Guidelines for Depression: More Gourmet Nights from NICE

Could it really still be Gourmet Night for Psychological Therapies? I have written previously comparing IAPT to Fawlty Towers’ Gourmet Night, in which a series of events led to the hotel restaurant offering three options on the menu: duck with orange, duck with cherries or ‘duck surprise’ (duck without either orange or cherries). As a result of lobbying becoming increasingly abstracted from the critical nuances of psychological therapy research in the noughties, the ‘Increasing Access to Psychological Therapies’ service (IAPT) established in 2008 ended up limiting choice of therapies to telephone CBT, computerised CBT or face-to-face CBT.  Whilst some IAPT services have expanded their choice of therapies a little, a look at the new NICE guideline for depression currently under consultation suggests IAPT (if it follows NICE) could be heading for a miserly change to the Gourmet Night menu by suggesting clients share their duck with up to 12 other customers: group-CBT.

IAPT was driven by a narrow interpretation of the NICE 2004 systematic review of treatments for depression.  The erroneous argument put forward by Layard and fellow CBT lobbyists was that a massive expansion of CBT services would reduce the nation’s welfare bill associated with mental illness, recouping the initial start-up cost of £170million.  This logic was somewhat misplaced, as NICE had employed standard systematic review methodologies uncritically with a narrow focus on ‘symptoms’ (like sadness, pessimism, worthlessness, agitation and frequency of crying ), assuming the same conclusions could then be applied to employment or functioning outcomes (like being employed, ability to deal with problems, being able to take care of oneself, feeling independent, feeling able to communicate with others, feeling able to get about, ability to attend education, work or social events). This assumption was wrong: an analysis of these outcomes would have painted a very different picture.

There were also a range of other, serious issues with the 2004 NICE review methodology as applied to psychotherapies for depression which myself and others have commented on elsewhere. Unfortunately it seems that the new NICE guideline for depression currently in draft form, will be making many of the same mistakes. As well as the problem of symptom versus functioning focus, I will highlight two more key problems: glossing over illness severity and ignoring the longevity of outcomes.

“If you don’t like duck, you’re rather stuck”

IAPT reports figures of around 50% of clients ‘recovering’, leaving the remaining 50% with few further options. However, there is also evidence that more than a third of IAPT clients have already received treatment in IAPT (i.e. they relapsed); and that many patients who initially recover will relapse or present to other services. IAPT figures for return to work are not routinely analysed but my own analysis of publicly available raw data suggests this may be less than 5%. It is important to bear in mind that the unquestioning linear extrapolation of NICE symptom-focused reviews into NHS policy fails to consider where on the depression severity spectrum NHS money is best spent. For example, IAPT does not get people back to work because its clients are generally not out of work. The evidence IAPT was based on were trials of mild and moderate depression. Yet it is people with severe chronic depression who are more likely to be out of work. The new NICE review does categorise research by duration of depression, severity and complexity but their methods of categorising were devised within the guideline development group and not subject to clinical or scientific review. Moreover, their meta-analyses do not take into account baseline severity when summarising the mean effect of treatments. This is important because it is much harder for a treatment to move a person with severe complex depression into an arbitrary category of ‘recovery’, than a person who starts out with mild or moderate depression. Treatments which are achieving quite good effect sizes from very severe starting points are not being credited by NICE because they are not moving enough participants all the way into the least severe band of severity. See my response to the NICE consultation for further explanation.

It’s a long term condition, stupid

Given that psychological difficulties (severe enough to have societal cost implications) are chronic, treatments ought to be only considered helpful if they show that benefits can last beyond the end of treatment.  Unfortunately, the new NICE review continues to insist on taking the treatment end point as the outcome of interest. This is presumably because most RCTs do not report follow up data – and yet the lack of availability of suitable evidence for treatment longevity should not be glossed over and replaced with findings from inadequate research galore which has failed to collect or report on follow up data for what is quite clearly a long-term condition. This would not be (and is not) acceptable for evidence used in NICE guidelines for other long term conditions such as diabetes or epilepsy. Where the longevity of symptom outcomes are ignored by NICE, the implications for the quality of health economic analysis are also serious, and these are further compounded by ignoring employment and functioning outcomes and baseline severity.

Of course the NHS should try to be more cost efficient and cost-effective through evidence synthesis. But low-cost therapies selected on the basis of a narrow vision of evidence hierarchy based on a disease model of illness is not the best way to do this in mental health. NICE needs to seriously rethink their approach to the new depression guideline if it is to be useful for clinicians and managers wishing to plug the evidence-policy gap.