The Seriously Flawed Psychological Therapy and Dementia Study


Psychological therapy for people living with dementia (PLWD), is the way forward according to a study by Bell et al (2022) focussing on Improving Access to Psychological Therapy (IAPT) clients. The authors reported a 40% recovery rate for people living with dementia, compared to a recovery of 47% in the matched control group without PLWD, they claimed ‘our findings are important for encouraging referrals of PLWD into primary care psychological therapy services, as our work suggests that these services are likely to be useful in treating anxiety and depression in PLWD’. But their study did not demonstrate a real-world benefit for treating PLWD sufferers in IAPT. This is spin in favour of IAPT. The study was funded by the Alzheimer’s Society, who not surprisingly endorsed the author’s conclusion.

Don’t Psychoathologise The Reactions of People Living With Dementia

The twin outcome metrics were the PHQ9 and GAD7, and to enter the study people had to be initially at ‘caseness’ on one of these measures, a score of 10 or more on the former or 8 or more on the latter. Recovery was defined as being below the threshold for ‘caseness’ on both measures at the last treatment encounter. But data was only analysed for clients who completed 2 or more treatment sessions. Data from the most recent IAPT Annual Report (September 2022) , indicates that almost half (46%) of IAPT clients attend one or no treatment sessions. Thus if Bell et al (2022) had included in their analysis those who attended only one or less treatment sessions, i.e they had performed an intention to treat analysis, the true recovery rate would have been significantly less than that claimed. Strangely the authors make no mention of this.

Bell et al (2022) do not consider whether reliance on two self-report measures as the sole outcome metric is problematic. No mention that independent assessments may be important, as the self-report measures that are completed are subject to treating clinician scrutiny and comment. Further there is evidence that patients score lower on the PHQ9 in order to motivate themselves Robinson et. al (2017).  This leads to a mismatch between scores and patients global ratings of change ‘How are you feeling in comparison to 2 weeks ago? [(1=feels much better, 2 = feels better, 3= feels the same, 4 = feels worse, 5= feels much worse]). The within subject effect size for the PHQ9 was 0.83 and for the GAD7 0.80. A less than 1 standard deviation reduction is just what would be expected of an attention placebo control condition, (absent in the said study) involving regression to the mean, with people presenting initially at their worst. The blindness of the authors to these limitations is breathtaking.  The authors declare no conflict of interest, but fail to state that one of the authors works for Icope an IAPT service.

It is likely that volunteer supporters for PLWD would be more welcome than psychological therapists and at least as useful.  The Alzheimer’s Society should look to such a body of volunteers. This is not to say that there will not be PLWD who are suffering from depression and anxiety and need evidence based treatment, but there is no sign of it coming from IAPT any time soon.  IAPT has a remarkable ability to secure funds and endorsements without its therapists having a unique knowledge or training in this case of dementia. There is a credibility gap between IAPT and its’ fellow travellers  and what happens on the ground. 


Dr Mike Scott