Since it was set up in 1999, the National Institute for Health and Clinical Excellence, or Nice, has rarely been out of the news. Turn to one newspaper and it is "the government's drug rationing watchdog," denying suffering patients access to drugs in order to save the NHS a few miserable pounds. Turn elsewhere and Nice is a brave attempt to bring rationality and best practice into the provision of healthcare.
When it opened for business (then just called the National Institute for Clinical Excellence) provision of healthcare in Britain was subject to a postcode lottery, with different drugs and treatments available in different regions (as it remains today). There was no way to evaluate what worked best, and there were often failures to speedily introduce new drugs and best practice. Nice was established to remove these inequalities and inefficiencies by providing "clear and robust advice" to NHS staff.
Since then, Nice's job has grown—in 2005 it took on responsibility for the promotion of good health. It now runs three centres—for health technology evaluation, clinical practice and public health excellence—employs 230 people in London and Manchester, and takes advice from 30 independent advisory groups on an annual budget of £30m.
Nearly all of the quarrels involving Nice have involved its appraisals of cost-effectiveness and whether the NHS should pay for a particular drug (if Nice says it should, the NHS must obey). Right now, Nice is embroiled in a row over drugs to treat dementia that it has judged that the NHS should not continue paying for. Patient groups are up in arms, as are drug companies Pfizer and Eisai, which have applied for a judicial review over this "wasted opportunity to provide best care for people with mild Alzheimer's."
But first, a few things that Nice does not do. It does not look at whether drugs are safe and effective; this is the responsibility of the European Medicines Agency, working in concert with Britain's Medicines and Healthcare products Regulatory Agency. And Nice does not set the price of drugs. The NHS has limited powers to cap profits on products sold to it through the Pharmaceutical Price Regulation Scheme (PPRS). And finally, Nice exercises less power in Scotland.
Alun Anderson: What makes Nice special?
Michael Rawlins: It is the first attempt in Britain to be explicit about the issue of cost-effectiveness in deciding what a health service can and cannot manage to do. And because we are one of the first countries to do this, Nice has attracted huge international interest. All healthcare systems are confronted with the same problems, and all of them are having to face the fact that they can't do everything that people want. The Australians can take credit for having started the ball rolling, as they have had a cost-effectiveness scheme for years, although it is limited to looking at new drugs. But Nice now has imitators in Germany, France and Greece, and possibly soon in Spain and Portugal. And there is a big debate in America over whether to set up something similar.
AA: Yet you are averse to the word "rationing," saying, "There is no role for Nice in the rationing of treatment to NHS patients"—a statement described as "weasel words" by health professor Alan Maynard.
MR: It is not that I am averse to the word rationing; it is because I remember postwar rationing. I still have my ration book. What we do in the NHS is not rationing as I remember it as a kid. Rationing meant that you got your allocation and that was it (it was sweets that concerned me). In the NHS you don't get an allocation of health service which you then use up. We keep on trying to look after you, irrespective of how much care you have had in the past. Insurance-based healthcare schemes, where you can spend up to a maximum and that's it, represent rationing as I understand it. If by "rationing" you simply mean that you need to prioritise, then use the word if you must.
AA: How do you measure cost-effectiveness to decide which treatments to make available and which not?
MR: We try to work out the value for money of different treatments. So we look at the costs of delivering a given treatment, including costs of drugs, whether you have to go into hospital to have the treatment and so on. Then we compare the cost with the health gain you get. What we are doing is making a comparison between this particular way of treating a patient and the other way, and seeing how much health gain you get from the treatment for how much extra cost.
There are two methods to measure health gain that we use, although one we very much prefer. We have in some instances looked at the "cost per life year gained," the cost in terms of incremental life years that a person on a certain treatment enjoys. This works for things that prolong life, but a lot of medicine is about improving the quality of life. So our preferred approach is to look wat the "quality-adjusted life year" (Qaly). A Qaly is a person's health-related quality of life graded from 0 to 1; if your treatment takes you from 0.6 to 0.8 for a year you have gained 0.2 of a Qaly. The measure works with life-saving treatments too, because if you are dead your quality of life is zero. Health economists have worked out lots of clever ways of measuring what the cost per Qaly is.
