Is David Cameron right to claim that people are dying because they cannot get access to cancer drugs?

He made the claim in the first TV debate last Thursday. Specifically, what he said was:  “I have a man in my constituency called Clive Stone who had kidney cancer who came to see me with seven others. Tragically, two of them have died because they couldn't get the drug Sutent that they wanted, that was on the market, that people knew was a good drug. That's a scandal in our country today.”

Sutent (sunitinib) is a kidney cancer drug that was appraised by the National Institute for health and Clinical Excellence (NICE) in February 2009. NICE concluded that for some patient groups, but not others, sunitinib appears to offer benefits compared with existing treatments in terms of overall survival, progression-free survival and tumour response. For people with a poor prognosis and those unsuitable for the standard treatment (immunotherapy) there was limited evidence and no conclusions about the clinical effectiveness of sunitinib as a first-line treatment in these groups could be made.

NICE approved sunitinib for use on the NHS for those patients in whom it showed a benefit. The fact that the seven patients who visited Cameron were not getting it indicates that they belonged to other groups for whom there was limited evidence of benefit. They may indeed have wanted the drug, but it is stretching a point to say that two of them had died because they didn’t get it. If NICE is right, indeed, there is no basis in evidence for making such a claim.

However, let us assume for the sake of analysis that the benefits had applied to them. Would Cameron’s remarks then have been justified? 

Professor David Spiegelhalter has made an interesting analysis of this question on the website Understanding Uncertainty. He points out that that survival benefits of drugs are expressed as a hazard ratio. If a drug has a hazard ratio of 0.8, then a patient taking it has 80 per cent of the chance of dying in the following month as he would if he were not taking it.

Furthermore, the hazard ratio can be used to work out the chances of an individual getting a drug outliving one not getting it. This is the result of a simple calculation. If the hazard ratio is expressed as h, then the probability of anybody getting the drug outliving anybody not getting it is 1/(1+h). (The proof is on the Understanding Uncertainty website.)

So how does this apply to sunitinib? Median survival time on the drug (for those it is appropriate for) is 37 months, without it 27 months. From this, making some assumptions, Professor Spiegelhalter works out a hazard ratio of 0.73 (27 divided by 37). From this it can be calculated that the chances of anybody getting the drug outliving anybody not getting it is 58 per cent (1/1.73).

This means that even a drug that on average provides a reasonable survival advantage cannot be guaranteed to do so for every patient. 

Cameron’s claims were therefore wrong in two ways. First, for the particular patients he identified, there was little evidence of possible benefit. And second, even if there had been, you cannot say that any given patient will benefit even from a drug known to be effective. While Cameron can be forgiven for not knowing the second reason, the first ought to have been clear enough. 

His sympathies were properly engaged; but to call it “a scandal in our country” was certainly overstating the case.