I appreciate your concern for technocratic management devices that can undermine doctors’ professional ethics. I think the Gruen Tender would actually re-inforce them and so strengthen them. If we do a good job we all like recognition for it. And it even makes us try harder if you’ll allow that.

If you’re wondering how it would work with emergency jobs and ambulances - the simplest answer is that it wouldn’t. (It would be interesting to ask if it could conceivably be used in tendering out ambulance services to various regions) But think about it with procedures that can be anticipated - I used the idea of an obstetric delivery. The doctors remain in charge. All they’re asked to do is to estimate their own error rate (and the errors would be measured by someone other than the doctor - and run past the patient). From this information we can work out which hospitals are doing the best job on what kinds of patients - they can swap notes. A lot of the improvement in medicine comes from people excelling and others trying to find out (and then imitate) how.

I love statistics. and can see what you are driving at. But the best
people to handle the morass of the human mind that drives medicine,
is still doctors, fallible and driven as they are. I have been watching
them for years and I would not have health decisions farmed out to
anyonelse

I don’t think you should write off the possibility that Gruen Tenders could emerge competitively. I agree they’re unlikely to cover the market in this instance. But a good service provider has an interest in using it to demonstrate their superiority - or even just as a gimmick, a ‘point of difference’ as they say.

Here is an stock tipper advertising an audited record of its performance.

http://www.fatprophets.com.au/content.aspx?page=Performance

Rex,

I’d sure like a GT to help me decide between real estate agents. I wouldn’t find it confusing at all. You just read the last column in the examples and pick the highest number!

In health I’d use it too if offered it. But the evidence suggests that at least at present, most consumers don’t do this kind of thing when it comes to health. But I think the GT has more going for it for health administrators and funders initially at least. They’re dolling out millions of dollars and where this can be effective - for instance where I’ve suggested it would be most effective, it could help them make allocational and funding decisions to a considerable degree.

The logic is faultless, but I think it is almost completely impractical.

Think of all the abundant information about such things as Mobile Phone Plans, insurance, superannuation, motor cars.. the list is endless, but does it help us decide?

For a selecet few maybe, but the vast majority take the default option, or the one their friend suggested, or just go with ‘the vibe’.

For your suggestion to have any value it must be very widely implemented in a field to give it a valid sample size, it must be independently and centrally collated to maintain levels of trust, and this must be funded from the process it seeks to measure.

It would be an expensive administrative burden for the benefit of only a few dedicated people. Most people will simply say its to hard to use.

All good points which probe the limitations of the mechanism. I agree that with low levels of undesired incidents it can take a while for observations to become statistically significant. But there are responses to some of your concerns.

For instance, why not aggregate ‘optimism factors’ over a range of procedures. One could aggregate it over a whole hospital, event though it was doing lots of different procedures. One could also produce more team and procedure specific optimism factors, but not use them until they became more statistically robust.

I think application of this idea to health care would be more difficult (but potentially more rewarding) than other areas. The reason is that health outcomes are multi-dimensional and often binary. For instance, did the patient die, yes or no. Were they cured, yes or no. Was the condition improved, yes or no. The fact that there are several possible success measures complicates both data gathering and interpretation. More importantly, yes-no data carries rather little information compared to numeric data. You would need many thousands of records to be able to reliably estimate a possibly time-varying optimism factor for say probability of patient survival. Any if you were trying to do this separately for each medical department and type of ailment you might never get enough data.

By way of contrast, in the real estate market there is only one obvious success measure from the seller’s point of view - highest bid price during the selling period. One difficulty I see with introducing such a system is that there is no incentive for the first real estate agent to join. Their optimism adjusted quotes will not be compared with any other agent, and they may well be compared by the market to some unattainable ideal. On the other hand, once most agents were part of such a system there would be great pressure on the remaining few to also join. I guess that economists have a name for this kind of industry, where there is a stable equilibrium with zero participation and also full participation. Anyway, the practical upshot would be that such a system must be imposed on the industry - at least initially. The RE industry already collects plenty of data on sales (but not on quotes) and they sell it, so there would be plenty of opposition to regulators stepping in to make such data free.

Casting the net much wider, I reckon there are more glaring examples of information problems that few economists seem to care about. In order of importance, the first is the difficulty in finding price information. I am thinking about a product such as a new car. While it is true that you can ˜shop around” it is very difficult to get a feel for what a reasonable price is. The dealers are completely inscrutable and you never know how many more thousand they might knock off. Most people will ask there friends and have a general idea that 10% below the list price is about par. But wouldn’t it be great if there was a website listing every new car sale in Melbourne over the last year?

The second is the lack of information about satisfaction levels of customers. I recently chose a builder for a mid-life-crisis-house-reno and was unable to find any reliable information on any of them and had to go on gut instinct. There are a few websites that allow people to post gripes but it is severely limited by libel laws.

And the third is the most important. Disinformation by which I mean advertising. This kind of links up with the second. There is a huge incentive for companies to lie to you about how good their product is, but who is motivated to publicise the products that let us down? We do indeed all build in our own optimism factor into the claims of ads but the images are still effective else there would be no industry.

If industry leaders were to take this up as a market signal, you could make it an industry standard. An independent firm could sell the service of logging and assessing the variability of firms’ tenders. Whether or not the benefits of doing this are greater than the costs depends upon the industry you’re looking at (asymmetry of information etc.etc.)