By Richard R. Watkins, MD, MS, FACP, FIDSA, FISAC
Professor of Medicine, Northeast Ohio Medical University, Rootstown, OH
A cohort study that included 48 hospitals in Michigan found that 12% of patients treated for community-acquired pneumonia were diagnosed inappropriately. Older age, dementia, and presenting with acute change in mental status increased the risk for misdiagnosis.
Gupta AB, et al. Inappropriate diagnosis of pneumonia among hospitalized adults. JAMA Intern Med 2024;Mar 25:e240077. [Online ahead of print].
Community-acquired pneumonia (CAP) often is challenging to diagnose. One of the main reasons is that, despite a multitude of guidelines and clinical evidence, no universally recognized definition of CAP exists. Several criteria used in the diagnosis, such as physical exam findings and an infiltrate(s) on chest imaging, can be subjective and are prone to observer bias. Inappropriate diagnosis and treatment of CAP can have detrimental effects, including antibiotic toxicities, spreading antibiotic resistance, and delaying appropriate management (e.g., treating congestive heart failure). Gupta and colleagues sought to understand the incidence and consequences of inappropriately diagnosing CAP in hospitalized patients, as well as the factors that lead to misdiagnosis.
This cohort study included patients admitted for CAP between July 2017 and March 2020 at 48 Michigan hospitals. The facilities were part of a quality initiative designed to improve care in hospitalized patients. Those eligible for inclusion were adults admitted to general medical floors with a discharge diagnostic code of pneumonia who received antibiotics on day 1 or 2 of hospitalization. Patients were excluded who were severely immunocompromised, pregnant, admitted for comfort measures, treated for an additional infection unrelated to CAP, left against medical advice, received care in an intensive care unit, were placed on mechanical ventilation, received antibiotics longer than 14 days, or were treated for a chronic obstructive pulmonary disease (COPD) exacerbation with doxycycline or azithromycin.
The primary outcome of the study was an inappropriate diagnosis of CAP. This was defined as any antibiotic treatment in a patient with fewer than two signs and/or symptoms of pneumonia or who lacked radiographic features of CAP. The secondary outcomes were 30-day all-cause post-discharge mortality, hospital readmission, post-discharge emergency department visit, Clostridioides difficile infection, or an antibiotic-associated adverse event. The investigators also compared outcomes based on receipt of short-course (three days or less) vs. long-course (more than three days) antibiotic therapy.
There were 17,290 patients treated for CAP during the study time frame, of whom 2,079 (12%) were misdiagnosed based on study criteria. The rate of misdiagnosis varied by hospital and ranged from 29% to 5%. Of the misdiagnosed patients, 1,531 (73.6%) lacked appropriate radiographic criteria, 507 (24.4%) had fewer than two signs and/or symptoms of CAP, and 41 (2.0%) lacked both. Like patients with CAP, those misdiagnosed often presented with new or increased cough (65.8%) and new or increased dyspnea (63.9%). Tachypnea and a leukocyte count > 10,000/µL or < 4,000/µL also were common in misdiagnosed patients (55.7% and 54.4%, respectively).
Compared to patients with CAP, the misdiagnosed were older (age 75 years or older), had dementia, were more likely to have public insurance, presented with altered mental status, had decreased mobility on admission, or had an inpatient hospitalization in the preceding 90 days. At discharge, the misdiagnosed were more likely to go to a skilled nursing facility. Furthermore, they received a median of seven days (range, five to nine days) of antibiotics. Most of the misdiagnosed (87.6%) received a full course of antibiotics. Being white, having a history of COPD, and presenting with an exacerbation of COPD increased the risk for receiving a full course of inappropriate antibiotics. Finally, receiving a full course of antibiotics was associated with more antibiotic-associated events (adjusted odds ratio [AOR], 7.23; 95% confidence interval [CI], 1.18-44.35; P = 0.03) compared to a short course.
COMMENTARY
CAP is a common illness that causes significant morbidity and mortality despite modern diagnostic testing and effective antibiotics. There are several reasons physicians continue to diagnose CAP inappropriately. One is the availability bias, which is the tendency to make decisions based on information that most readily comes to mind. Because they encounter CAP so frequently, physicians may be inclined to not fully employ their critical thinking skills. This is understandable given the hectic hospital environment with its myriad competing imperatives. Another is the nonspecific nature of CAP signs and symptoms, which often overlap with other acute cardiopulmonary conditions (e.g., COPD and congestive heart failure). Because of this ambiguity and the poor outcomes associated with CAP, physicians may feel the urge to overtreat rather than risk missing a CAP diagnosis. Finally, physicians may feel pressure to meet quality metrics surrounding antibiotic administration.
The study had some limitations worth mentioning. First, omissions of signs and symptoms of CAP in the physician documentation may have resulted in misclassification of patients with an appropriate diagnosis of CAP as inappropriate. Second, relying on physician documentation of antibiotic-associated adverse events may have led to underreporting since the causality may not have been appreciated. Third, there may have been unrecognized confounding variables that affected the study as a result of the retrospective design. Finally, patients with altered mental status may not have been able to accurately verbalize signs of CAP. The study by Gupta and colleagues is an important addition to the evidence on managing CAP. By quantifying the rate of misdiagnosis, this should serve as a reminder to clinicians about the importance of an accurate diagnosis. It also underscores the critical need for better diagnostic methodologies.
