In our field this refers to scores that are deemed to separate “help seeking” and “non-help-seeking” scores on some measure. The traditional terms are about separating “clinical” and “non-clinical” but I have switched from that to this “H-S” versus “N-H-S” to avoid the rather disease model implications of “clinical” versus “non-clinical” but the reality is that there are no perfect and short terms as probably what we are looking for is a cutting point to use in screening with the measure that would separate “those currently not receiving input but likely to benefit from input (cost-efficiently)” from those who don’t fall into that group. The brutal reality for the world of psychosocial interventions is that there are no issues, no measures and no interventions that really have a strong basis on which to make this distinction. However, everyone, politicians, managers and therapists alike seem to want this so we have … cutting points!
The term can be used for cutting points dividing a score range into more than two ranges but it’s almost always used for dichotomising as above.
Details #
There are two main methods to find cutting points:
- The “Clinically Significant Change” part of the RCSC (Reliable and Clinically Significant Change) paradigm
- Using the Receiver Operating Characteristic/Curve (ROC) method, with or without weighting the costs/benefits of false negatives and false positives.
The CSC method is simpler than the ROC method so it’s more widely used in our field and we have always used it when giving cutting points for scores on CORE measures for different countries/languages of the measures. The ROC method, in principle, is more adaptable when costs are considered.
If you wish to understand these issues do look at the existing glossary entries below (existing ones have active links) and then at the screening model shiny app in the online resources. When I have time I will create entries here for the CSC method and the ROC and probably some Rblog and other shiny apps as this is a pretty important area. (And I often wish we simply dumped cutting points, see the issues in the dichotomisation glossary entry below!)
Try also #
- Clinically Significant Change (CSC)
- Dichotomisation
- False negative (in screening)
- False positive (in screening)
- Receiver Operating Characteristic/Curve (ROC)
- Screening
- Sensitivity (in screening)
- Specificity
Chapters #
The issues are touched upon throughout the OMbook but not particularly expanded anywhere.
Online resources #
Dates #
First created 19.iii.25.
