“Rationing, however you want to try it, is an issue for all countries.”
If there is a quintessential British value, it is surely a belief in fair play. In the national psyche, queue-jumping ranks alongside major crimes, and it sometimes seems that variations in schooling, healthcare, and other local services are less acceptable than a poor service for all. There is even a special phrase, “postcode lottery” (postcode is the U.K. equivalent of zip code), to describe the outrage people feel when, for no reason other than where they live, some win and some lose.
It was outrage over postcode lotteries in healthcare that in 1999 gave rise to NICE, the National Institute for Clinical Excellence. Renamed the National Institute for Health and Clinical Excellence when the job of evaluating public health measures was added to its portfolio, this agency was assigned a task worthy of Solomon: to decide which medical treatments the National Health Service (NHS) should pay for—and which ones patients should be denied. Nominally, the new agency had three equally important jobs: to keep a lid on costs, to promote innovation in healthcare, and to ensure equity by standardizing decision-making across England and Wales (Scotland has its own agency). In reality, though, the political necessity of the last goal—equity—has become paramount.
Now, after almost a decade, how has it fared? What can other countries learn from the NICE experience? The verdict, in a word, is mixed. Professor Peter Smith, the director of the University of York’s Centre for Health Economics, says NICE has done pretty well with modest resources—and blazed a trail for health technology assessment—but suffers from several problems as well. “Whatever country I speak in, they have all heard of NICE,” he says. “Rationing, however you want to try it, is an issue for all countries.” But the agency is too slow, he says, and spends too little of its time evaluating old healthcare technologies to see which ones are obsolete. Both problems are the result of its limited staff and budget. He would like to see it take a more imaginative approach, with quick approval for obvious wins, or for companies willing to offer their products cheaply until more evidence of the benefits is available.
He cautions, though, that creating an independent agency like NICE cannot absolve politicians from the duty to make painful choices. Healthcare budgets are finite, so decisions to spend in one area inevitably mean cutbacks in another, however much politicians wish that were not the case.
These tensions could have been predicted from the start. In 1999, then-Health Secretary Alan Milburn faced questions about the case of a Sussex resident who suffered from motor-neurone disease, for which his doctor wanted to prescribe riluzole (Rilutek). If he had lived in neighboring Norfolk or Essex, he would have received it free, but in Suffolk the drug was not covered. NICE, said Milburn, would end such anomalies. What he did not say, though, was whether riluzole would become available everywhere, or nowhere. Nor, if the former, did he say which other treatments would be denied to balance the books.
The most high-profile part of NICE’s work is in deciding whether new health technologies are sufficiently good value for money to be provided on the NHS. The unit of measurement? The dreaded QALY, or “quality-adjusted life year.” Sure enough, the QALY is a creation of economists, in which a year of life is discounted based on the state of a patient’s health. A year spent in impaired health—no matter what the disease or cause of illness—counts for less; the more severe the impairment, the more the life year is discounted.
There is no explicit upper limit above which NICE regards the cost of a QALY as too high, in part because if there were one, it would, some fear, become regarded as a price target by pharmaceutical companies. Even so, seasoned observers say that the organization works to an implicit maximum of £30,000 (about $59,000) per QALY.
To date it has carried out more than 100 “technology assessments,” and another 50 are under way. Some compare a number of different treatments for the same disease and rule on which are better value for the money; some consider one treatment alone. The subjects can be suggested by anyone: members of the public, doctors, government ministers. It is, however, the Department of Health’s decision which ones are carried out.
The economic analyses use data provided by licence holders and manufacturers, usually under contract by health-economics groups in university departments. Once they are completed, a panel of experts meets to agree on guidance. If it can’t agree, drug companies and patient groups can make presentations. “Awful,” is how one observer describes these meetings. “They’re so confrontational. You get distraught patients standing up and shouting: ‘You’re condemning me to death.’”
Many NICE rulings are extremely controversial and have highlighted the inherent difficulty of its Solomonic brief. In 2005 there was an outcry when the agency ruled that none of four drugs licensed for the treatment of Alzheimer’s disease (donepezil, rivastigmine, galantamine, and memantine) offered sufficient benefit to be funded by the NHS. In 2006, after being provided with further trial data by the license holders, it relaxed this ruling and said three of them could be used—but only once the disease had progressed to “moderate” or “severe” levels. The fourth, memantine, was still regarded as not worth using for anyone.
Eisai, the Japanese pharmaceutical company which holds the license for donepezil (Aricept), promptly took NICE to court with the support of Pfizer, which markets the drug, and the Alzheimer’s Society registered as an interested party in the suit. It was an attempt to broaden the criteria for covering a drug so as to include patients in the earliest stages of dementia. But only one of the six points in Eisai’s suit was upheld, and that a minor, procedural one. The guidance still stands, and early-stage Alzheimer’s patients are not granted these drugs by the NHS.
