What efficacy measures are clinically relevant and should be used in Cochrane reviews of acute migraine trials? An alternative viewpoint

Abstract

I briefly review the purposes of efficacy measures, which go far beyond supporting new drug development. I use vignettes to illustrate the importance of functional recovery during the migraine attack, and argue that headache relief provides this. Sustained headache relief (SHR) is therefore a very worthwhile outcome when the alternative is a day of debilitating pain. As a measure, SHR may not be ideal for new drug development but it is informative to individuals, health care providers and politicians, and serves cost-effectiveness analysis better than any other. Cochrane are absolutely right to use it in systematic reviews along with the IHS-recommended measures.

Introduction

The clinical officer greeted his patient: “What can I do for you, Mr Owusu?”

“Some mornings I wake up and I have very bad pain in my head. But I have to plough my field. It goes bang, bang, bang in my head. It hurts me more when I move, and the daylight hurts my eyes, but still I have to plough my field. Can you give me something to make it go away?”

“I don’t know. I will try.”

“If you cannot make it go away, please give me something to make it less. If the pain is only a bit bad instead of going bang, bang, bang in my head, even if it stays all day, I could see to steer my oxen, and I could plough my field. That is what I need to do.”

*****

“Good morning, Mrs Smith. How were the new tablets?”

“Better than the last lot, Doctor. They got rid of that awful thumping. But I still had a dull headache for most of the day. Haven’t you got anything stronger?”

“You’ve tried everything. Would you like some more of the same?”

“Well, they did make a big difference. I was able to get the children off to school, and do the shopping – and most of the housework. I’d have spent the day in bed otherwise.”

Viewpoint

I entirely agree with Tfelt-Hansen (1) that rigorous efficacy measures are needed for new drug development; that new and usually more expensive drugs must demonstrate clearly that they outclass existing treatments. I agree that the International Headache Society (IHS)-recommended measures – pain free at two hours (PF2) and sustained pain freedom over 24 hours (SPF) – show that current treatments are far from perfect, and for this reason there is a need for new and better drugs.

Mr Owusu cares nothing for these matters. No new drugs will reach him in his lifetime.

Efficacy measures have other purposes than supporting new drug development, of more immediate relevance to most people with headache. Of course people with headache want to be pain free quickly, and to remain pain free, but in a realistic world it is not possible to measure efficacy only against a standard of perfection. When people with headache are trying to choose for themselves between currently available treatments, how are they helped by a measure that tells them they have a one in eight chance of achieving a particular outcome (1)? When politicians are asked to allocate resources, will they do so to treatments with numbers needed to treat (NNTs) of eight? Why, when they yield so little insight into how current treatments perform in people, should PF2 and SPF be the only measures that Cochrane reports?

While pain is unpleasant, most people – like Mr Owusu and Mrs Smith – are more concerned that it interrupts function. Health states that are the sequelae of diseases are rated and compared in terms of the disability they cause (2,3). This approach can be questioned, but it has a very large and respectable body of opinion behind it (4), and cannot be dismissed. In the case of headache, mild pain, as is achieved in sustained headache relief (SHR), clearly has less functional consequence than moderate or severe pain: In the Global Burden of Disease Study 2010, the ‘average’ (moderate to severe) ictal pain of migraine was accorded a disability weight of 0.433; the ‘average’ (mild) pain of tension-type headache only 0.04 (5). Health policy is informed by cost-effectiveness analysis, which cannot put a monetary value to pain, but can to disability and its consequence of lost productive time. For this purpose it can be argued that mild pain is not disabling, that headache relief therefore equates to functional recovery, and SHR to functional recovery for the whole duration of the attack.

Which makes it a curious and narrow view that SHR lacks clinical relevance, not to be used even as a secondary efficacy measure in systematic reviews (1). SHR may be an imperfect outcome, but for Mr Owusu and Mrs Smith – and anybody anywhere who has to get up and do a day’s work – it is a very worthwhile outcome when the alternative is a day of debilitating pain.

Tfelt-Hansen comments that Cochrane should ‘not even be suspected of wrongdoings’ (1). The statement invites the unfortunate inference that Cochrane has left itself open to being suspected of such. This is unfair. In its reviews, Cochrane can, of course, report only what is available to be reviewed. The policy in the past of the Pain, Palliative and Supportive Care Review Group (PaPaS) was to review virtually every outcome available, whatever its worth, leaving the reader to judge what was of value. This led to somewhat indigestible reviews. PaPaS has moved during the last three years towards a selective policy, including in the main text only those outcomes considered most important and putting others that might be of interest in appendices. For acute headache outcomes, PaPaS does not entirely follow IHS recommendations and does not state that it does; as main-text efficacy measures it takes PF2 and SPF from IHS (6), together with headache relief at two hours, and adds SHR for the several very good reasons described above. Tfelt-Hansen may disagree with this, but he has no argument that Cochrane does wrong.

Conclusion

Efficacy measures serve a variety of purposes. Cochrane, and PaPaS in particular, rightly take a broad view of these in their reviews since Cochrane’s purpose is to inform health care (at all levels) and health policy (7). SHR has its limitations, and may not be the best measure in new drug development, but it is hugely clinically relevant, informative to individuals, health care providers and politicians, and serves cost-effectiveness analysis better than any other measure so far proposed.

Clinical implications

Sustained headache relief is highly clinically relevant to most people with migraine.

It is an important efficacy measure for many purposes (though not all), and Cochrane should continue to include it among those evaluated.