Automated Cardiac Dysrhythmia Analysis

Lewis J. Thomas, Kenneth W. Clark, Charles N. Mead, G. charles Oliver, Kenneth L. Ripley, Bruce F. Spenner, Bruce F. Spenner

Research output: Contribution to journalArticlepeer-review

39 Scopus citations


Automated analysis of abnormal cardiac rhythms (dysrhythmias) is well established for use in real-time ECG monitoring in hospital intensive-care units and for high-speed processing of long-term ambulatory ECG recordings; yet considerable difficulties persist. The many facets of ECG signal acquisition have not been sufficiently well standardized, with a result that definitive signal characterization continues to be troublesome. Analysis algorithms rely heavily on time-domain feature extraction or correlation techniques, although incursions have been made into other domains. Progress continues to be made in improving analysis accuracy, but no algorithm is without its weaknesses. Most system implementations employ some degree of human interaction to compensate for analysis deficiencies. Performance evaluation of implemented systems requires extensive effort, and results to date are clouded by a lack of evaluation standards and the absence of a widely accepted evaluation database. The American Heart Association is sponsoring work which promises to put future evaluations on firmer ground. Research continues to address all of these issues because of a strong belief in the clinical utility of automated dysrhythmia analysis. The rationale for that belief is clearer for the analysis of long-term ECG recordings than it is for in-hospital monitoring, but results are available to show that patients are treated more vigorously if such monitoring is employed, and newer therapeutic approaches have increased the importance of reliably detecting rare but significant events.

Original languageEnglish
Pages (from-to)1322-1337
Number of pages16
JournalProceedings of the IEEE
Issue number9
StatePublished - Sep 1979


Dive into the research topics of 'Automated Cardiac Dysrhythmia Analysis'. Together they form a unique fingerprint.

Cite this