = 63.8%). The parallel analysis found that the SEP had one 7-item factor with α = 0.94 (M = 5.7 (SD = 1.9) out of 10). The SEPG also had one 7-item factor with α = 0.95 (M = 5.5 (1.9). The OEP had one 6-item factor with α = 0.89 (M = 3.5 (SD = 0.8) out of 5). The SEP and SEPG, were correlated to the OEP each other (both p < 0.01). Actively practicing PAs had the highest composite mean SEP and SEPG scores. First-year PA students had the highest mean scores for the OEP. There was a weak association between the mean SEPG score and the number of correctly identified prescriptions (rs = 0.18, p = 0.04). CONCLUSION: The SEP, SEPG, and OEP show preliminary evidence of reliability and structural, construct, and known-group validities using simulated prescriptions. These tools may be able to be used by educators and implementation scientists as one method to show the effectiveness of future interventions to reduce incidence of prescribing errors.

Are Physician Assistants Able to Correctly Identify Prescribing Errors? A Cross-Sectional Study.

PURPOSE: (1) To describe how often physician assistant (PA) students correctly identify prescribing errors and (2) examine between-cohort differences on ability to correctly identify prescribing errors. METHODS: This was a cross-sectional study of 2 cohorts of PA students at one institution. Students were presented with 3 hypothetical prescriptions, 2 of which contained a prescribing error. For each prescription, students were asked to (1) identify whether an error occurred and (2) indicate the type of error. A simple Poisson regression model analyzed the data. RESULTS: We received responses from 130 students (72.6% response rate). Approximately 12% (12.3%, n = 16) correctly identified whether all 3 prescriptions were correct. The median number of correctly identified prescriptions was 1 (interquartile range = 1). There was not a statistically significant between-cohort difference identifying the correct number of prescriptions (ß = 0.27, P = .10). CONCLUSION: Physician assistant students' prescribing error identification was similar to previous research in medical and nursing students. Efforts to improve prescribing training are critical to ensure patient safety.