The effect of erroneous computer interpretation of ECGs on resident decision making.
Med Decis Making. 2009 May-Jun;29(3):372-6
Authors: Southern WN, Arnsten JH
BACKGROUND: The use of computer interpretations of electrocardiograms(ECGs) as an aid to physician interpretations is widespread. Computer misinterpretations are common and negatively affect physician interpretations. OBJECTIVE: To measure the effect of computer ECG misinterpretations on clinical decision making. DESIGN: Quasi-randomized trial. SETTING: Resident teaching conferences. PARTICIPANTS: Included 105 internal and emergency medicine residents. INTERVENTION: After a brief case presentation, residents were asked to interpret an ECG and choose appropriate management. Residents chose from a concealed stack of handouts; some contained an erroneous computer interpretation of the ECG (citing acute ischemia), and some contained no computer interpretation. MEASUREMENTS: ECG interpretations and management decisions by residents whose ECG did or did not include an erroneous computer interpretation were compared using chi-square tests. RESULTS: The presence or absence of erroneous computer interpretations of ischemia did not significantly affect residents' ECG interpretations (P = 0.62). However, the residents whose ECGs included erroneous computer interpretations were more likely to recommend revascularization than the residents without (30% v. 10%, P = 0.01). Of those residents who read the ECG as diagnostic of ischemia, those with the erroneous computer interpretation were more likely to recommend revascularization than those without (54% v. 25%, P = 0.048). LIMITATIONS: A single ECG was used. CONCLUSIONS: Erroneous computer interpretations of ECGs affected residents' clinical decision making in the absence of an effect on the actual interpretation of the ECG. Measuring the impact of computer misinterpretations by examining only physician interpretations will underestimate the effect of computer misinterpretations on clinical decision making.
PMID: 19470722 [PubMed - indexed for MEDLINE]