Patients’ groups complained that NICE did not take enough account of the social, as opposed to clinical, benefits of delaying symptoms of Alzheimer’s; that by focusing on clinical cost-benefit analysis alone it ignored the significant benefits to others—both financial and emotional—that would result from delaying the onset of severe dementia, even if only by a little. This highlighted a common tension between the way in which NICE thinks and the way members of the public do. It seems natural to many people to argue for including the benefits to third parties in calculations of cost-effectiveness—but what about patients whom no one cares for? NICE’s approach values the Alzheimer’s sufferer who lives alone and has no friends or relatives just as highly as the one with a large and loving extended family: according to Sir Michael Rawlins, chairman of NICE since its inception, “A QALY is a QALY is a QALY.”
Another highly controversial NICE ruling concerned the use of ranibizumab (Lucentis), licensed by the European Medicines Agency in March 2007 to treat wet age-related macular degeneration (wet AMD). This disease, in which blood vessels grow abnormally and leak into the center of the retina, damaging sufferers’ central vision, is one of the most common causes of blindness, with around 26,000 new cases in the United Kingdom a year. The treatment, though, is expensive: around £1,000 for each monthly injection, with as many as 10 or 15 usually needed. And that is just for one eye.
NICE’s draft guidance, published in June 2007, was that only the most severely affected fifth of patients should be eligible for treatment on the NHS, and then only for their less-seriously damaged eye. This caused an uproar. NICE received 13,000 complaints on the topic, the most it had ever received on a single issue. And, sure enough, the commotion had an effect. In December it published revised guidance: all patients should be eligible, and for both eyes. It is now considering a proposal by the manufacturer, Genentech, to pay for up to 14 injections per patient, per eye, after which, if further treatment is needed, Genentech will provide the drug for free.
What really made this case stand out, though—apart from the way prioritizing the better eye and abandoning the other one highlighted the essential heartlessness of health economics so starkly—was an unusual conflict of interest. Genentech, the company that produces Lucentis, also produces a related drug, bevacizumab (Avastin), licensed in 2004 to treat colorectal cancer. It works by starving certain sorts of tumors of their blood supply, and when it first came on the market, ophthalmologists thought that perhaps it might be able to work on the abnormal growths in the eyes of wet-AMD sufferers. With nothing else to offer patients facing blindness at the time, they took to dividing up single doses of Avastin into as many as 40 parts, and injecting it, off-label, directly into sufferers’ eyes.
The treatment turned out to be not only gratifyingly effective but also cheap, costing as little as £20 per injection. But, frustratingly, NICE cannot evaluate the use of Avastin for wet AMD because it can only consider medicines for indications on their license. Genentech has no intention of applying for such a license, so, unusually, the NHS is carrying out its own head-to-head trial of Avastin versus Lucentis for wet AMD. If Avastin turns out to be an acceptable alternative to the pricey Lucentis, there may be a fudge of some sort to enable NICE to look at Avastin for this off-label use. Then the current arrangement might well be overturned—which may explain Genentech’s willingness to compromise and offer co-payment for patients to be treated with Lucentis. An ironic final twist in the tale: NICE has already ruled that Avastin is too expensive to be used on the NHS for the indication on its label.
In the United States, Genentech reached agreement last December on the issue of Avastin/Lucentis. The situation in the U.K. is more fraught, precisely because of the existence of NICE, which can’t rule on Avastin for wet AMD. If and when NICE rules that Lucentis is cost-effective, then ophthalmologists will have to use it rather than Avastin, because NICE’s rulings are compulsory.
As far as promoting innovation in healthcare is concerned—one of its core purposes—most observers agree that NICE has been a failure. This is for two reasons. The first is that improving the way doctors practice would mean persuading them to drop outmoded treatments—but NICE is already struggling to evaluate new drugs and treatment, and rarely gets around to looking at old ones. The second is that its technology assessments, although generally regarded as excellent, are inordinately slow, and this seriously delays the use of new treatments. A 2006 study by the Karolinska Institutet showed that new cancer drug use in the U.K. was consistently and significantly well below that of its European Union neighbors.
And this situation is only going to get worse. “I estimate that there are about 40 cancer drugs that will be licensed in the next two years,” says Karol Sikora, a consultant oncologist and professor of cancer medicine at Imperial College London. “NICE won’t be able to cope with this onslaught. And even if it could, it’s not going to accept any of them. The NICE evaluation made Avastin £88,000 per QALY [for the treatment of colorectal cancer], so you can’t use it on the NHS, but in every other European country you can.
“If a drug isn’t NICE-approved, that’s a great excuse for primary-care trusts [the local NHS budget-setters] not to pay for it,” he continues. And the time spent waiting for NICE to rule can be considerable. “There are currently four cancer drugs that have been licensed, and doctors want to prescribe them—but they’ve not been assessed by NICE and no one knows when they will be. It puts doctors in a difficult position when the drugs are approved and on the shelves, but we can only get them privately. We can see the toys, but we can’t touch them.”
