Survey-based assessment of emergency room physician practices for patients with suspected infections and sepsis
Background:
Acute infections and sepsis, as leading causes of morbidity and mortality, represent a major burden to healthcare systems. In the United Kingdom and the United States, respectively, an estimated 3.2 and 15 million patients are assessed annually for acute infections and sepsis in accident and emergency departments (EDs).
Current diagnostic procedures for patients with suspected acute infections or sepsis lack sufficient sensitivity and specificity. Little research exists regarding physician utilization and perceived efficacy, utility, and satisfaction with these procedures. This study aimed to identify: 1) prevailing diagnostic practices for suspected acute infections and sepsis in the ED; 2) emergency physician perspectives regarding the value of existing procedures; and, 3) the need for innovative diagnostic tests for acute infections and sepsis.
Methods:
In May 2017, a request to complete an online questionnaire was disseminated to 9,000 US-based ED physicians. 79 responses were evaluated from physicians across 24 states whose geographic distributions largely coincide with population density, infectious disease incidence, and sepsis incidence distributions throughout the US. Physicians were questioned separately about patients with suspected acute infections and patients with suspected sepsis (a more-severe subgroup of the former).
Results:
Most respondents (91%) had 11 or more years’ experience, and most (93%) practiced in metropolitan or suburban settings.
For patients with suspected acute infections, physicians ordered: complete blood count with differential (CBC) (86%); urinalysis (UA) (77%); basic/comprehensive metabolic panel (CMP) (74%); chest x-ray (CXR) (70%); blood cultures (BCX) (59%); urine cultures (UCX) (57%); lactate (54%); procalcitonin (PCT) (4%); and C-reactive protein (CRP) (1%).
For suspected sepsis patients, physicians ordered: CBC (100%); lactate (100%); BCX (100%); UA (99%); CMP (99%); CXR (96%); UCX (95%); PCT (20%); and CRP (20%).
Physicians commonly utilized SIRS (77%) to assess severity in suspected sepsis cases; SOFA (19%), qSOFA (19%), SEP-1 (14%), MEWS (13%), and APACHE (13%) were used more rarely.
Although many (67%) respondents were satisfied with their ability to diagnose sepsis, most (95%) believe current diagnostic procedures require marked improvement.
Respondents also opined on a diagnostic test currently in development, HostDx Sepsis (Inflammatix, Inc). This test informs on the presence, type (bacterial vs. viral), and severity of infection by reading the host immune response (mRNA patterns from whole blood). The test’s algorithm combines the expression levels of 30 genes into clinically actionable scores to predict the likelihood of bacterial infection, viral infection, and 30-day mortality. Upon presentation of published levels of performance, most physicians (81%) perceived the test to be clinically useful and 97% found its performance robust. Most physicians (92%) recommend incorporating HostDx Sepsis into their hospital’s protocol and would order the test (on average) 13 times per week.
Conclusions
Emergency physicians currently rely on multiple imperfect tests to diagnose acute infections and sepsis. Interestingly, CRP and procalcitonin use was limited, and most physicians desire improved diagnostics. HostDx Sepsis was perceived as having favorable performance and high clinical utility. Understanding how physicians utilize and perceive current diagnostic procedures will help advance existing ED assessment methods, thus improving care for patients with acute infections and sepsis.
