Wednesday, 7 May 2008

NICE Drugs

The National Institute of Health and Clinical Excellence (NICE) guides on clinical practice through Technology Appraisals (does this drug/gadget/procedure work) and Clinical Guidelines (what does the evidence base from all research suggest should be done) which commissioners (some practices occasionally and Primary Care Trusts in particular) then accept as How It Shall Be.

As such, what NICE says matters. If NICE say a drug or treatment should be available, a Primary Care Trust better have good reasons why it isn't, if they don't fund it. To date, when challenged, I don't believe there have ever been good enough reasons (and patients have been supported in accessing whatever NICE suggest).

NICE isn't wholly independent, it's an arms-length government body. So it's kind of independent, but it can't bite the hand that feeds it. Naieve though I may be, I do see it as pretty independent (rather than peddling government policy) because the decisions and rationale given has shown this to be the case.

There have been a lot of discussions about the recent court ruling. The Times says the courts uphold limiting the drugs, with comment siding more with NICE. The Telegraph says the courts recognised NICE was flawed, with comment siding more with the drug companies. The Guardian gave a balanced view, the ruling upheld NICE guidance but the process was weak. The BBC comment was about a drug appeal victory that bemused me.
NICE have issued their own response here.

What is all this about?

Well, NICE look at evidence. If we just make it up as we go along, experienced clinicians could do that themselves. NICE exists to cut through that and guide us on what the best evidence at present suggests. NICE balance this with some opinion (from patients, carers, health and social care professionals, charities, support groups, drug companies, anyone with an interest). This then informs implementation guidelines they generate.

The evidence for Alzheimer's disease drugs is clear. They work. NICE acknowledged this in their first appraisal and in the subsequent review.
Drugs cost money. NICE work out quality life years and benefits that drugs can give, making a decision (that's the same on all appraisals, not specific to dementia care) on whether the bang for the buck is good enough. If quality of life is only modestly improved or if it costs squillions to achieve benefit, the cash in the NHS could be better used elsewhere and NICE won't recommend wholesale use of the drug.

The contentious bit is when to start and when to stop the drugs. You get good benefit in the middle stages of the disease. You sometimes get some benefit at the later stages, but not in cognitive domains (i.e. it can treat behavioural symptoms, it doesn't improve cognition).

Peterson's work in 1999 showed that mild cognitive impairment will often get better. In any given year, if you've mild cognitive problems (memory, orientation, concentration, sequencing, whatever), many get better, many stay the same and only 12% (1 in 8) go on to develop dementia. So 7 out of 8 don't, even 12 months on.
Ritchie in 2001 found conversion rates from mild cognitive impairment to dementia of 11% to 12% with a huge proportion recovering to normal. So, again, only 1 in 8 got worse, most got better.

NICE reckon if you're found to have mild cognitive impairement you shouldn't be on medication. I guess we'd all agree with that.
NICE reckon if you're found to have mild dementia you shouldn't be on medication. I disagree with this, I think once a diagnosis of dementia (not mild cognitive impairement) has been made, treatment should be started. In my corner we do, the PCT and GPs are happy with this, so all is well. Nationally this isn't the case and in many areas patients need to deteriorate to moderate dementia before drugs are offered. Therein lies the contention. In what other illness do you have to irreversibly and needlessly deteriorate before a helpful medication is started?

NICE say in dementia care (like most other guidance) that their guidance should apply to about 2 out of 3 patients. Thus, a lot of the time it won't apply. Given this latitude, if commissioners accepted and funded this too, all would be well.

Overall, the ruling isn't that relevant, to my mind. The courts said, don't change any guidance but do be more open about how you made your decisions.

This concerns me far, far, far more.

What NICE have had to do is give Big Pharma their algorithms for quality life measures and outcomes for whether a drug is cost effective or not. With this, a drug company can now have research and costing that, owning the algorithm, they can generate a result just below the line so NICE will recommend it. If it was well below the line, heck, they can charge more. If above the line, pay for "better" research. Effectively, drug companies can now generate answers for NICE that they know NICE will have to accept. As I started saying, what NICE says matters and is adopted as policy. Now, with this algorithm, Big Pharma can know that their product will have to be recommended by NICE. If Big Pharma can tell NICE that, using their own processes, their drug comes up as meriting widespread NHS use, drug companies now have a mechanism to indirectly affect national prescribing policy.

This concerns me much more than wrangles over MMSE and dementia cut off guidelines that any half savvy medic has been getting around, anyway.


PhD scientist said...

Yes, a ridiculous ruling handing enormous advantages to the Pharma companies at a stroke.

Personally I think reciprocity should apply here - if NICE has to tell the PharmaCos the details of its model, then why aren't the PharmaCos compelled to reveal exactly how much money they are making per dose sold? I for one would love to know the basis of their pricing - i.e. does herceptin really cost that much to make, or is it priced at "what the US market will bear"?

The opacity of the drug companies about pricing is already something they used to advantage in dissembling about immense profits they make on successful drugs, particularly in US markets. When US consumers complain about drug costs, the Pharma spokesmens' first response nowadays is "R&D is expensive, and people in markets with socialised medicine and central control of purchasing - like Europe, Canada and Australia - aren't paying their fair share." This plays well with some US consumers as they can then feel done down by European anti-free marketers.

The line is, of course, widely disbelieved by people in science who don't work for the drug industry. The widespread assumption there is that the PharmaCos make money in essentially all markets they choose to sell in, but make obscenely large profits in the US market.

And talking of the US market, I somehow doubt New Labour's friends at Kaiser Permanente will be handing Pfizer the details of their cost-benefit model...

Neil said...

Of course Big Pharma sell the product at the price the market will bear. 1. As I've seen some of the figures, and 2. Duh! They are a business, not a charity. Would you expect Coca Cola to do it any differently?

PhD scientist said...

Well, fine Neil. But then why should the state-mandated independent body trying to work out whether the drug really offers value for money be forced to reveal all the details of how they assess this? If they were a private insurance provider in the US they would not be compelled to reveal the details, for the same reason the Pharmas don't have to reveal the ins and outs of their pricing models.

Tainted_Halo said...

Hang on...
So Big Pharma will know how to demonstrate that their drugs are of most benefit - presumably this is by showing most benificence in well structured RCTs.

If the drug co's are going to do this - and NICE still have to replicate or at least validate the work - how is this a bad thing?

Ok.. I'm thinking of a number and I want you to write it down - But isn't it easier if I just tell you so you can write down what I want?

PhD scientist said...

tainted_halo: sadly "It doesn't work like that".

The underlying point is that the drugs are shown by the trials to be of marginal clinical benefit. The debate is all about whether the money saved by the drugs helping people to stay living independently (IF they do - note "marginal") outweighs the large cost of treating lots of people with a not-terribly effective drug. For instance, imagine that lots of people will have to be treated to get one of them to benefit by staying independent for one extra year (akin to "Number Needed to Treat"). Then the argument comes down to the costs and therefore to the cost model.