[Investigation methods in clinical cardiology. IV. Clinical measurements in cardiology: validity and errors of measurements].
Measurements represent an essential part of clinical activity. Very often, however, relevant disagreement in clinical measurements becomes apparent. The sources of this variability are the subjects (patients) that are measured, the measurement instrument itself, and the observer. The assessment of the quality of measurement usually relies on the evaluation of its reproducibility and its validity. The reproducibility is basically measured as the inter-observer concordance, the intra-observer concordance, and the test-retest concordance. The specific parameter used to its quantification (intra-class correlation coefficient, kappa index, graphic methods, etc.) depend on the kind of variable to be measured. The validity of the measurement is the degree to which the measurement is really measuring what we think it should. If an acceptable standard is available, then so called criterion validity is usually assessed. Otherwise the validity should be assessed by other ways that use subjective criteria (content validity and face validity) or empirical criteria (construct validity).