Read on Twitter
New paper on skip-out questions in #depression diagnosis.
In many clinical interviews for depression, we query for core symptoms sad mood & anhedonia. If neither are present, we "skip" the other 7 symptoms (eg sleep, appetite) bc 1 of the core symptoms is required for diagnosis🧵
In many clinical interviews for depression, we query for core symptoms sad mood & anhedonia. If neither are present, we "skip" the other 7 symptoms (eg sleep, appetite) bc 1 of the core symptoms is required for diagnosis🧵
This is a practice to save time, but begs the question: are there people with many so-called "secondary" depression symptoms who do not have the core symptoms? This is what our new paper in JPCS led by the wonderful @orla_mcbride is about.
We also look into common data-analytic procedures to deal with skip-out data, such as imputing missing data with 0, meaning that if you don't have sad mood/anhedonia and we don't ask you about insomnia, we decide you don't have insomnia. Yep, bizarre, but it's really common.
Broadly, we find that symptoms *are* endorsed in much higher frequency in the presence of sad mood and/or anhedonia, seemingly justifying DSM's core distinction. There are two "howevers" here, however:
However 1: given that all symptoms are positively intercorrelated, you could make the same argument for most symptoms. E.g., all symptoms much higher endorsed when you have insomnia vs if you don't have it.
However 2: while symptoms were much less frequent when not endorsing core symptoms of depression, they were far from 0. Therefore, substituting those with e.g. 0 (conceptually or statistically) is flawed.
Keep in mind that secondary symptoms present in absence of core symptoms are *meaningful* in this data.
When you take a self-report depression questionnaire & are asked about insomnia, you may indicate insomnia in the absence of depression (e.g. ongoing construction on street).
When you take a self-report depression questionnaire & are asked about insomnia, you may indicate insomnia in the absence of depression (e.g. ongoing construction on street).
This was not the case here: secondary symptoms indicated were only recorded by clinicians after ruling out things symptoms due to medical illness, symptoms due to medication, insomnia due to construction etc.
Hope you enjoy the piece—feedback very welcome. Big thanks to Ken Kendler & Steve Aggen for providing such an amazing dataset, Orla for leading this work over many years (we did it Orla!!! 👊), & my student Jelle for contributing to his first paper!
For those who don't have full text access, the PDF is (in line with the Taverne Amendment, thank you Dutch government) available on my website.
eiko-fried.com/publications
eiko-fried.com/publications
• • •
Missing some Tweet in this thread? You can try to
force a refresh