The difficulty is where to set the threshold between cost-ineffective and cost-effective. In theory, you could rank all your interventions, starting with the most cost-effective and going down to those that are least cost-effective, and then draw a line where you would run out of money. But this is not practical, and it also assumes that utilitarianism rules, that we want only the greatest good for the greatest number and do not take into account the question of equity. By equity I mean that Nice's decisions should seem fair in a way that a purely economic calculation—making a judgement through ranking treatments quantitively—may not be able to achieve. Health economists say that you could weight Qalys in some way, so-called equity-weighting, but no one has worked out how to do it.
We have said that any treatment below about £20,000 per Qaly should be considered as cost-effective, but once you start going above £25-£35,000 per Qaly then we need special reasons for approving it. The threshold is not absolute; on one occasion we approved a treatment at £48,000.
We achieve equity by asking our appraisal committee to use its judgement. For example, we looked at the flu drug Relenza. If you gave it to everyone with flu, you would typically get a day's reduction in the symptoms. The costs work out at £38,000 per Qaly. Now, Riluzole for motor neuron disease prolongs life and also costs £38,000 per Qaly. Both are expensive, but Riluzole delays the need for tracheotomy for six to nine months, and that is very important to those patients. So we said yes to Riluzole and no to Relenza. For a day's flu relief, you might be better spending money on a bottle of whisky.
AA: If there is a threshold of £25,000 per Qaly, then Nice is saying that a treatment that costs above £25,000 to provide an extra year of full health—or two years at half quality—may be too expensive to be approved. One extra year of my life is being valued at around £25,000. I feel it's worth more!
MR: Yes, but we need some measure. There is a finite amount of money available for healthcare. And Nice is not alone in putting a value on life. The department of transport values a human life at £1.38m. This enables them to decide whether putting in a traffic scheme that will save lives is cost-effective. If you convert the department of transport's cost per life saved back into cost per life year gained, it comes out around £32,000 —not so different from our figure. The World Bank has suggested the figure should be based on GDP per capita, which in Britain is around £18-19,000 a year.
AA: If you follow the health economist's approach, then surely some groups of people are worth spending more on than others. If I were a chain-smoking, hard drinking old hack with a dire life expectancy, should I have the same entitlement to health expenditure on a year's life as a young child?
MR: We have been concerned about this right from the start. We have to take into account the values of the society in which Nice operates. So what we have done is set up a citizen's council, a representative group of people from England and Wales. We advertised for the council and got 35,000 replies and 4,500 formal applications, from which we then selected 30 people. The council meets twice a year for three days at a time. We pose them questions and provide them with witnesses, engaging both sides of an argument. They really get immersed in the subject. When we asked them to look at whether there were circumstances in which age should be taken into account in Nice's decisions, we knew it would be a difficult problem, so we asked them to do two three-day sessions. Eventually they concluded that you should not take age into account: that there should be no difference whether a patient is aged five, 25 or 75. I had imagined that they would say children should get some preference, on the grounds that when you are in your seventies you have had a "fair innings," and you should pass some of the money on to the kids. I would have voted that way. But they heard the full "fair innings" argument, and they rejected it. Recently I talked to a group of Japanese who said that they might spend more on elderly people because of their respect for the old. One cannot necessarily translate values from one country to another.
The citizen's council also looked at smokers and rejected different treatment for them, because the association between ill health and smoking is statistical, not individual. Although statistically speaking a smoker is at greater risk, you don't know in any given case whether smoking is the main cause of disease. But you could insist that a patient give up smoking because it could stop a treatment working, or refuse a liver transplant to someone who won't stop drinking. And before we give out pills to help obese people, we ask that they try to lose weight by dieting.
