Stacy A. Poe

Diagnostic errors by radiology residents in interpreting pediatric radiographs in an emergency setting

BackgroundThere are few data regarding the frequency and type of diagnostic errors made by radiology residents and fellows (“trainees”). However, increasing interest in reducing medical errors highlights the need to analyze which areas of medical knowledge are most problematic for physicians-in-training, including radiology trainees. Once these areas are identified, they can be emphasized during training.ObjectiveTo quantify the diagnostic errors made by radiology trainees interpreting radiographs from a pediatric emergency department.Materials and methodsA total of 23,273 dictations of emergency radiographs performed over a 1-year period at a pediatric hospital were analyzed for corrections after staff interpretation and for type and incidence of missed abnormalities by radiology trainees. Errors were categorized by type of pathology and anatomic region.ResultsOf the 80 errors detected, 90% were false negatives and 69% were recurrent. Most errors (69%) involved the diagnosis of fractures and/or dislocations. Sixty-one percent of all recurrent errors involved buckle, Salter II, avulsion, and transverse fractures—yet these cases constituted only 3% of all cases seen during the study period.ConclusionThe most common errors made by radiology trainees can be identified. By targeting these errors, training programs can improve the quality and relevance of the education they provide.

Trial of metronidazole vs. azithromycin for treatment of cyclosporine‐induced gingival overgrowth

Abstract:  Gingival overgrowth usually characterized by increased cellular growth of gingival fibroblasts appears to be multifactorial. In patients receiving CyA for more than 3 months, the incidence can approach 70% and can be attributed to pharmaceutical immunosuppression. Case reports have reported regression of overgrowth with both metronidazole and azithromycin. The goal of this study was to determine the efficacy of metronidazole and azithromycin in reducing CyA‐induced gingival overgrowth. Twenty‐five patients were included in this double‐blinded randomized study. All patients were receiving CyA as medically indicated and diagnosed with gingival overgrowth by a dentist. Patients were randomized to receive either 5‐days of azithromycin or 7‐days of metronidazole given at baseline only. The extent of gingival overgrowth was measured at 0, 2, 4, 6, 12, and 24 wk. Fourteen patients at CCF and 11 patients at CCHMC were studied. Repeated measures anova was performed to assess differences within and between groups. Gingival overgrowth at baseline was not statistically different between groups. The mean degree of gingival overgrowth after treatment was different across all time intervals (p = 0.0049) showing azithromycin to be more effective than metronidazole. Therapy with azithromycin offers an effective alternative to the management of CyA‐induced gingival overgrowth.

Upper Airway Volume Segmentation Analysis Using Cine MRI Findings in Children with Tracheostomy Tubes

Objective The purpose of this study is to evaluate the airway dynamics of the upper airway as depicted on cine MRI in children with tracheotomy tubes during two states of airflow through the upper airway. Materials and Methods Sagittal fast gradient echo cine MR images of the supra-glottic airway were obtained with a 1.5T MRI scanner on seven children with tracheotomy tubes. Two sets of images were obtained with either the tubes capped or uncapped. The findings of the cine MRI were retrospectively reviewed. Volume segmentation of the cine images to compare the airway volume change over time (mean volume, standard deviation, normalized range, and coefficient of variance) was performed for the capped and uncapped tubes in both the nasopharynx and hypopharynx (Signed Rank Test). Results Graphical representation of the airway volume over time demonstrates a qualitative increased fluctuation in patients with the tracheotomy tube capped as compared to uncapped in both the nasopharyngeal and hypopharyngeal regions of interest. In the nasopharynx, the mean airway volume (capped 2.72 mL, uncapped 2.09 mL, p = 0.0313), the airway volume standard deviation (capped 0.42 mL, uncapped 0.20 mL, p = 0.0156), and the airway volume range (capped 2.10 mL, uncapped 1.09 mL, p = 0.0156) were significantly larger in the capped group of patients. In the hypopharynx, the airway volume standard deviation (capped 1.54 mL, uncapped 0.67 mL, p = 0.0156), and the airway volume range (capped 6.44 mL, uncapped 2.93 mL, p = 0.0156) were significantly larger in the capped tubes. The coefficient of variance (capped 0.37, uncapped 0.26, p = 0.0469) and the normalized range (capped 1.52, uncapped 1.09, p = 0.0313) were significantly larger in the capped tubes. Conclusion There is a statistically significant change in airway dynamics in children with tracheotomy tubes when breathing via the airway as compared to breathing via the tracheotomy tube.

Venous pressure monitoring does not accurately predict access failure in children

Abstract. Access failure is a significant cause of morbidity and mortality in hemodialysis patients. Routine monitoring of arteriovenous (AV) fistulas and grafts could increase access longevity. Dynamic venous pressure monitoring is a surveillance test advocated to detect early signs of vascular thrombosis. Venous pressure measurements obtained, per DOQI recommendations, in children undergoing hemodialysis with an AV fistula or graft were reviewed. Baseline venous pressures were established by calculating the mean of venous pressures obtained without an antecedent thrombosis. A paired t-test was performed comparing mean baseline pressure measurements with pressures immediately preceding each thrombosis episode. Since some patients had multiple thrombosis episodes, the assumption of independence was not met. A second paired t-test was performed comparing mean baseline pressures with the mean pressure measurement per individual, obtained immediately preceding a thrombosis episode; 335 venous pressures were collected in ten pediatric patients. Eighteen thromboses occurred in five patients, in whom a total of 241 venous pressures were measured. Venous pressures did not correlate with thrombotic events (P=0.4284). Specific thrombotic events for each patient were correlated with mean patient-specific venous pressures and showed no correlation (P=0.3229). Dynamic venous pressure monitoring is not an adequate predictor of access thrombosis in pediatric patients.