Blinkers for bad news

The past month has seen the controversy around Avandia (rosiglitazone) and its possible side effects go through more twists and turns than a soap opera (see ’Does GSK have a case of the Vioxx?’). There’s room for a good argument on both sides of the debate – but there are much broader implications for the pharmaceutical industry beyond this one drug. 

You could argue that if a widely-used drug, which is going into a population already at risk of cardiovascular trouble, carries a risk of exacerbating that very condition – well, people need to know about it and then act appropriately. And if the only way to find this out is to resort to a meta-analysis of existing studies, then that’s a reasonable price to pay for a timely warning. Someone needs to do it, if the industry and the regulatory agencies aren’t going to. 

Or, you can respond by saying that in a population with existing cardiovascular risk, meta-analysis may not be the statistical tool needed to see if that risk changes slightly after taking this particular drug. If the largest studies in the mix are negative, a contrary result obtained by combining the smallest ones needs to be viewed with some scepticism. And with trials already years underway to answer the specific cardiovascular question hanging over this treatment, is it better to wait for a solid answer or to sound the alarm now? 

I lean more toward the second view, but I do agree that the more sets of trained eyes looking at a data set, the better. It’s also crucial that some of those eyes should be outside the organisation that generated the data in the first place. A company should be sensitive to signs of trouble with its own drugs (fear of lawyers, you’d think, would be enough to ensure that), but wishful groupthink can keep everyone believing that looking for the bad news is someone else’s responsibility. It’s worth noting that the drug industry has no particular concentration of this problem; it’s one of those things that unfortunately goes hand in hand with the presence of human beings. 

Whatever your view, the situation highlights much wider concerns. One of the basic reasons for the drug industry’s profits, and for its problems, is that it’s very hard to be dispassionate about health care. But we need to be. A hyper-rational alien being, looking at Avandia (or indeed Vioxx) would first try to calculate the risk of harm from the drug in question, and then try to balance that against the good it could do. 

It wouldn’t be easy for the alien, and it’s even harder for us. How do you read the scales when weighing the value of pain relief, in the case of Vioxx (or glucose control, with Avandia) versus a small (but proven, in the case of Vioxx) risk of heart attack? These things aren’t even in the same units. 

But I’d argue that we’re going to need to work out some conversion factors, because side effects are not going to go away. Anyone who has done drug research knows what crude tools even our best pharmaceuticals are. Combine that with the enormous human diversity of the patients taking them, and something bad is always going to happen, somewhere, to someone. We’re minimising that, compared to how things used to be. It can be hard to see the effects of the work, though, because we’re also improving our ability to recognise trouble, and we’re becoming a lot more sensitive to it when we do uncover anything untoward. 

No one wants to hear that every drug has side effects, or to think about what that phrase means. The general public may have trouble getting past the first lines of the Hippocratic Oath: primum non nocere. If our rule is going to be ’first, do no harm’ – no harm at all, ever, to anyone – we’re not going to be able to keep a lot of modern medicine. Physicians tend to realise this very early in their careers, but not everyone has the benefit – or burden – of their experience. 

So we’re back to the human tendency to avoid looking for bad news. In the same way that companies have to be hard-headed about even their most cherished and profitable products, people need to face up to what those products can and can’t do. It might be prettier to stay in the world of advertisements, where everything works quickly and cleanly. Or maybe in the world of trial lawyers, where nothing is ever supposed to go wrong and it’s someone’s direct fault if it does. But we can’t. We live here. 

Derek Lowe is a medicinal chemist working in preclinical drug discovery