Disagreement between emergency physician and inpatient physician diagnosis of infection in older adults admitted from the emergency department.

OBJECTIVES: Older adults with infection are at increased risk of misdiagnosis while they are patients in the emergency department (ED) due to the common presence of nonspecific signs and symptoms. The primary objective was to determine the proportion of admitted older adult patients thought by the emergency physician (EP) to be infected, as compared with the diagnostic impression of inpatient physicians. The secondary objective was to determine the agreement between EP and inpatient physician diagnosis of specific infection types.
METHODS: The authors conducted a prospective, observational, convenience sampling of a cohort of ED patients ≥65 years old admitted to the hospital with diagnoses of acute infection. EPs noted at least one suspected source of infection. Inpatient diagnosis of infection was determined by chart review of the inpatient chart. Outcomes included the presence of any infection and of specific infectious sources diagnosed within 48 hours of admission. EP and inpatient diagnoses were compared using proportions, positive and negative likelihood ratios (LR+ and LR-), and the phi coefficient.
RESULTS: The study included 103 patients diagnosed with a suspected infection by the EP. Nineteen patients (18.4%, 95% confidence interval [CI] = 11.5% to 27.3%) were not diagnosed with any infection by the inpatient physician. For specific infection sources, ED diagnosis of bloodstream infection often did not agree with the inpatient diagnosis. Sensitivity was 40.0% and specificity 78.4% with an LR+ of 1.85 and LR- of 0.76. The phi coefficient was 0.15. EPs overdiagnosed pulmonary infection, with 72.1% specificity and an LR+ of 3.24. EP diagnosis had good accuracy for skin and soft tissue infection (sensitivity = 78.6% and specificity = 96.6%), with adequate LRs (LR+ of 23.3 and LR- of 0.22). Urinary tract infection (UTI) was underdiagnosed in the ED (sensitivity = 58.3%), but it is unclear if this is due to true ED underdiagnosis or due to overdiagnosis of UTI in the inpatient setting.
CONCLUSIONS: In older patients admitted from the ED, the provisional ED diagnosis and the inpatient diagnosis of an acute infection often disagree. In this sample, 18% of older ED patients diagnosed with infection during an ED stay were not diagnosed as infected by the inpatient physician. Regarding infection types, EPs were poor at diagnosing bacteremia and overdiagnosed pulmonary infections. EP diagnosis of skin and soft tissue infection generally agreed with the inpatient physician. There was also disagreement regarding presence of UTI, but the true nature of this difference is unclear from the data obtained in this study.