Criteria of validity of a method of measurement

Whatever its scope of application, medical or not, a method of measurement must, to achieve a well defined goal, to have the following characteristics:

  1. Sufficient specificity
    By specificity, one understands that the method used explores a well defined parameter and is not disturbed by the possible variation of another ignored parameter.
  2. Sufficient limit of detection
    By limit of detection, one indicates the lower detectable variation of a given parameter. This limit depends much on the importance of the background noise, i.e. of the variability of the signal in absence of the modification. For example, if blood pressure from a subject varies much during repeated measurements, in absence of treatment, it will be more difficult to highlight the effect of an antihypertensive treatment.
  3. Sufficient reproducibility
    Reproducibility indicates the variation of the figures obtained during the measurement of the same parameter during several essays. For example, a sphygmomanometer can go wrong, the conditions under which blood pressure is measured can change, inducing a bad reproducibility and thus errors.
  4. Sufficient accuracy
    The accuracy of a method indicates the difference between the measured value and the actual value. If a sampling contains 1 mg/L of a determined drug, more the value measured by a method is closely related to this number, more the method is right.

A target and the points of impact of shootings illustrates these possibilities:

Not reproducible
Wrong and reproducible
Right and reproducible

These criteria, with the help of a certain number of adaptations, can be applied to clinical events: enhancement of a complaint, advent of an undesirable effect…

Your turn
User session
Bookmark, share this page
Bookmark and Share