AA: To what extent is the threshold for cost-effectiveness set by the NHS budget? Even with the threshold you have, wouldn't it be possible to approve so many treatments that the NHS would run out of money?
MR: No, the threshold does not reflect the budget. There is cost-effectiveness, and there is cost. They are different. We are specifically prohibited from looking at the budgetary impact of our decisions, so if a drug were cost-effective we would say it was cost-effective. If we made that judgement and a huge number of people turned out to need the treatment, we could bankrupt the health service. The government might then have to overrule us, or reconsider the size of its budget.
AA: So what we regard as "cost-effective" is rooted in the amount of money we are used to spending?
MR: Absolutely. That's why in Poland, for example, the threshold would be lower.
AA: You are on the record as saying that despite the rapid increase in British health spending to 9.3 per cent of GDP, it will have to rise further. And the consensus is that healthcare costs will rise ever higher. In the US, healthcare is already at 16 per cent of GDP and rising at more than double the rate of inflation. The pharmaceutical industry is creating fewer new drugs while charging sky-high prices. And it's hard to feel sympathy for the industry when the outgoing CEO of Pfizer, Hank McKinnell, is to receive a $200m payoff, despite Pfizer's market value slumping during his five years as boss. What can we look forward to?
MR: It is not working at present. We are not getting the affordable new drugs. Many people in the pharmaceutical industry agree. The industry is going to have to change. And it has got to happen fast, within the next five to seven years. There are lots of drivers for change: the companies themselves, regulatory authorities, and the people who pay—whether it is the NHS here or the Health Management Organisations in America—are going to say "we cannot afford this." Everyone will have to change their expectations. There is a lot of dialogue going on about this, but we have to accept that Britain is just a small player, just 4 per cent of the world market for pharmaceuticals. It is really the Americans who rule.
AA: There are many steps in creating a new drug. There's the laboratory discovery phase, toxicological testing in animals, phase I clinical trials that look at safety and dosage in healthy human volunteers, then phase II and phase III clinical trials that evaluate the effectiveness of the treatment on people with the disease. It takes more than a decade for a drug to travel from lab to doctor, and studies suggest that the average total bill for developing a new drug is $800m, with $450m going in the clinical phases. Where along this line can we look for savings?
MR: Twenty years ago, most drugs were developed in-house by big companies like Glaxo. Now a significant proportion are developed outside, often by academics backed by venture capitalists. The pharmaceutical company buys in at a later stage to develop the drug. So as well as the profits for the drug company, you've also got the cut for the venture capitalist.
One thing we must reform is the requirements of the drug regulatory authorities. Toxicology testing is an archaic process. A lot of the routine drug toxicology tests are not evidence-based at all. You might have thought that there would be a huge literature showing that tests work: that, for example, if funny little round cells develop in the liver of rats during the tests, we can assume that the drug is going to be toxic to liver cells in humans. But there isn't. Drug testing regulations are hard to change, though, as you have to get Japan, Europe and North America to agree.
My other criticism of the process is that the clinical trials have become bigger and bigger, and I don't think it has been worth it. I was horrified to discover that during the six years I was chairman of the Committee on Safety of Medicines [a statutory advisory body], the number of patients enrolled in clinical trials doubled. Companies feel they have got to set up bigger studies, and clinical trials are expensive. Regulatory authorities are nervous about safety, so want to see the drug in the widest possible group before agreeing to it. We are going to have to be smarter with clinical trials, using smaller numbers and Bayesian statistics that can get more information from smaller data sets.
Both the US Food and Drug Administration and the European Medicines Agency (EMEA) recognise the problem, but on the other hand, when you get a failure like Vioxx—an arthritis drug that Merck had to withdraw in 2004 because it increased the risk of heart attack—everyone starts to say that you need many more patients in the clinical trials before a drug goes on the market. If you want zero risk, then drugs will be more and more expensive. And if more time is spent on testing, then companies will have a shorter period of patent protection after the drug goes on the market. They will want to recoup their money during the period, so that jacks the price up too.
