AN EXERCISE in CLINICAL REASONING with COGNITIVE AUTOPSY Ankita Tandon, DO; Francisco Alvarado, MD (Presenters); Kellee L Oller, MD, FACP
Total Page：16
File Type：pdf, Size：1020Kb
SHARE NOVEMBER 2019 V42, NO.11 MORNING REPORT A BEWILDERING BACTEREMIA: AN EXERCISE IN CLINICAL REASONING WITH COGNITIVE AUTOPSY Ankita Tandon, DO; Francisco Alvarado, MD (presenters); Kellee L Oller, MD, FACP (*Discussant in italics) Dr. Tandon ([email protected]) is a PGY-2 Internal Medicine resident and Dr. Alvarado ([email protected]) is a PGY-3 Internal Medicine resident both at the University of South Florida. Dr. Oller ([email protected]; @kelleeoller) is an academic hospitalist for the University of South Florida Morsani College of Medicine, practicing at Tampa General Hospital, and program director for the Internal Medicine Residency Program. 60-year-old Hispanic man with a history of insu- was suspected, no specific etiology was identified. He was lin-dependent diabetes mellitus presented for six transitioned to oral ciprofloxacin and discharged home to A days of abdominal pain, nausea, and vomiting. complete a two-week course of antibiotics. Broadly, his symptoms can be categorized into two With unexpected shifts in a patients’ hospital course initial diagnostic categories: gastrointestinal etiologies or it is important to step back and evaluate if the newly a gastrointestinal manifestation of systemic disease. The identified symptoms were present on admission or if they differential diagnosis of gastrointestinal etiologies can were directly related to the hospital stay itself (such as be subcategorized by organ, for example: peptic ulcer a new medication started since admission or a hospital disease, pancreatitis, small bowel obstruction, biliary ob- acquired infection). As the patient was in the hospital for struction, acute hepatitis. Diabetic ketoacidosis (DKA) 24 hours or less, it is likely that this infection was present or adrenal insufficiency are examples of how abdomi- on admission, just not evident. nal pain and vomiting may be symptoms of a systemic Klebsiella bacteremia can result from infections of process. the genitourinary tract (e.g.. pyelonephritis), pulmonary His abdominal pain was diffuse, constant, and exacer- infections, and the gastrointestinal tract (e.g.. cholangi- bated with oral intake. He denied fever, diarrhea, constipa- tis). These more common etiologies have been excluded tion, melena, or hematochezia. He had been unemployed as sources for this patient’s bacteremia. Interestingly, for one year and was unable to afford his medications. He Klebsiella has been an increasingly recognized culprit denied drug use, but reported consumption of 4-6 beers of primary liver abscess syndrome (without evidence of per day. On examination he was tachycardic and mildly associated biliary infection) in the United States. In the hypotensive. On labs, his blood glucose was elevated to community setting, Klebsiella liver abscess syndrome is 396 mg/dl along with an anion gap metabolic acidosis typically seen in patients with poorly controlled diabe- and elevated blood ketones. Hemoglobin A1C was 14%. tes.1 Reassuringly, he underwent CT imaging without His complete blood cell count was remarkable for a mild evidence of biliary dilations or liver abscess, and classic leukocytosis. Liver function tests were normal. A CT scan elevation of alkaline phosphatase or other liver function of abdomen/pelvis showed nonspecific stranding surround- tests are not present. A source is identified in 84% of pa- ing the hepatic flexure of the colon as well as distal gastric tients with community-acquired Klebsiella bacteremia, antrum and duodenum. He was started on intravenous unlike in this case.1 fluids, insulin, and pantoprazole. With these interventions, Two weeks after his initial presentation he returned to he reported improvement in his symptoms and his labora- the emergency room with subjective fevers, nausea, eme- tory derangements normalized the following day. sis, and decreased tolerance of his oral antibiotics. He was On the second hospital day, he developed a fever of afebrile and hypotensive requiring vasopressors. Labs now 102ºF. His symptoms from admission were improving demonstrated white blood cell count of 20.5x10*3 ul, and he denied new symptoms. His procalcitonin was alkaline phosphate of 323 U/L, and mildly elevated trans- elevated at 2.14 ng/ml. Other laboratory data remained aminases. A CT of the abdomen and pelvis was repeated. unchanged, and urinalysis and chest radiograph showed The previously reported antral and duodenal stranding no infection. The following day blood cultures resulted was resolved; however, a 7 x 5 cm hepatic abscess was Klebsiella pneumoniae. While a gastrointestinal source continued on page 2 1 SGIM FORUM NOVEMBER 2019 V42, NO.11 SHARE MORNING REPORT (continued from