Case Presentation: 71-year-old female with history of DM2, HTN, breast carcinoma in situ status post mastectomy, B-cell lymphoma (R-CHOP in 2016 & RT, currently in remission), autoimmune hepatitis related cirrhosis on Azathioprine presented to the hospital with a 4-day history of progressively worsening productive cough, mild hemoptysis, fever and shortness of breath. In the Emergency Department, she was hypoxic requiring 4 lpm oxygen but otherwise stable hemodynamically. CT lungs showed mostly perihilar, but also peripheral infiltrates without overt consolidation. She was hospitalized and started on antibiotics for community acquired pneumonia, but antibiotic coverage was expanded to cefepime and vancomycin on day 2 of hospitalization due to progression of respiratory distress. She was intubated and transferred to the ICU on day 3. She underwent bronchoscopy the same day. Mucopurulent secretions were noted, and BAL fluid was sent for further analysis. BAL or induced sputum for Pneumocystis Jirovecii staining was not sent. Legionella Ag was negative and LDH was noted to be markedly elevated at 3000 U/L. On hospital day 4, Fungitell test was strongly positive at 1700 pg/ml and Voriconazole was started. ANCA panel was negative. On hospital day 7, BAL studies which were sent on day 3 turned positive for PJ PCR, following which voriconazole was stopped and Bactrim and steroids were started. Unfortunately, she developed a high-grade drug fever due to Bactrim and then was switched to Primaquine and Clindamycin. Hospital course was further complicated by hemolytic anemia, severe thrombocytopenia, and upper GI bleed. Her respiratory failure progressed, and her family transitioned her to comfort measures.
Discussion: A practicing hospitalist often encounters pneumonia in immunocompromised patients, and this case highlights the inherent challenges in treating such patients. Performing a traditional pneumocystis stain (Gomori Methenamine Silver Stain) (turnaround time of 1 day) instead of waiting for results of BAL PCR for PJ (usual turnaround time varies between 1-5 days) for 4 days would have prevented the diagnostic delay. Delay in diagnosis is considered as “diagnostic error” by the National Academies. We applied the “DEER Taxonomy tool” to analyze potential reasons for diagnostic errors and found suboptimal weighing of history and laboratory data, delay in ordering the needed tests, delay in ordering a consult, delay in considering the diagnosis as some of the contributory factors. Premature closure and anchoring bias probably contributed to this diagnostic error as well. The vignette highlights the need for debiasing strategies such as slowing down thinking to transition to ‘type 2 thinking’, metacognition and use of checklists.
Conclusions: As hospitalists taking care of medically complex patients, being aware of cognitive biases and having a high index of suspicion for Pneumocystis pneumonia especially when one encounters the constellation of classic perihilar infiltrates on chest imaging and elevated LDH may help timely treatment of critically ill patients. Depending on the practice setting, realizing that the traditional staining of BAL or induced sputum for pneumocystis may offer quicker and useful results, despite lower sensitivity, when compared to the gold standard PCR test could potentially avert a diagnostic error.