I while back I attended a very informative talk by Ken Harvey of Latrobe University. It was about prescribing software for medical practitioners.
Your doctor probably has a computer on their desk by now the Federal Government gave grants of $10,000 in 1999 to GP practices to assist them to computerise. When you get a prescription you might have noticed that there are ads on the software that your doctor uses. Do you think that is a good idea?
Medical Director, the prescription package that became dominant, takes strategically placed advertisements. The screen into which the doctor puts your various details systolic/diastolic blood pressure, age, cigarettes a day, total cholesterol etc then ranks your cardiac risk by percentile in your age and sex cohort. Wonderful stuff! Think of how fantastically such expert systems could improve healthcare.
As part of the service, the doctor gets an advertisement for Lipitor (and so do you if you’re looking on). When Ken Harvey gave the speech he claimed that the chemical name of the drug is barely legible (in breach of the relevant industry code of industry self-regulation) but you do get to read the slogan “Power you can trust”. The Pfizer website informs you that it’s the # 1 cholesterol-lowering drug in America.
About 50% of doctors report that they are uneasy about this but around 80% of them have the software from which the ads beam out.
The debate will, predictably degenerate into an exercise of moralising. Pharmaceutical companies bad. Regulation good. It doesn’t surprise me that pharmaceutical companies do this and to the extent that the rules permit what they do, they’re behaving consistently with their obligations to their shareholders. Nor does it surprise me that the activists who don’t like what is going on speak moralistically that’s the way to get on the tele. But as a policy question, my guess is that these ads are not appropriate and the companies should come up with a better code of conduct (and also stick to it).
In fact, the companies have responded a little. Section 3.9 of their industry code (pending approval with the ACCC as I type this) forbids advertising on the kinds of screens that I’ve just mentioned, ones that one might expect to be turned towards the patient. So that’s a good thing I guess. But it seems to me that advertising of any kind in a system for selecting pharmaceuticals is neither ethically too flash, nor economically efficient.
But I think we lose a lot if the debate just runs according to this fairly standard ‘business versus consumers’ script. What a fantastic resource we might have if such software were to become as good decision support software as could humanly be managed. Then it would help steer doctors towards the best evidence-based practice possible at any given time.
Such software could help them avoid the dangers we know of (existing software is good at a lot of this already pointing out known drug incompatibilities and alerting doctors to patients’ allergies for instance). A bit of a squiz at Medical Director indicates that they’re also building the software’s connections to important medical advisory texts.
The software is already very useful in generating data for pharmaceutical companies to use all de-identified and so not infringing people’s privacy. And its vendors are building it as you’d expect into more than software prescription software. It’s being developed to send orders for prescriptions to pharmacies, pathologists and so on.
So perhaps it’s all good, or mostly good. But I’ve got a few concerns. Firstly as the ads demonstrate, the software vendors have conflicts of interest. The ads prompt doctors with new information but in a very biased way. Is there a better way?
Secondly, as I understand it (though I am not sure) Medical Director ¢â¢ behaves as dominant software suppliers tend to, which is to say providing competitor platforms with interoperability is not high on their list of priorities. Yet much of what is being built is personal information about patients. Shouldn’t they have some right to interoperability? And wouldn’t it be a lot more efficient?
Standard competitive approaches get us a fair way towards a better world providing one accepts that:
1. the kind of advertising I’ve discussed should be sharply curtailed or prohibited and
2. interoperability is mandated with access regimes requiring release of appropriate source code. (I presume this is not hugely technically problematic but don’t know enough to be sure. I don’t know if the dominant software provider actively obstructs interoperability it sees as a competitive threat as for instance Microsoft does. All my suspicions tell me they would but I don’t know.)
This would surely improve things. At least it would get the software evolving to meet the needs of doctors which would be a good start. But I think there are other stakeholders who should be represented actually I mean involved not represented (if it can be done effectively).
I suspect that open source software gives us some clues about how to proceed. I don’t mean that prescribing software should necessarily be mandated as open source (though it would probably be sensible for hospitals to build open source prescribing software as part of larger software architectures that could migrate from there as open source software and compete with proprietary software.)
The real issue is how to enrich feedback loops. Who knows what mature medical software could do? Some ideas:
1. Data feedback could be ‘live’ so that information emerged very quickly. This would be helpful regarding problems with drugs, with the software with appropriate warnings and bug fixes resulting. It could even give us very accurate information about the state and spread of various diseases through the community.
2. Doctors could be put in touch with doctors treating similar cases and with relevant experts in Australia or why stop there the world.
3. When patients were diagnosed with something, they could be provided with a print-out or e-mail of reputable web-addresses to go to find out more about their condition, and be invited to join user groups of drugs, and again, invited to provide feedback on those drugs efficacy, side effects, means of dealing with the side effects and so on.
I’ve only started here. Its easy to dream this stuff up, but if you know a bit about open source software you know that amazing things are possible and further that where interactivity, morale, high skill and specialisation are important there are circumstances where voluntarism is not just much cheaper than proprietary approaches, but technically superior perhaps vastly superior over time.
I expect that, though it’s possible that the state could play a useful funding role (then again funding can sometimes mess things up and the state hasn’t done much deliberate funding of open source software) the most exciting things would not come from governments but from rich interactions between profit-driven providers (pharmaceutical companies, software vendors or open source providers selling services to maintain open source software), medical professionals both publicly spirited and less so, patients and other participants each collaborating.
