I lay claim to being TTE’s lowest common denominator - a history major with zero science background. Carl even against the formidable mass of my ignorance you’ve managed to kindle a dim light of understanding.
When I worked in mental health services we were plagued by managers and lawyers who
insisted we used screening tests for suicide and homicide (a risk assessment).
They failed to recognise that the low prevalence of these behaviours nullified the predictive value of the tests. About 0.01% die by suicide each year (11 deaths per 100,000 population per year).
So even if the sensitivity and specificity of the test were as high as 99% (which of course it wasn't) the test would identify 10 out of 1000 people as a suicide risk but probably none of these 1000 people would kill themselves that year.
I lay claim to being TTE’s lowest common denominator - a history major with zero science background. Carl even against the formidable mass of my ignorance you’ve managed to kindle a dim light of understanding.
Thanks this is clear and helpful.
When I worked in mental health services we were plagued by managers and lawyers who
insisted we used screening tests for suicide and homicide (a risk assessment).
They failed to recognise that the low prevalence of these behaviours nullified the predictive value of the tests. About 0.01% die by suicide each year (11 deaths per 100,000 population per year).
So even if the sensitivity and specificity of the test were as high as 99% (which of course it wasn't) the test would identify 10 out of 1000 people as a suicide risk but probably none of these 1000 people would kill themselves that year.
Thank you Carl….got it after 2 views….11 minutes = greater understanding of sensitivity and specificity. Every day’s a school day here in Hobbiton.