Best Practices - Concurrent Mental Health and Substance Use Disorders

Appendix E

Terms for Substance Abuse and Mental Health Screening

1. The sensitivity of a screening tool refers to its ability to detect the condition or health concern that it is intended to detect.³⁷⁷ A highly sensitive tool for mental health or substance abuse will, therefore, not miss too many people who have either a mental health or substance use disorder as determined by a subsequent diagnostic assessment.
2. The specificity of a screening tool refers to its ability to avoid saying that someone may have the problem or health concern when in fact they do not.³⁷⁷ A highly specific tool for mental health or substance abuse will, therefore, not identify too many people as possibly having either a mental health or substance use disorder only to have this ruled out by subsequent diagnostic assessment.
3. While the goal is usually to develop screening tools that are high in both sensitivity and specificity, one may lean toward high sensitivity and use later assessment to rule out the false positives. If, however, subsequent assessment of many false positives is seen as inappropriate (e.g., too expensive) one would lean toward higher specificity. Normally, one tries to strike a balance.
4. Sensitivity and specificity, however, tell only part of the story about the operating characteristics of a screening tool. It is also important to determine the predictive value of a screening tool.³⁷⁷ Positive predictive value refers to the probability of being positive on the criterion measure, if one is identified as at risk with the screening tool; negative predictive value refers to the probability of being negative on the criterion measure if one is identified as not at risk with the screening tool. This means that a tool which has a certain level of sensitivity and specificity will actually perform better or worse when used in settings that differ substantially in the prevalence of the condition or health concern being detected. Thus, a substance abuse screening tool can have a higher predictive value when used, for example, in a setting providing services to the homeless with a very high percentage of people with substance use disorders, compared to an early intervention program for schizophrenia where the percentage is not as high. Thus, a screening tool that works for one setting, or sub-population, may not be the best choice for another.
5. There is an important distinction between population screening, where everyone in the service would be given the screening questions or procedures, and strategic screening, whereby the questions or procedures are applied only for certain sub-groups already thought to be at higher risk. This is getting at the issue of predictive value mentioned above - the payoff will be higher when you use the screening tool with those subgroups with the highest prevalence of the problem. The choice between population or strategic screening is often a matter of the available resources, and the consequences of missing people who really should have been screened positive for further diagnostic assessment. The consequences of missing a substance abuse disorder can be very substantial for some mental health disorders such as schizophrenia or bipolar disorders.