How to get doctors to do what is right.

Although most people think of British medicine as 'socialized medicine' a proportion of healthcare in the UK works just like it does in the US - it is insurance based. Often the insurance is provided by the employer, but taking out a family policy is certainly within the means of a large proportion of the population. The same doctors who operate in the NHS spend some of their time operating in the insurance based sector (we call it private practice over here), and the NHS contracts allow them to do so. In this sector, there is no restriction of what medicines may be prescribed and no rationing. Waiting lists are virtually non-existent. Rates of pay for the doctors are much higher than in the NHS and many private practitioners earn well over $2 million a year. Of course, if you operate in a field like pathology there isn't much demand for your services.

The big difference between the UK and the US is that only about 20% of Brits are insured compared to about 80% of Americans.

Both countries have taxation-funded safety nets for those who are not insured. In America it is called Medicare, Medicaid, the VA hospitals and County Hospital ER units; in the UK it is called the NHS. The strange thing is that the 20% using these safety net provisions in the US costs the American taxpayer a greater proportion of the GDP than the 80% of the population using the safety net costs the British taxpayer.

That sounds impossible, but it is probably down to the method of funding which requires 20% of the cost to be attributed to administration in the US compared to only 6% in the UK. It is also down to the way that doctors are paid in the US - by item of service whereas in the UK they are paid a salary on a sessional basis.

I remember my time in private practice. Chemotherapy regimens used to have a 28 day cycle. I was paid by the insurance companies around $800 for all the chemotherapy I administered over a 28 day period. More intensive 21 day cycles then came in. It meant that 6 courses could be administered in 18 weeks rather than 24 weeks; but I would only be paid for 4 28-day periods instead of 6. Where was the incentive to give the more intensive chemotherapy, which would inevitably mean more work for me in supportive care but for a smaller fee?

At one time the $800 fee was only for intravenous chemotherapy; where is the incentive to give it by tablet? I could justify myself by saying that if I gave it intravenously I could be sure it had been taken, but this sort of thing left a nasty taste in my mouth (whereas the patient avoided the nasty taste).

Welcome to the world of perverse incentives.

The only type of physician to make a large income from private practice were those with a procedure. Gastroenterologists had gastroscopies and colonoscopies. Respiratory physicians had bronchoscopies. Cardiologists had cardiac catheterizations. Dermatologists had skin biopsies and liquid nitrogen therapies. The poor old geriatrician or pediatrician had none of these.

A recent Op-ed article in the New York Times was entitled "Paying Doctors to Ignore Patients". It gave the example of a patient check-up that included an X-ray, a urinalysis and a physical. Three procedures: three fees. Each fee covers not only the doctor's time and skill, but also his overhead. If he buys an expensive bit of kit like a CT scanner, once the thing is paid for, the high overhead fee attracted becomes pure profit. It pays him to do as many CT scans as he can. In fact the best way to make money is to do as many expensive procedure as possible and as few cheap ones (like simple consults) as he can get away with.

This year the spending on physician services in America will top $500 billion. Doctors who own their own CT scanning equipment order two to eight times more scans than those you do not own their own equipment. A good question to ask if you are offered one is, "Do you own the scanner?" A 2002 study suggested that doctors are ordering roughly $40 billion worth of unnecessary imaging tests each year.

One way of reducing costs is to pay doctors a standard stipend for each patient, graded according to the severity of their conditions. The incentive is then to do as little as possible for each patient and patients might suffer. There is a happy medium where the patient has the right number of tests to make the diagnosis without wasting money on unnecessary ones; but how do you incentivize doctors to do this?

A possible way is to look at outcomes. It turns out that this is quite complex. I spent a couple of years trying to devise outcome measures for hematological patients in the NHS. At the end of our committee's lifespan we decided it was impossible; there was just too much variation between cases. It's all very well for gall bladder operations - you could easily set a tariff for them, but every admission for AML is different. The cheapest way of managing it was to kill them off in the first week. Those who had a good CR could be reasonably inexpensive, but the cases that had short remissions and several courses of treatment including an allograft, cost the earth. You couldn't guarantee that any particular unit would have the same casemix as any other unit. A unit that had 67% long term survivors might actually be doing worse than a unit with 35% long term survivors because of differences in casemix. At that time we couldn't even differentiate between cases in terms of chromosomes, let alone flt3 or NAM status.

In another field, a famous study showed that the response to publishing the outcomes of cardiac surgery in new York Hospitals, was an unwillingness to accept referrals of more complex cases.

So what does incentivize doctors? You could publish outcomes but experience suggests that this is ineffective. Often the providers 'game' the system - that is manipulating practice to improve what is recorded without improving the quality of service. A good example would be the target of patients spending no longer than 4 hours in the ER, reached by redesignating the trolley that the patient is lying on as a 'bed' and the corridor where the trolley is parked as a 'ward'. However, even when gaming takes place there is a tendency for actual performance also to improve. If they take the risk in gaming the system, they have certainly been incentivized.

It might be thought that the wallet is the greatest incentive to better performance, but this is often not the case. We all know that in Canada it is a sense of altruism that makes doctors perform better - though most doctors for whom this motive predominates have already joined Médecins Sans Frontières. A sense of Professionalism might dominate - there is a sense that belonging to a professional group compels compliance with a set of values that does not depend on personal 'goodness'.

If outcomes are measured it does instill a sense of competition amongst providers. The cardiac surgery Olympics would be one where everyone would covet the gold medal. But as we know, there are bound to be some taking steroids. However, when an institution regularly performs in the top ten all sorts of rewards accrue from patient choice to recruitment of young graduates. Even those who can't be Harvard or Yale would seek to avoid censure.

There are individual incentives within an organization such as promotion to more responsible positions, better health insurance, better company car, longer holidays and less obvious benefits like less managerial oversight and more freedom to operate without constraint.

Outcome measurement is here to stay. The problem is that it is much easier to measure process than outcome and the measurers themselves certainly need to up their game.

What do patients really want? Very few want to be charging around the country seeking the Manchester United of oncologists, only to switch to Liverpool when they are top of next year's league tables. What they really need to know is that my local hospital is performing well, that it makes the right decisions, has access to the latest drugs and operations and that I will be treated well when I get there.


I thank my son Richard for his research in this area and for helpful discussion as I was thinking about this article.