1) As one of the authors of the meta analysis (Jauhar et al
http://bjp.rcpsych.org/content/204/1/20), at which you say you want to "poke holes", I will make a few comments about the meta analyses discussed. It is CBT clinicians/researchers running the RCTs who select symptom-reduction as the outcome- not meta-analysts. Indeed, of more than a dozen published meta analyses in this area, all have reported on symptom reduction, including NICE
2) I am not aware of any meta analysis showing that CBT benefits 'levels of satisfaction' or anxiety in psychosis – perhaps the author could reference this (it certainly wasn’t Wykes et al). Further regarding Wykes et al, those authors did examine "levels of functioning, and levels of low mood" - And found no impact of CBT on these outcomes in studies that they defined as having 'acceptable levels of 'quality' (see
http://schizophreniabulletin.oxfordjournals.org/content/34/3/523.long)
3) "The current evidence on CBT for psychosis being able to reduce positive symptoms is actually not as bad as some would have us believe, with comprehensive meta-analyses showing some reduction of positive symptoms (and those studies which informed the NICE guidelines)."
The 'comprehensive meta analysis' referred to here is Stafford et al
http://www.bmj.com/content/346/bmj.f185 is not, in fact, a meta analysis of people ‘with’ psychosis, but one examining if CBT prevents people ever developing psychosis. It is also factually incorrect to claim that this meta analysis produced evidence for "reduction of positive symptoms". Stafford et al are unequivocal in stating " There were no significant effects of this treatment on quality of life, symptoms of psychosis (total, positive, or negative), depression, or mania."
Regarding “those studies that informed the NICE guidelines”, the NICE meta analysis is hardly now or was ever 'worth the paper it was written on'. The ‘update’, on those who have psychosis, was published in Feb 2014, but includes studies only published up until 2008 - ignoring 6 years of generally better controlled largely negative studies. Moreover, the 2009 NICE meta analysis is massively flawed - it contains 110 separate meta analyses and so, by chance will, of course, find (spurious false positive) effects. Many of their so-called ‘meta analyses’ were run on just 2 or even 1 study -this is cherry-picking as an artform!
4) The article concludes " In my mind when we pick apart what CBT for psychosis aims to do and what it looks like when it's being done, measures of distress, level of functioning, and patient satisfaction in relation to their individual goals for therapy are perhaps better rulers to select than whether or not someone is still hearing a voice."
These may be better outcomes, but who can say? ...little or no research has addressed these outcome! Although NICE might talk about distress as a target for CBT for psychosis, they present zero evidence on it. CBT for psychosis advocates, researchers and clinicians rarely measure such alternatives (and in a few instances where distress has been examined, it is not impacted by CBT). CBT was designed to address symptom reduction and so, its unsurprising that it has been promoted by advocates as a quasi-neuroleptic. If the evidence goes against their 'own' focus on symptom reduction, it is hardly surprising that CBT advocates might want to attribute the focus to others and consequently change their target outcomes