In the US, the amount of money spent on marketing new drugs is very high. As much as 30-35 per cent of turnover is spent on marketing. In Britain it is much less, as marketing is fairly tightly capped by the PPRS, but we are talking about a global industry in which we are only a small player.
AA: One common complaint about the pharmaceutical industry is that too much effort goes into developing very similar "me too" variants of existing drugs, or trying to modify drugs just to extend the life of their patents, rather than developing new drugs.
MR: Generally speaking, drug companies want a breakthrough because that blockbuster creams off a large market. On the other hand, a "me too" is easier to discover, as you know the mechanism; if it is a cholesterol-reducing statin, for example, you know which enzymes are inhibited and you know what to do. If you are a commercial organisation and your salary as the chief executive is coupled to the share price, then it is tempting to take a lesser risk by doing something not quite so innovative.
AA: We have heard a lot about personalised medicine—drugs matched to an individual's genetic make-up—as the next great development in medicine. How will that affect costs?
MR: Personalised medicine is likely to increase costs. A drug for me and a drug for you aren't going to be exactly the same, so there will have to be more work. The market for each drug will be smaller, so the profits will be smaller unless the price is higher. Some people claim that genetics will make it easier to develop drugs. I don't think that's right.
AA: Coming back to Nice and the media, you came under criticism when you appeared to be too slow in approving Herceptin, an expensive new breast cancer drug. Some patients took legal action against their primary care trusts (PCTs) because the drug was not available where they lived.
MR: We've been received pretty well by the media apart from the Daily Mail, which doesn't like anything we do—but the Daily Mail has an idiosyncratic view of healthcare. The media want news, so they get interested only when we say no to something. The appraisals programme [on drugs] is the bit that gets the attention, although I think it is our guidelines programme [on clinical practice] that really makes the difference. Whether or not particular drugs are available often does not make as big a difference as the whole pathway of care of individual patients.
In the case of Herceptin, patients did take action against their PCTs for not funding the drug. But we couldn't do anything until the regulators had decided to license it. Herceptin can cause heart failure and therefore you have to monitor patients with regular echocardiograms. That is a significant cost. But once we had the EMEA judgement, we fast-tracked the process and approved Herceptin in three weeks. That means that the NHS must offer it to eligible patients.
One of our faults has been that we are somewhat cumbersome and very rigorous. We do accept that you cannot wait for a year or two for Nice to make up its mind once a thing is on the market.
AA: Once Nice has provided an appraisal, it is not always enforced. One of the reasons for setting up Nice was to end the "postcode lottery" of healthcare, but it does not seem to have entirely gone away.
MR: That's right. There are two aspects to this. The first is that sometimes a drug we approve is an alternative to something else, so we can't make the new treatment mandatory. Second, though there is some postcode prescribing of anti-cancer drugs, a lot of it is not down to cost but to doctors not taking up the drugs. We are not good at taking up new things in medicine.
AA: Some doctors do complain that they can't keep up with all Nice's advice.
MR: Yes, that's why we produce a summary—but our summaries are too long too!
AA: Could a patient just go along to their doctor with the Nice guidelines?
MR:That's exactly why we produce our material in a patient-friendly format as well as for a professional audience. I want patients to go along and say, "Why aren't I being treated like this?"
AA: You have been running Nice since 1999. Do you feel confident that Nice is itself cost-effective?
MR: I don't know how to work it out. Healthcare has improved and death rates from cardiovascular disease and cancer are down, but many things have happened, not just Nice, so as an academic I would find it hard to pinpoint our contribution. Richard Smith, the former editor of the British Medical Journal, has been quite hostile to Nice, but recently said we were a great British cultural export along with Shakespeare, the Beatles and the Teletubbies.