page 2) identified (Figure A). Broad spec- A B trum antibiotics were initiated. He underwent CT guided drainage of the hepatic abscess. Culture was positive for Klebsiella pneumoniae. The recurrence of this patients’ symptoms prompting readmission with septic shock could represent antibiotic failure in the setting of re- sistance, antibiotic intolerance since ([A] Computed tomography imaging with liver abscess within the right hepatic lobe. [B] discharge due to emesis, or lack of Review of the images from the admission two weeks prior, the abscess was present and source control (such as an abscess). larger. Photos courtesy of the patient.) The repeated CT imaging sheds light on the etiology for his Klebsiella illness script, to recognize and achieve the interventions needed to mitigate bacteremia from his first admission a diagnosis. In contrast, complex and learn from diagnostic errors are and explains his failure to improve diagnostic dilemmas may result in more metacognitive. Examples of with therapy. It would be helpful to uncertainty or a delay in diagnosis or interventions include metacognitive review the CT images from the first treatment. Physicians are encouraged training, use of decision support admission to evaluate if early signs of to retrospectively perform a cognitive tools to broaden differential diag- abscess were present. autopsy of their diagnostic perfor- noses, and checklists that serve as In fact, the images from the first mance following diagnostic inaccu- de-biasing tools. As the table cat- hospitalization did reveal a lesion in racies.2 This works to guide self-feed- egorizes the cognitive breakpoints right hepatic lobe. While the “con- back while avoiding maladaptive impacting the case, the right column clusion” on the radiologist report feelings or guilt which may naturally highlights how a checklist approach only documented the stranding at the occur with diagnostic error but may may have prompted better clinical gastric antrum and duodenum, within impair constructive learning. decision making.4 the full report are details of a 4 x 7 x In the reflection of this case, we Errors in the radiology field are 5 cm lobulated well-circumscribed hy- will simplistically review the faulty often classified into perceptual (an podense lesion felt to represent a com- knowledge, data gathering, and syn- abnormality is present but not seen) plicated or septated cyst (Figure B). thesis that led to the delay in diag- or interpretive (an abnormality is nosis (see the following table). Errors identified but its importance is inter- Discussion in synthesis are much more preva- preted incorrectly).5 Illness frequently follows patterns and lent than errors in data gathering The radiologist reviewed the clinicians can rely on their compiled and knowledge combined.3 Unlike initial CT scan with the indication knowledge of a condition, termed an errors attributed to systems issues, continued on page 3 Utilizing a Checklist Approach to Improve Diagnostic Certainty 4 Errors or Events Leading to Delay or Missed Diagnosis How a Checklist Approach Knowledge Data Synthesis Could Provide Solutions to (Least Prevalent) Gathering (Most Prevalent) Achieve Diagnosis Diagnostic Inertia: On Obtain your own complete Hospital Day 3, the Hospitalists medical history and transitioned with same reperform a focused and working diagnoses purposeful physical exam Short cuts in data gathering: Interpretive Error/Framing Bias: Differentiate hypotheses with reading only the conclusion Radiologist assessment of liver further history, exam, and of the radiology report lesion before the development diagnostic tests of fever and bacteremia Association of poorly Correctness assumption of Context Error: in which Take a “Time-Out” to reflect. controlled diabetes as data previously gathered abdominal pain was attributed What are we missing? What risk factor for Klebsiella to peptic ulcer disease (based else could this be? primary liver abscess on CT stranding) despite Acknowledge uncertainties bacteremia and anticipate follow up and patient education 2 SGIM FORUM NOVEMBER 2019 V42, NO.11 SHARE MORNING REPORT (continued from page 2) of “a patient with abdominal pain Acknowledgements: Healthcare Research and and vomiting.” The way in which a The authors would like to thank Dr. Quality (US); 2005. problem is represented can influence Shanu Gupta for her critical review 3. Graber ML, Franklin N, Gordon the cognitive process, an effect called of the manuscript. R. Diagnostic error in internal framing bias. For instance, the imag- medicine. Arch Intern Med. es may have been interpreted differ- References 2005;165(13):1493-9. ently if the indication for testing was 1. Tsay RW, Siu LK, Fung CP, et 4. Ely JW, Graber ML, Croskerry instead “a patient with fever, leuko- al. Characteristics of bacteremia P. Checklists to reduce di-