Finally, if we think of this software and the knowledge embedded in it, it partakes very considerably of the nature of a public good which is one of the reasons that open source can so often out-compete proprietary software without any subsidy at all.
Ironic isn’t it? Just as economic reformers began to take private competition as paradigmatic of efficient economic activity (and in most of the moves they made they were pretty much on the money in doing so) along come a whole lot of new areas where the characteristics of public goods predominate and the market (or a proprietarily driven, and profit-driven market) is not efficient. This is true of all those info-goods and services that can be replicated for nothing on our computers and the internet but also in the biosphere with its massive public bads both of the new four-letter acronymed variety AIDS, SARS, all the bird flus and more traditional bugs with new drug resistance.
There’s probably a case for some increases in public funding of these things. But it can be over-rated. And what’s more interesting and I think more productive is trying to think of new ways of facilitating the natural evolution of ways of seizing the opportunities that new technologies are bringing our way.
I’m always a bit puzzled by this theory that you select your best and your brightest, via a highly competitive process and put them through several years of med school and post-grad training only to have them quail into mindless, compliant, script-writing zombies the first time some pearly-toothed pharma product manager hoves into view.
Surely, we can anticipate that most of them will have sufficient intellectual rigor and focussed expertise to be guided by clinical and prescribing guidelines, primarily, rather than an ad flickering on their computer screens.
It’s also puzzling that Australians respond to direct advertising by purchasing billions of dollars worth of ‘complementary medicine’ annually – most of which is ineffective or unnecesssary – and no-one bats an eyelid.
Puzzling indeed Geoff.
And the dollars being invested each year in the ads (which rely on the absurd theory that those who are, as you say, selected competitively and trained within an inch or their lives, might be influenced by such ads) just deepens the mystery!
Nicholas, a really nice application of your “open source” line of thinking. A meme perhaps?
I’ve passed the post on to a few doctor friends
Well I think the ads are very definitely about promoting new product and pharmaceutical companies have very large marketing budgets. If some clinician sees the ad and checks the product info out, looks at the trial data, examines conference presentations, talks about at journal club, etc, the company might engender interest. If they get enough advertorial happening in MSM – obscure phase II trial data is always good for this – they might also engender interest. The point is that relatively few Docs will simply respond like Pavlov’s Dog and the companies are aware of this.
I think what’s important here is the relationship between doctors’ private incentives and society’s. For me, the existence of advertising says something about Doctor’s less than optimal uptake of new technology. The take up was too slow under conventional pricing, so the developers had to look for alternative means to lower the price (and get doctors to use it). If pharmacutical companies want to come to the party then all well and good. But I agree it becomes complicated by the potential for negative flow on effects if doctors are too easily swayed (which I agree is pretty doubtful).
While I agree open sourcing the application could work, perhaps one more easier solution would be just to fix the source of funding that is used to get doctors to adopt and grow platform. I’m not really sure why the pharmacutical companies need to be involved at all – if it is a public good, why not mandate a fixed fee (like a tax) to come from patients each time?
Nick, an enormous amount of work has already been done on developing a new health system, and there’s more on the way.
http://www.healthconnect.gov.au/whats_new.htm
There are many complicated issues, including differences between state health legislation, secure verification of people and professional jealousies between doctors and pharmacists.
The work seems good, and aims to create a marketplace for software developers to compete. This will almost certainly provide software alternatives that don’t bother users with advertising. [1]
Your concerns about data and interoperability are without foundation. Data and important functionality are defined used public (i.e. open) schemas and APIs, so that any conforming application can work with them. [2]
Note that open source has nothing to do with this. Open standards are competely separate from open source.
1. Medical Director’s use of advertising suggests some sort of market failure. Other market sectors favour versions free of advertising.
2. Medical Director seems to be reasonably interoperable as is. It imports many formats and exports data as XML files, which theoretically could be read by other applications.
Nicholas;
I too have mused in this direction, apparently also unaware of the HealthConnect project. Still, I think that there should be some emphasis on disconnecting the storage of reported symptoms from diagnoses, so that remote expert systems can perform data mining, oversight, etc etc to supplement the current approach.
http://chester.id.au/posts/8
There are considerable numbers of people interested in and already contributing to Open Source Health Software, particualrly Electronic Health Records (EHR). Open Sourc esoftwareis important in this area for reasons of interoperability, future proofing and not the least costs. Developing economies need almost exactly the same health care and software that we need but are clearly hampered by costs. FLOSS goes a considerable way toward erasing much of this cost.
A program for GPs or patient health record could basically be the same the world over and with FLOSS the core development / programming costs do not have to be paid for again and again and again with each new nation or in Australias case each new state that comes online.
The big leading health care countries like us have shown little interest in FLOSS for health due mainly to governments lack of foresight and an almost wilful ignorance of Free and Open Source Software.
I’ll look up some of the links to OSS in Health and dump them here if you are interested Nic?
Oh yeahand Geoff I’ve never understood how society gets upset about GPs small foibles when we have a protected industry, local Pharmacies, all a vital part fof our health sytem, openly flogging, iridology, aromatherapy, bach flower remedies, weight loss scams, anti aging snake oil and others such as chiropractors giving x rays and reiki. bah.