Search PubMedSearch

Biomedical subjects

I Nordgaard-Andersen

Publications and source records attributed to I Nordgaard-Andersen.

2 recordsLinked to original sources

[Automatic computerized ECG interpretation].

The introduction of computerised interpretation of an electrocardiogram depends on the reliability of the interpretation. Previous investigations of computerised ECG diagnosis have been concerned with the diagnostic significance of selected QRST morphology changes and/or arrhythmia. In the present investigation, a full interpretation of the recorded values has been performed in terms of diagnosis and conclusions for each ECG. The investigation included 1,000 ECG's from 500 in-patients, at a Central Hospital. The ECG's were recorded on Mingograf 740 from Siemens-Elema, AB. All data from each computerised ECG were compared with the results of an ordinary visual analysis. For the rhythm diagnosis, mistakes were found in 11% of the ECG's, and in 2-3% of the mistakes the wrong diagnosis was considered of clinical significance. For the diagnosis of the QRST morphology, mistakes were found in 29% of the electrocardiograms with considered clinical significance in 20% of the ECG's. Agreement between the diagnosis "normal ECG" in the computer-based ECG interpretation and the visual analysed ECG with the diagnosis "normal ECG", was found in 99.4%. In the diagnosis of cardiac rhythm, agreement was found in 89% of the ECG's, whereas agreement concerning the QRST morphology was 71%. The differences were mainly because of too narrow boundaries in the diagnosis of normal ECG's in the investigated computerprogram. It is concluded, that the value of automatically computerised interpretation of an electrocardiogram depends to a large extent on the criteria encoded in the computerprogram employed.

Adolescent

[Repeated ECG recording, benefits].

The investigation included 500 inpatients, in a Central Hospital. For each of the patients, two ECG registrations were made each time an electrocardiogram was requested and without any modifications between each registration. All electrocardiograms were examined for variations in cardiacrhythm and QRST morphology. In 19.8% of the patients, this double registration provided some new information. The cardiacrhythm was changed in 16.4%, and this was of clinical significance in 15.2% of these. The figure corresponded to 1.6% of all the ECG's. The QRST morphology had altered between the registrations in 5.4%. Among the alterations, 46.8% were of clinical significance, corresponding to the presence of clinically important information in 3% of all the patients investigated, if an ECG registration was repeated within one minute. In one out of every fifth ECG registration it would be possible to obtain some new information, if the registration was repeated within one minute.

Adolescent