Erpelt W
Minerva Med. 1976 Oct;67(48):3089-94.
Multistage screening in Hesse (W. Germany) and the visual processing of about 12,000 ECG's in accordance with the USPHS (Riedel programme) are reported. A good fit was found with data obtained from a similar study in another part of the country. Comparison is then made with the visual and computerised processing of about 6,000 ECG's in accordance with the same code, using four classes: "normal", "conditionally within the limits of normal", "suspect" and "definitely pathological". The encouraging result on comparison between "normal" and "pathological" in the 6.3% consisting of false negatives (only 2.1% definitely pathological) suggests that "normal" and "conditionally normal" ratings by the computer can be accepted, whereas considerable possibilities for cardiological diagnosis would remain in the case of pathological traces, which would always be accepted, where as considerable possibilities recognised standards set down by the USPHS code are considered reliable. They can be used under hospital or screening conditions. On the other hand, the "scale of values" programme offered as a model requires further checking. Notes on future development are also offered.