Sunday 16 August 2009

Fun with numbers

An article in The Times last week contained the following statement:

... five-year survival rates for prostate cancer in America are more than 90 per cent, compared with just over 50 per cent in Britain — at least according to the 2008 Concord study.

Here's a reader's letter rebutting it:

Sir, Chris Ayres (Commentary, Aug 13) cites prostate cancer survival rates to demonstrate the superiority of the US healthcare system over the NHS. A better comparator is death rates from prostate cancer, which are 322 per 100,000 in the US and 254 per 100,000 in the UK (OECD Health Data, 2004).

The higher survival rate from prostate cancer in the US is a function of the more extensive use of PSA screening there than in the UK. The value of PSA screening is debatable but it results in the detection of many cases of prostate cancer that would not in any event lead to death, and so shows higher survival rates...

Ian J Hartill, Chandler’s Ford, Hants.


I suppose the best way to envisage this is to imagine that [exactly the same number] out of 100,000 men in each country get the disease, half the time it would be fatal (without treatment) and half of the time it wouldn't be fatal.

The Yanks love going to the doctor for check-ups, so all their cases are diagnosed, including the non-fatal cases, which increases the average life-expectancy of all men who are known to have the disease, but not that of those with the fatal type disease.

We Brits on the other hand don't like going to the doctor, especially if it involves having something shoved up our bums. So the disease only gets diagnosed if it's more serious, which by definition is more likely to kill you.

So the 'survival rate' (from the first excerpt) is indeed lower in the USA (as it includes considerably more non-fatal cases) and the number of deaths attributed to the disease is lower in the UK (because it excludes those men who just grin and bare it and then die without ever being diagnosed).

11 comments:

PeterJ said...

An additional problem is that the PSA test isn't really diagnostic, and produces unnecessary treatment in the US which wouldn't happen here. These false positives, where patients wouldn't develop prostate cancer in their lifetimes if left alone, are counted as prostate cancer survivors.

There's strong resistance to PSA screening from doctors in the UK because of this, as unnecessary treatment can lead to patients suffering lifelong incontinence, impotence or other side effects for no good reason.

Pogo said...

It's also worth noting that virtually every man over the age of about 65 has some degree of prostate cancer and that every one who dies after that age will therefore die with prostate cancer, but not necessarily from it.

If the medical system in the USA manages to keep a proportion of men alive that in the UK would have died from other reasons - eg most of the other "faster or more serious" cancers - it's not unreasonable to assume that eventually more of them are going to die from slow-moving prostate cancer.

Mark Wadsworth said...

PJ, Pogo, that all makes it even more difficult to decide how to measure which system is "better", if you control for all these variables.

PeterJ said...

Mark, it certainly makes prostate cancer a bad yardstick for any comparison. Perhaps that's why it is chosen as one by people trying to put together a particular general case?

I'll have to try looking up the figures for other cancers both with and without screening issues (breast cancer and kidney cancer, say).

neil craig said...

The normal economist's assumption is that the demand for a free good is infinite & it thus appears surprising that Americans, who pay for it, are more likely to go to a doctor than Brits who get it for free.

It suggests there is a conflicting rule that, The Americans have paid for it through insurance. It suggests there is a conflicting rule that, once people feel they have paid for a good they will go to much greater lengths to make sure they get it but that we don't value something free.

Mark Wadsworth said...

PeterJ, I look forward to your findings. But whatever you choose, it is always a moving target (as with prostate cancer), because the likelihood of the condition being treated successfully changes people's behaviour, and hence increases the likelihood of treatment being required.

For example, if there were no 'cure' for diabetes, then diabetics would die much younger and have far fewer children (or voluntarily choose not to have children). But diabetics can be helped reasonably cheaply, so they live longer and have more children, who in turn are more likely to be diabetics.

NC, it's a quandary. But we pay for it via taxes and are always told it's "free" (so OT1H would be expected to overconsume, but as it's "free" we don't value it - hence all the missed appointments).

In most other countries it's paid for via insurance (even if the insurance is really a tax) and so people think "I've paid for it, I might as well use it" - it is still more-or-less free at the point of use.

PeterJ said...

An early trawl through some available statistics suggests that this is going to be a long job - there is so much to control for that a single meaningful comparison figure for any individual cancer looks impossible to reach. General incidence, gender, race, age, nutrition, early diagnosis rates, availability of primary care... all of them vary in the statistics from place to place within countries, let alone from country to country.

Still, I did confirm that selective single comparison figures are being cherry-picked for political purposes as I suspected.

Anonymous said...

I don't like going to the GP for the simple fact I can never get to see the GP I want to see. Our family GP is never in on a Wednesday and in Jan/Feb takes six weeks off to go snowboarding in Canada.

If I'm ill, I can phone the surgery first thing in the morning and if I'm lucky get to see a GP the same day. If I want to see my GP of choice, it is usually 10 days later.

Dan Hannan is right, the NHS is rubbish.

Dick Puddlecote said...

We Brits on the other hand don't like going to the doctor.

Yep. I don't go to my doctor unless I'm in mortal agony (so never) anymore. Simply because before I get to tell them what is wrong with me, they have looked at my file and badgered and hectored me on lifestyle choices to the point I wished I'd not bothered.

If I was paying for it directly and could bugger off elsewhere, I imagine the welcome would be more warm.

I tend to bang on about such things ;-)

AntiCitizenOne said...

Yep, visiting the Doctor isn't completely "free" if you count your time, frustration, flexibility.

Visiting a GP is also not very frustration free and the NSH "mess around" is there to control demand.

James Higham said...

Which is the reason stats should be viewed with deep suspicion. Also, what is the sample size?