Here's the press release:
http://www3.imperial.ac.uk/newsandeventspggrp/imperialcollege/newssummary/news_18-10-2013-12-51-57
The study tried to be single blind. I'm not sure how well this could have overcome problems with response bias, considering the nature of the condition being treated, but kudos for trying imo.
Looks like reported differences diminished over-time, with standard care result improving, while +CBT results got worse. At 24 months, the difference is reduced to 2.24.
For their 'normal' anxiety outcome measure, I'm not sure if this means that there was not a significant difference at 24 months (their final time period). They do no explicitly say, and I've not got the appendix:
1 year, 27 (14%) of the 194 patients assessed who had received CBT HA had levels of health anxiety in the normal range (HAI score of 10 or less) compared with 14 (7%) in the 193 patients in the c ntrol group, with the odds ratio of achieving a normal level of health anxiety between CBT HA and the control group being 2·15 (95% CI 1·09–4·23, p=0·0273), with significant differences also being shown at 3 months and 6 months (appendix).
For the outcome measure they had regarded as their second most important, the use of medical resources, they excluded the patient who used the most resources from the CBT group, as an outlier.
Even so:
Total health and social care costs including the cost of the intervention were lower in the CBT
HA group (mean £7314) than the control group (£7727). In analyses adjusted for baseline cost, however, the adjusted mean difference between the two groups was £156 (95% CI –1446 to 1758, p=0·848). Although equivalence was not achieved, there was no evidence of a significant difference in cost between the CBT
HA and standard care groups. Imputation of missing data did not alter this finding.
In their protocol, they said:
The difference in mean costs between active treatment and control groups together with the 95% confidence interval will be derived. The equivalence margin is pre-specified as £150. The equivalence will be declared between active and control groups in terms of secondary economic outcome if the 95% confidence interval falls within (-£150, +£150).
In the BBC article one of the researchers said:
"Health anxiety is costly for healthcare providers and an effective treatment could potentially save money by reducing the need for unnecessary tests and emergency hospital admissions," Prof Tyrer said.
This is from their discussion:
No evidence of the effectiveness of CBT HA in terms of the secondary outcomes of social functioning or quality
of life was evident, with a corresponding lack of evidence of cost effectiveness in terms of quality adjusted life
years (QALYs; appendix). This might suggest a longer timeframe might be necessary to demonstrate the full
effects of improvements in health anxiety, particularly as complex and expensive investigations, even if mainly
activated by health anxiety, often cover a long time scale.
Also, their patients were those keenest on CBT:
A further weakness is that most of the patients who
were potentially eligible for the study declined to take part
and so the population treated might not be representative,
but we have no reason to believe that those who declined
to take part were fundamentally different from those who
agreed. As many people with hypochondriasis and health
anxiety attribute their bodily symptoms unequivocally to
medical pathology, 32 and therefore feel that only medical
expertise can help them, attitudes, both from staff and
patients, need to change before the treatment can be given
more widely. But if change does not occur, and standard
medical care fails to be aware of health anxiety, an
important, largely hidden, but eminently treatable cause
of morbidity in medical clinics is likely to persist.
Given their results, I'm not sure they have that much reason to be concerned about patients thinking that psychiatry has little to offer them. One of the authors of this paper already has a book out: 'Tackling Health Anxiety: A CBT Handbook'. Seems like a bit of a fumble to me.
Some sharp comments in the commentary:
At 12 month follow up in the current trial,6 30 patients (14%) receiving cognitive behaviour therapy and 16 (7%) patients in the standard care group had levels of health anxiety in the normal range, so screening of more than 28
000 patients attending hospital based medical clinics led to the recovery of 14 people who wouldn’t have got better with routine care. The cost of screening was not included in the analysis: combined with the absence of effect on quality of life that raises a question about cost
effectiveness.
Sounds like their answer is more lumping, including those with 'medically unexplained symptoms', alcohol problems, poor treatment adherence, etc.
Also, they say things like this:
For primary depression and anxiety, approaches such as cognitive behaviour therapy self help resources are now the recommended first step, 5 but similar developments have not generalised to physical health care.
Personally, I don't care what the recommended first step is, I care about what the evidence shows. So long as the NHS is still recommending homoeopathy as a treatment to patients, I don't really care much about their recommendations.
My summary: CBT for health anxiety is, even for those patients keenest on this intervention, of little use. No significant impact on QOL or health care use, and at 24 months was only associated with a HAI score 2.24 lower than those who had only received standard medical care (from baselines of 24·9 (4·2) and 25·1 (4·5) ). It could well be worthwhile for some patients, but it really doesn't warrant cheery headlines, or tut-tutting over patients disinterest in the effective psychological treatments available to them. This from the press release seems totally unwarranted: "The findings are good news for the 10 to 20 per cent of hospital patients who excessively worry that they have a serious, undiagnosed illness." The idea that these results would justify providing CBT to 10 to 20 per cent of hospital patients is just madness.
Really, I'm shocked by how ineffective this intervention was, considering the way in which those with hypochondria are often talked of in the medical literature. I had assumed that CBT was massively more effective than it seems to be, and based on this study, I'm not sure that we have any reason to think it would be preferable to 'befriending' or some other minimal intervention that avoids making claims of expertise.
The hype was nowhere near as bad as what we saw with PACE though, and I don't think that the way they presented their results was that bad. I wish researchers would be more willing to say: 'Oh dear, our treatment isn't very good. We'd expected it would do better than this.'