Computerized electrocardiogram (ECG) interpretation has been developed and used in an attempt to reduce the human readers' time and work demands, reduce and standardize the coding used by different ECG classification systems and minimize the inter- and intra-observer variation in human interpretation. The past decade has witnessed a rapid increase in the use of computer programs to interpret all types of ECG tests recorded routinely. However, the use of computerized ECG interpretation has not resulted in major improvement in diagnostic accuracy of human interpretation. The methodological quality of the available published evidence is limited in several aspects and conclusive evidence could not be found: - on the diagnostic accuracy and reliability of using computerized interpretation of resting ECG as an automated laboratory test for screening normal ECGs in asymptomatic adults; and - on whether computerized interpretaion of resting ECG can replace the interpretation by a skilled professional in an ambulatory clinical environment. The available published evidence suggests that the computer programs with the best performance may be as accurate as the human interpretation in diagnosing normal ECGs. However, computerized interpretation of ECGs should be used with the awareness of the risk of false positive and false negative findings.