Survey of observer variation in transesophageal echocardiography: comparison of anesthesiology and cardiology literature.
OBJECTIVE: Transesophageal echocardiographic examination tends to be somewhat observer and experience dependent, and observer bias can arise easily when data are calculated and interpreted by unskilled, nonblinded, or single observers. The study plan was to see whether authors have adequately described how observer bias is minimized in their studies. Thus, a study was conducted systematically reviewing methods reported in transesophageal echocardio graphy articles in peer-reviewed anesthesiology journals versus those reported in peer-reviewed cardiology journals. INTERVENTIONS: After MEDLINE searches of the literature published from 1997 through 1999, the authors investigated 56 anesthesiology reports and 56 randomly selected, year-matched cardiology reports. An 8-item questionnaire was developed that examined several factors: the number of observers and their experience levels, whether observers were blind to clinical data, whether low-quality images were excluded, the use of on-line or off-line analysis, and observer variability. MAIN RESULTS: The analysis revealed inadequacies in reporting of important information that relates to bias and quality in 91.1% of anesthesiology and 98.2% of cardiology articles. Observer variability was not reported in 50.0% of the anesthesiology reports and 67.9% of the cardiology reports; however, difference between the 2 bodies of literature was not significant. The journal impact factor was significantly higher for the cardiology literature than for the anesthesiology literature (2.42 [0.386-10.893] v 1.07 [0.664-3.439]; median [range], p < 0.001). CONCLUSION: Articles reviewed had at least some inadequacies in reporting the methods to minimize observer bias in both the anesthesiology and cardiology literature. Reporting methodology standards in TEE examinations remain to be established.