Three direct consequences of changing a diagnostic threshold:
- Change in size/number of identified population
- Change is average severity of identified population
- Change in homogeneity (similarity) of identified population
If we make the diagnosis more inclusive/less restrictive/lower threshold/cast net wider:
- A larger number of children will be classified as having this condition (identified population increases)
- The average severity of the identified population decreases (we have included milder cases who were not included before)
- The homogeneity of the identified population decreases (they are more different from each other than they were before)
If we make the diagnosis more conservative/set higher threshold for identification:
- A smaller number of children will be identified as having this condition (identified population decreases)
- The average severity of identified cases increases (we have thrown out the milder cases, so the mean severity shifts downwards)
- The homogeneity of the identified population increases (we have excluded mild outliers who typically show more variability in presentation)
Notice also: the children have not changed–only our identification of some of the children as having a particular problem/disorder/characteristic.
Many other changes will spin off of these fundamental effects: we may have more “awareness” of a problem among our children, funding patterns for the exceptionality may change, resources may shift. I view all of the secondary effects as being driven by the more fundamental effects on perceived size of the group and perceived severity of the problem; the increased homogeneity helps establish a picture in the mind of the professional or general population.