Some worry, too, that by delaying widespread uptake of new drugs, NICE is obstructing the pharmaceuticals’ development pipeline. “So far there’s no evidence of this,” says Smith, “but it’s certainly a concern. There’s a debate to be had about the best ways to encourage innovation but at the same time ensuring that you don’t just hand the pharmaceutical industry an open check book.”
As for its third mandate, eliminating the dreaded postcode lottery, it has not come close to succeeding—not by a long shot. Even though its rulings are supposed to be mandatory, the drugs it approves can take some time to come into widespread use in the NHS, as primary care trusts cite budgetary constraints to delay implementing its guidelines or to introduce extra restrictions of their own. And even when its rulings are adhered to, that can mean other, less high-profile, treatments lose out. Treatments for mental illness and learning disabilities, all the things nobody lobbies for, get cut back to balance the books.
As for whether NICE is saving money, well, that was always a moving target. Its brief but eventful life has coincided with the growing trend of patients using the Internet to diagnose their own ailments as well as the use of expensive drugs like statins by large numbers. It has, however, had some success in driving hard bargains for the NHS, as with Lucentis for wet AMD. Another example is Velcade, a treatment for multiple myeloma. An initially negative assessment from NICE has been relaxed because of an agreement brokered by the manufacturer, Janssen-Cilag, under which the NHS will pay for four rounds of treatment for all suitable patients. After this, the patient’s level of serum M-protein, a marker for the disease, will be measured, and if it has fallen by half or more, the NHS will continue treatment; if not, the NHS will be reimbursed for what it has already spent.
And these are not the only examples of manufacturers giving ground in order to get on the right side of NICE. In December, Genentech wrote to oncologists to offer a price-capping arrangement for Avastin: they can apply for a refund for any patient who spends more then £14,000 a year on the drug, for any approved indication. Such innovative pricing is a way for manufacturers to keep up their list prices in order to stop parallel exports, but to still offer sufficiently good value that they are not shut out of the U.K. market.
More subtly, NICE has moved the question of healthcare costs into the open in a country that has long allowed its citizens to remain startlingly ignorant on the subject. In 2005 Andrew Dilnot, an economist at Oxford University, made a television program about the future of the NHS as a taxpayer-funded service, for which he followed the treatment of a truck driver who had been badly injured in a crash. The man spent many weeks in hospital, at first in intensive care and then, after numerous operations, undergoing intensive physiotherapy. None of it cost him a penny. Toward the end of the program Dilnot asked him how much he thought his treatment had cost the taxpayer. His answer underestimated the trust cost (£130,000) by almost an order of magnitude. “The NHS was set up to be ‘universal, comprehensive, and free,’” says Dilnot, “a charming idea, but naïve. There are decisions being made all the time about what treatments to offer. NICE has made it more obvious that this is going on.”
So what lessons can be drawn from the NICE experience?
First, whatever the carping, healthcare is, always and everywhere, a money pit, and in a state-funded system with limited resources, it is useful to make decisions about coverage on a rational, data-driven basis. The technical quality of NICE’s assessments, say doctors, is world-class. Even if they don’t appreciate its role in rationing medicines and healthcare technologies, they do appreciate having an independent, even-handed, and trustworthy body to turn to when they are weighing different courses of treatment. Second, those carrying out health technology assessments should be allowed to evaluate pharmaceuticals for off-label uses. Third, they should have a role in commissioning their own trials when drug companies, for commercial reasons, are not willing to do so, as with Avastin for wet AMD. Fourth, they should have some freedom to choose their brief, independent of political pressure, and a budget to carry out the necessary investigations. Fifth, they should be fast: Drugs that are likely to be major developments should be assessed for cost-effectiveness in parallel with licensing.
And a final lesson: the money should follow the health assessor’s decisions. In the U.K., with its taxpayer-funded system, there is no mechanism for bumping up the budget in response to costly central decisions; politicians, in sidestepping the hard decisions about which treatments to grant and which to deny, have also avoided the budgetary consequences of making those decisions. When NICE eventually ruled in favor of Herceptin, for breast cancer, some primary care trusts called for it also to rule on what other treatments should be dropped to balance the books. “Nobody has suggested what treatments we cut in favor of Herceptin—not the media, medical advocates of the drug, the courts who upheld patient appeals, or NICE. It would be especially interesting to know what the secretary of state for health would like us to cut,” wrote one group of doctors and health economists in the British Medical Journal at the time.
But a country where healthcare is funded at least in part by insurance premiums—whether state-run and subsidized, or not—could avoid this pitfall: a costly ruling from the health technology assessor could also trigger a proportional increase in premiums. Then everyone involved could have a NICE, grown-up conversation about costs and benefits—and the inexorable link between the two.
Helen Joyce is a Britain correspondent for The Economist, writing mostly on education, health, and science policy. Before joining The Economist in 2005 she edited Plus, a magazine about math and its applications published by the University of Cambridge, and Significance, the quarterly magazine of the Royal Statistical Society. She has a Ph.D. in mathematics from University College London.