RESUMO
PURPOSE: Clinical reasoning, a complex construct integral to the practice of medicine, has been challenging to define, teach, and assess. Programmatic assessment purports to overcome validity limitations of judgments made from individual assessments through proportionality and triangulation processes. This study explored a pragmatic approach to the programmatic assessment of clinical reasoning. METHOD: The study analyzed data from 2 student cohorts from the University of Utah School of Medicine (UUSOM) (n = 113 in cohort 1 and 119 in cohort 2) and 1 cohort from the University of Colorado School of Medicine (CUSOM) using assessment data that spanned from 2017 to 2021 (n = 199). The study methods included the following: (1) asking faculty judges to categorize student clinical reasoning skills, (2) selecting institution-specific assessment data conceptually aligned with clinical reasoning, (3) calculating correlations between assessment data and faculty judgments, and (4) developing regression models between assessment data and faculty judgments. RESULTS: Faculty judgments of student clinical reasoning skills were converted to a continuous variable of clinical reasoning struggles, with mean (SD) ratings of 2.93 (0.27) for the 232 UUSOM students and 2.96 (0.17) for the 199 CUSOM students. A total of 67 and 32 discrete assessment variables were included from the UUSOM and CUSOM, respectively. Pearson r correlations were moderate to strong between many individual and composite assessment variables and faculty judgments. Regression models demonstrated an overall adjusted R2 (standard error of the estimate) of 0.50 (0.19) for UUSOM cohort 1, 0.28 (0.15) for UUSOM cohort 2, and 0.30 (0.14) for CUSOM. CONCLUSIONS: This study represents an early pragmatic exploration of regression analysis as a potential tool for operationalizing the proportionality and triangulation principles of programmatic assessment. The study found that programmatic assessment may be a useful framework for longitudinal assessment of complicated constructs, such as clinical reasoning.
RESUMO
INTRODUCTION: The routine use of a single aerobic bottle for blood culture in pediatric patients has become commonplace, as anaerobic bacteria are not frequently involved in clinically significant infections. The aim of this study was to assess the usefulness of routinely performing anaerobic blood cultures in pediatric oncology patients. METHODS: Prospective study was conducted on pediatric (<18 years) patients affected with febrile syndrome after receiving chemotherapy for hematological or solid malignancies. Samples were inoculated into pediatric aerobic and standard anaerobic bottles (BacT/Alert automatic system). Strains were considered clinically significant, or deemed as contaminants, depending on isolation circumstances and clinical criteria. RESULTS: A total of 876 blood cultures from 228 patients were processed during the 21-month study period (January 2014 to September 2015). Baseline diagnosis included 143 solid tumors and 67/18 cases of leukemia/lymphoma. Bacterial growth was detected in 90 (10.2%) blood cultures for 95 different isolates, of which 62 (7.1%)/63 isolates were considered clinically significant. Among the latter, 38 (60.3%) microorganisms grew in both aerobic and anaerobic bottles, 18 (28.6%) only in aerobic bottles, and 7 (11.1%) only in anaerobic bottles. Gram-negative bacilli (33; 52.4%), mainly from the Enterobacteriaceae family, were the most frequently isolated microorganisms. Overall, only 3 out of 90 isolates (3.3%) were strict anaerobes (Propionibacterium acnes), and all of them were deemed contaminants. CONCLUSION: Strict anaerobes did not cause significant infections in febrile pediatric oncology patients, and anaerobic blood culture bottles offered no additional advantages over aerobic media. Our results suggest that routine blood cultures should be solely processed in aerobic media in this group of patients.