30 July 2009

Transparency Saves Lives

Here's a wonderful demonstration that the simple fact of transparency can dramatically alter outcomes - and, in this case, save lives:

Outcomes for adult cardiac patients in the UK have improved significantly since publication of information on death rates, research suggests.

The study also found more elderly and high-risk patients were now being treated, despite fears surgeons would not want to take them on.

It is based on analysis of more than 400,000 operations by the Society for Cardiothoracic Surgery.

Fortunately, people are drawing the right conclusions:

Experts said all surgical specialties should now publish data on death rates.

Follow me @glynmoody on Twitter @glynmoody and identi.ca.

4 comments:

guy said...

Interesting that these stats are more useful to the providers (surgeons) than the consumers (patients) in this 'market'.

And yet compare this with the wreckage left by the misuse of stats in the education system (distorted league tables, abandonned SATS tests, school admissions fraud trials etc. etc). What's the difference here? From the article it looks like the surgeons' initial scepticism was probably influenced by the corrosive effect such stats have had on education. But perhaps the 'market pressure' argument is so bogus in this case that there really isn't any patient 'choice' --- if you're desperate for a heart op are you really going to shop around? Patients (like parents) don't want choice, they want reliable quality of service.

Medical stats like these are pretty black and white (how many treated, in what condition, how many survived for how long). Educational stats are pure grey.

glyn moody said...

Interesing points. I also can't help feeling that people actually try a tiny bit harder when they know that someone is watching....

EJ said...

The past few years I had direct experience in the Netherlands with the mess created by trying to "measure" educational data (even something that seems as simple as attendance...). I think that the positive results remarked upon here are a kind of shock effect that actually does not last very long. The moment money is attached to the numbers, hospital administrators (and not, in most cases, the medical staff) will find ways to manipulate the numbers. There is an endless variety of ways you can manipulate numbers, even numbers that seems as clearcut as "deathrate". Ask any epidemiologist. They know al about numbers in medical settings. The predictable course of events after the discovery of the manipulation of numbers isthat more numbers will be requested, to check on those collecting the numbers, or to check on the numbers themselves. What follows is an endless arms race that soon provides only work for administrative staff, an no benefit whatsoever for the treatment of patients.
The problem is that for the umpteenth time we are trying to reduce quality to quantity. By doing that quality looses much of its "quality". And there is no easy way out. I you ask a patient about his doctor, he will most likely, knowing nothing about medical things and in no way capable of giving a sensible opinion on the quality of that work, just tell you whether the staff were nice, talked to him and explained what they we going to do. Probably hospital staff knows about the quality of their colleagues. But will they tell outsiders? And all this is compounded by the fact that "outcome" in medical treatment is mostly a subjective measure, given the fact that most medical conditions are not like broken legs that heals, or not. They get a little better, or a lot better, or get better for a while, etc. What will the measured outcome be?
I personally think that numbers can be helpful, but maybe we will have to come up with some kind of "ethical numbers". meaning numbers that are not directly connected to money but reflect the commitment of the hospital to do as much "good" as possible, and cause the minimum of "harm" possible. It should also reflect the fact that doctors and other medical staff are not perfect and make mistakes. We should take these out of the hands of the courts (maybe with the exception of extreme cases of neglect) and find a way to deal with medical errors publicly without crucifying the people involved. A system like this would probably satisfy most patients too. Many patients who were affected by medical error just want to hear "sorry", and be listened to. That does not happen now because of the fear of legal action.
So transparency in numbers might be helpful, I think, but not as a measure coupled to money. transparency should be IMHO be the core of an ethical reputation system.

glyn moody said...

@EJ: interesting points. I'm sure you're right about the corrosive influence of money here, and I hope they don't bring it in. Perhaps I'm more of an optimist about using reputation as an incentive.