Read on Twitter
ECT & depression review - Reading this review I was surprised that *none* of the included RCTs met the criterion
"One or more other treatments (antidepressants, CBT etc.) had been tried and did not work prior to ECT" https://www.bbc.co.uk/news/uk-52900074
"One or more other treatments (antidepressants, CBT etc.) had been tried and did not work prior to ECT" https://www.bbc.co.uk/news/uk-52900074
For 'Decliners described' - the criterion is "Any description of people who were approached but declined to
participate"
Why are studies failing on this rating if they had no decliners?
First one I checked is West https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1504188/pdf/bmjcred00643-0021.pdf any evidence of declining?
participate"
Why are studies failing on this rating if they had no decliners?
First one I checked is West https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1504188/pdf/bmjcred00643-0021.pdf any evidence of declining?
Or Lambourne & Gill 1978 https://www.cambridge.org/core/journals/the-british-journal-of-psychiatry/article/controlled-comparison-of-simulated-and-real-ect/20185DD2A8837C64A6950699E3CDB0D8 Is it fair to give 'negative score' here - evidence of screening but I cant see any evidence of decliners
Same is true for Freeman et al 1978 - it seems that Read et al are interpreting
'Decliners described' "Any description of people who were approached but declined to participate"
to mean paper says nothing about decliners whether there were any or not!
'Decliners described' "Any description of people who were approached but declined to participate"
to mean paper says nothing about decliners whether there were any or not!
How fair is it to give a negative score to ECT trials where there were no decliners described....because there were no decliners?
I mentioned earlier another criterion
They report that
all studies fail on "Any “Quality of Life” ratings administered & scores reported (e.g. HONOS)"- but all trials were completed before HONOS was even devised
How fair as a criterion to beat the ECT trials?
They report that
all studies fail on "Any “Quality of Life” ratings administered & scores reported (e.g. HONOS)"- but all trials were completed before HONOS was even devised
How fair as a criterion to beat the ECT trials?
Its not surprising that the inter-rate agreement is modest "This translates to a kappa score (which allows for agreement by chance) of .58, in the “fair to good” range (.40- .75; Fleiss, 1981)."
Though .58 is 'moderate agreement' (pace Cohen)
Though .58 is 'moderate agreement' (pace Cohen)
Another thought...if adding the Cochrane Risk of Bias (for kudos) - why not run the actual RoB tool rather what looks like a somewhat idiosyncratic version of it
Of course, one key feature of 'reporting selective outcomes' in the Cochrane Risk of Bias is the idea of registering intentions - the review authors did not apply it here to the RCTs - but then they didn't pre-register their review intentions either
Only small things perhaps, but they start add-up...I could never get a review published if I used only 1 search engine (as the authors here)
To be fair, this is not a meta-analytic review, nor is it even a *systematic* review - and even the failing on the latter alone would undermine any claims to warrant change of practice on basis of such a review
All studies rated as not being blind (in terms of having no previous ECT) - this appears to be applied in studies where it simply fails to say if any had received previous ECT - there is a difference IMO
Much fairer and normal practice in such reviews to have: Yes, No and undetermined or unsure
'No' implies that the reviewer is 'sure' and in some of the cited studies, this cannot be the case
'No' implies that the reviewer is 'sure' and in some of the cited studies, this cannot be the case
Ulett et al (1956), Brill et al (1959) and Harris & Robin (1960) - main criticisms are no inclusion of a depression measure - does anyone know what validated depression measures existed in late 50s?
Most serious weakness of Brill et al (1959) is "All males, mostly middle-aged"
Mostly middle-aged with at a mean age of 35? and a range from 18-68
Curious -Would a veterans hospital in the 50s have had female patients?
Mostly middle-aged with at a mean age of 35? and a range from 18-68
Curious -Would a veterans hospital in the 50s have had female patients?
Contrast with most serious criticism of Wilson, Vernon,
Guin, and Sandifer (1963)
"No older people, or men"
Guin, and Sandifer (1963)
"No older people, or men"
With regard to 'blinding' - is it possible that someone undergoing ECT could be blind to the intervention?
and is this then contrary to another criterion
"Any self-report or patient ratings administered and scores reported"?
and is this then contrary to another criterion
"Any self-report or patient ratings administered and scores reported"?