ABSTRACT
SETTING AND PARTICIPANTS: 66 first-year internal medicine residents at Northwestern Memorial Hospital were randomized to Group A or Group B. Curriculum participation was mandatory for all first-year internal medicine residents, but participants were given the option to exclude their answers from the study. DESCRIPTION: Prior to 2020, there was no formal radiographic curriculum for internal medicine residents at our institution. Additionally, the COVID pandemic necessitated a paradigm shift in medical education from in-person teaching to remote learning. Accordingly, we created a novel virtual learning curriculum to teach common CXR findings to first-year residents. Objectives of the curriculum include 1) providing first-year residents with a systematic approach to reading and interpreting CXRs, and 2) prompting pattern recognition via proper identification of common CXR findings. We created a randomized cohort study with cross-over design to evaluate the efficacy of our curriculum. First-year internal medicine residents at McGaw Medical Center of Northwestern University were randomized into two groups (Group A/B). In phase I, only Group A was administered the 11-week curriculum. Learners received 2-4 weekly CXRs focusing on a modified ”ABCDE” approach. Each weekly lesson was designed to be completed in 15 mins via smartphone or laptop. Multiple choice standardized assessments were administered before (Pre-Test) and after (Post- Test #1) administration. In phase II, Group B, but not Group A, was given the curriculum;both groups then completed Post-Test #2. This phase assessed curriculum efficacy (Group B) and learning retention (Group A). EVALUATION: Independent and paired-sample T tests were used to compare scores between and within groups. Group A scored higher on Post-Test #1 following curriculum administration, compared to on the Pre-Test (pre: 44 ± 15%;post: 59 ± 17 %;p= 0.005). Group B scored similarly on the Pre-test and Post-Test #1 (pre: 50 ± 14%;post: 44 ± 17%;p= 0.25), but higher on Post-Test #2, following their curriculum administration (60 ± 17%) than on Post-Test #1 (p= 0.04). There was no statistically significant difference in Post-Test #2 scores between Groups A and B (55 ± 17% and 60 ± 17%, respectively). In Group A, self-assessed confidence with CXR reading was higher at the time of Post- Test #1 than Pre-Test (72 ± 13%;55 ± 12%;p= <0.01). DISCUSSION / REFLECTION / LESSONS LEARNED: This study suggests that our novel remote learning curriculum is a practical, effective adjunct to standard residency education for reading CXRs. Notably, residents who received the curriculum demonstrated higher scores and had improved confidence with reading CXRs. Study limitations include small sample size and participant attrition. Future studies include applying our remote learning framework to other imaging studies.
ABSTRACT
Rationale: The frequency of superinfection with Aspergillus among patients with severe COVID-19-related respiratory failure is unknown. The association of bronchoalveolar lavage (BAL) fluid markers, such as galactomannan, with the presence of Aspergillus infections is also unclear. Methods: Our cohort included all patients with COVID-19 admitted to the Northwestern Memorial Hospital ICU from March to November 2020 who were intubated and underwent a BAL;patients were identified as having COVID-19 by the presence of a flag in the electronic medical record. We evaluated BAL fluid data for fungal organisms and markers. Data analysis was performed in Excel and Prism, with non-parametric values compared by two-tailed Mann Whitney tests. Results: We identified 274 patients meeting study entry criteria. The median (interquartile range) age was 62 years (47-69), and 93 (34%) were female. 714 BALs were performed on these patients, and 457 galactomannan tests were sent from these BALs. Only six (2.2%) patients grew Aspergillus on BAL fluid culture;one patient never grew Aspergillus but given his significantly elevated BAL galactomannan of 7.81, was empirically treated. The median (IQR) optical density index of galactomannan for the BAL samples that grew Aspergillus was 6.69 (3.37-8.87) compared with 0.08 (0.06-0.13) for those that did not, p<0.001. Using our lab's reported galactomannan index cutoff of >0.5 as positive, with patients who grew Aspergillus on culture as gold standard, the test sensitivity was 91.7% and specificity was 94.6%;using the ATS recommended cutoff of >1.0 as positive, the sensitivity was 75.0% and specificity was 97.5%. The median (IQR) age of patients who grew Aspergillus was 71.5 years (64-80.5) (p=0.02 compared with the overall cohort). Only two were immunocompromised (diffuse large B-cell lymphoma and another was status post liver transplant). Of note, two patients who grew Aspergillus had visible mold or plaque seen during bronchoscopy;three had cavitary findings on imaging;three died. Conclusions: Aspergillus superinfection is uncommon among ventilated patients with COVID-19 but can occur in patients without typical immunocompromising risk factors. BAL fluid galactomannan levels are lower in patients without Aspergillus, and had reasonable specificity especially if cutoff of >1.0 was used. Future analysis should focus on additional factors that differentiate between these two groups.
ABSTRACT
Rationale: Coronavirus disease 2019 (COVID-19) can cause severe respiratory failure that worsens despite maximal medical management. When to initiate extracorporeal membrane oxygenation (ECMO) and how to manage these patients on ECMO is not clear. Here, we present our experience with venovenous ECMO to support patients with COVID-19 and compare it to historic patients supported with VV-ECMO for other causes of respiratory failure. Methods: Patients admitted to our tertiary academic medical center in 2019 and 2020 who received VV ECMO support were included in this retrospective chart review. We examined patients with and without COVID-19 infection. We placed COVID-19 patients on ECMO who failed supportive care with mechanical ventilation using a high PEEP low tidal volume strategy, prone positioning, and neuromuscular blockade. Data analysis were done in Excel and Prism. Non-parametric data were compared with unpaired, two-tailed Mann-Whitney tests. Results: ECMO was provided to 26 COVID-19 patients and 38 patients without COVID-19. Median (interquartile range) age of COVID-19 patients was 49.5 (40.5-56.25), compared with non-COVID-19 patients: 53.5 (30.5-60.25), p=0.28. COVID-19 patients had a significantly higher BMI: 32 (30.1-35.9) vs. 26.4 (23.6-29.4), p<0.001. There were 27% female COVID-19 patients compared with 37% female non-COVID patients (p=0.43). COVID-19 patients had similar PaO2:FiO2 ratios as non-COVID patients on day of cannulation: 74 (69-112) vs 78 (60-205), p=0.65. COVID-19 patients had longer ventilator duration pre-cannulation (not including time spent intubated at outside hospitals prior to transfer to our center)-1.9 (1.4-7.0) days vs 0.7 (-.2-1.0) days, p<0.001. COVID patients spent more days on ECMO compared with non-COVID patients: 20.7 (7.3-36.5) vs. 11.5 (3.8-26.8), p=0.14. Twelve (46%) of the COVID-19 ECMO patients died, compared with 9 (25%) of the non-COVID ECMO patients, p=0.10. Conclusions: In patients with severe SARS-CoV-2 pneumonia induced ARDS who fail maximal supportive therapy with mechanical ventilation, outcomes are similar or worse than patients historically receiving VV ECMO support for respiratory failure. These findings highlight the need to determine the optimal timing of ECMO initiation and management in patients with severe SARS-CoV-2 pneumonia.