PRECARD 2.1: Bringing Epidemiological Data into the Clinic
One of the recent acquisitions of the Medical Museion is this copy of the PRECARD Version 2.1 risk assessment software. The software is used by general practitioners to calculate the risk of cardiovascular disease in the patients. Distributed freely in the form seen below, it was used by around 30% of all Danish GPs in 2003.
The software was developed by Troels Thomsen and colleagues at the Centre for Preventive Medicine at Glostrup University Hospital in Denmark, in collaboration with the Danish Heart Foundation. Initial funding was provided by Bristol-Meyers Squibb, who held exclusive distribution rights until 2002. The basis of the software were the results from large scale epidemiological studies conducted in Denmark, including the Østerbroundersøgelsen and the Glostrup Population Studies. PRECARD is the model for the HeartScore software, developed by the European Society of Cardiology, which aims to provide detailed risk assessment tools in the area of cardiovascular disease to European countries in general. At the moment, HeartScore is not able to perform as detailed risk evaluations as PRECARD since the Danish data for cardiovascular disease are more elaborate than elsewhere.
The software was collected by dr. Susanne Bauer as part of her research project “Mapping the visual and epistemic cultures of epidemiology: Individuals, populations, and the biomedical body” conducted at the Medical Museion as part of the “Recent Danish Biomedicine, 1955-2005″-project. In this context, PRECARD is an example of how information technology and visualization of epidemiological data are used in preventive medicine. Such tools may work to change the doctor-patient relationship towards joint decision-making based on probabilities, yet also enroll individuals into the logics of probabilistic health risk management.
06 Nov 2007 Søren
The issue of collecting software has vexed me because of the issue of what do you do with it after you have it? Do you keep obsolete computers to run it if someone wants to see it working in 50 years? Is that even possible, let alone desireable? At the medical museum, we go back and forth between accessioning this type of thing into the Archives as information vs. Historical Collections as an object.
Mike Rhode
speaking for himself and not the US Army
Yes - should software be classified as objects or as archive material? In our research&collecting project on recent biomedicine, we encounter such risk assessment tools, decision support software, digital anatomies, interactive visualisations etc. and are trying to integrate this digital part of biomedicine’s heritage.
On the longer term museological perspective, I think it will be crucial to also keep the hardware – just like many labs keep an old computer to run their favourite DOS programs.. Keeping the hardware could be coordinated with museums of technology that are active in collecting the history of computing. Precard was alsocollected with exhibition plans in mind. Here, I’d see software as historical objects rather than as archival information.
In view of the emerging field of software studies, couldn’t museums introduce an entirely new category for this?
I guess we could ask our colleagues in the history of Swedish information technology project — do you have any good idas on this, Isabelle?
This is definitely a very tricky problem. I am reluctant to say that since we have collected the PRECARD software, we should also maintain in our collections a suitable computer with a compatible operating system, and that we should do this for every piece of software that we collect. One reason is that computers take up space, but more importantly keeping them might give a false sense of security since computers might stop working if stored for long periods of time. I think Susanne’s suggestion that we should coordinate our activities with museums of technology that actively collect computers and have the skills to keep them in working order is the best way to proceed.
That said, we will probably find that we will collect quite a few computers in years to come. Much recent biomedical analytical instrumentation features a quite ordinary PC as part of their setup. When such instruments are offered to the Medical Museion, CD-roms with the original software is often missing, and a vital part of the instrument is therefore located on a harddrive. In these cases, my policy is to bring in the computer (stripped of screens and keyboards as long as these do not display any direct relation to the platform they were part of), and curate it as part of the general object. Even if CD-roms are available, I will tend to keep the computer since the software from a particular laboratory or department will usually be set up in a specific way that is worth preserving.
And that is exactly where the question of whether to see software as objects or archival material becomes crucial. As Susanne says, the PRECARD CD-rom is definitely an object. The material side of it is definitely part (though certainly not the most importatnt part) of why it is worth having in the collections of the Medical Museion. Yet any modification that a GP may have done to the setup is perhaps best suited for the archive. Still, I think Susanne is right in questioning the categories. After all, any clear distinction between objects and information is difficult to maintain, and interactive software highlights this.
Late answer to this. Although my colleagues from the Tekniska museet - who curate our project of history of IT (1950-1980) - know more than I do, I understand that the principle in our project has been to not actively collect software that cannot be operated, and instead focus on tangible and decipherable artefacts and devices. Especially, manuals or instruction books are preferred to software itself.
I forward the question to the head curator of the project and hope for continued discussions on that theme.
And here comes an answer from Tekniska museet in Sweden (Swedish National museum of Science and Technology).
This is indeed a problem. The museums have always collected human tools, but today those tools are often in the form of digital software. The computers are only what the software defines them to be. In that perspective it is indeed desirable to be able to collect software, but the problems connected with this practise seems to be too big.
The museums collect artefacts for the future, and that is not only a future some few years from now - we are talking distant future here. This is where the problem starts to become really tricky. Few museums have the resources to collect and maintain hardware operational to run all kinds of software from all times on all kinds of media plus migrating software on ageing media, such as CD-ROMs, floppy-discs and magnetic tapes. Imagine the kind of staff that a museum would need for that! Imagine the know-how they would need to have! Imagine the number of computers, I/O-devices and operating systems that would need to be maintained!
When we started the project to document and collect Swedish computer history, one of the first things we realised was that in the long run it would just not be possible for us to collect software and keep it available. Pretending otherwise would just not be honest! So, as Isabelle pointed out, we decided to collect software manuals and other descriptions.
Yes, software collection is a really problematic issue for science museums. A few years ago I was looking after an archive of computing history, and was confronted with the issue quite often. Here are a few thoughts:
First, it’s not new. One of the collections we took in to the archive was the documents of a large electrical engineering firm. Their archive method was state of the art for the 1950s: microfilms embedded in punch cards. It enabled tens of thousands of documents to be stored in a relatively small space. It was probably as big an archive collection relating to an electrical engineering company as you will find. But without a punch-card sorter the collection was, and is, inaccessible. The format of the information made it so.
Second, we are all familiar with the rapid change of information technologies, including data formating. I have problems accessing files I wrote 10 years ago, let alone further back. So we can all see the immense problems for future historians and today’s archivists.
There are two solutions. Either you have the resources and position of a national archive, in which case you have the clout to demand that people supply data and software in formats you can archive, and you have the resources to maintain it and make it available. This is the brute force approach and it can work. the UK National Archives are doing this. However, while most institutions, including most archives, do not have the resources or power. Here the second solution is the only way forward: rely on a bottom-up provision of time, tools, energy and interest. Old arcade games are kept because there is a community who is interested and willing to transfer them to virtual environments where they work. The same is true of programs for the first stored-program computers (see Martin Campbell-Kelly’s wonderful EDSAC simulation as an example). In general, archives have to give up being the containers of documents for the future, and start being the coordinators and prompters of bottom-up archiving. You’ll spot this, of course, as the Open model.
Finally, this is important. To keep the hardware but lose the software is the same as preserving the architecture but losing any knowledge of what went on in a building. What would we know of political history if we just had the buildings the parliaments met in but no record of the debates?
I fully agree that just the hardware won’t be sufficient: In the medical field it is the software that can document how ideas and visions of the body change, how medical decisions are made etc. or telling stories on how instruments depend on software. Re the practical challenge, see a new post: http://www.corporeality.net/museion/2007/12/04/curating-and-preserving-medical-software-inspiration-from-computer-history/