An example emerges

Just after I'd sent out the last chapter to the people in the virtual cafe, I had a real life meeting with one of the readers, Dan Winchester. Dan lives just a few miles from me and we often meet for a beer and a chat in a local pub. Dan's specialty is databases and the design of backend systems that provide Web sites with their organisation.

Dan was somewhat bemused by the strategy I was suggesting. He didn't see the need for bots or personal agents as he felt everything that needed to be done could be handled from the server side. He considered it would be far less complicated if clients simply complete a form on a Web page and leave everything else up to the Web site – which could use a suitable suite of programs, coupled to a database, to provide all the functions the clients would need.

When I explained how it might be preferable to put the programming on the client side if the client's help were needed in obtaining the solution they wanted, he immediately raised the objection that this was different from the way everyone else was using the Web. He was of the opinion that because the approach would involve having to teach clients to do something new, there would be a resistance to its use. As he explained," People don't want to learn anything new. They like intuitive applications that are simple to use and to understand. They want everything done for them".

Although this is a valid argument, it applies only where it is possible for computer programs and algorithms to satisfactorily perform the tasks that are needed by a client. For many client needs, particularly those involving highly volatile data, it just isn't economically possible to create a server side solution because too much human input is required. This makes it imperative that the solution is handled on the client side, where a human is available to handle the parts of a solution that are not amenable to any algorithmic approach.

Dan Winchester wasn't convinced. He couldn't imagine a situation where clients would need to be an integral part of a solution – and, at the time, I had no specific example to give him. However, the very next week, an ideal example did arise: when I was invited along to a special meeting in the UK Houses of Parliament.

Here was an instance of developments that seemingly arise out of the blue when pursuing an opportunity. Prior to starting work on this book, I'd never had any contact with the medical business. Yet, after the question about a cancer database coming up in the virtual cafe, I'd found myself discussing with several people in the oncology division of a major drug company the possibility of applying the people space concept to help cancer patients find cancer treatment trials. These contacts saw a great value in this approach, not only from the point of view of the patients but also from that of the drug companies, who were having to spend quite considerable sums of money trying to find suitable patients for their treatment trials. These communications with the drug company resulted in me getting an invitation to the meeting in the UK Houses of Parliament.

The meeting was called for 5 PM, Tuesday 27th February 2001, in committee room 21 in the House of Commons. It was entitled "Developments in Information Technology and Cancer Care" and the subject matter to be discussed was "How developments in information technology will affect the quality of care received by cancer patients". In the chair was the Member for Parliament: Dr. Ian Gibson MP.

These kind of meetings are regularly used as one of the means by which the British Government obtains information about the various areas of specialty the Government has to deal with. At this particular time, there had been much press comment on reports that the standard of care and treatment for cancer in Britain had fallen below that of some of the other countries in the European Union and the Government had declared their intention to catch up within five years.

There were three main speakers who had been invited to the meeting to describe the approaches they were taking in using information technology to help improve the quality of cancer patient care.

One of these speakers described how his department was in the process of creating a registry of all cancer patients in the UK. They had recorded the details of some 8 million patients already and were attempting to add to this registry the 250,000 new cases that occurred every year. The speaker produced some graphs to illustrate how they'd used their database to deduce that patients who were treated by more experienced physicians did rather better than patients who were treated by less experienced physicians. Not exactly a startling revelation to come out of such a gargantuan amount of data.

Another speaker, who was a project manager for the National Health Service Information Authority, described how they were setting up a database to record the treatment of every cancer patient in the country. An outline was given of the many practical difficulties that were involved in getting this information into a database.

The most ambitious scheme was proposed by the third speaker, who was attempting to set up a Web site to advise cancer patients of the various treatment options that were available. It was explained that many patients wanted to be informed enough to play a part in the decision making concerning their own treatment and needed a reliable source of up to date information.

The literature handed out by this speaker to the attendees of the meeting explained there were 20,000 medical journals in various languages that between them publish 2 millions papers every year. These papers were said to contain details of many clinical trials but some were too small to give clear answers and most reviews of the evidence presented in the papers tended to be highly biased. Additionally, it was pointed out that a search for "cancer" on the Internet search engine GOOGLE.COM produced 5,960,000 references. The speaker told the meeting that this amount of information would be confusing and too ambiguous for most patients to be able to make any use of.

The speaker explained that there were multiple decision points in a patient's journey. Questions needed to be driven by patients as well as their carers. There would be a requirement for various levels of information to allow shared decision making between patients and their physicians. This would create a need for unbiased, systematic reviews of all cancer treatments, which would have to be continuously kept up to date. The suggestion was for there to be "kite marks" given to papers, to provide indications of their authenticity.

All through these presentations there were murmurs of support and approval for this work, with many of the attendees making copious notes. Nobody seemed to be aware of the practical limitations of these proposals. It seemed to be assumed that the mysterious power of the computer could overcome the horrendous problems associated with dealing in enormous numbers of people and vast quantities of volatile data.

After the presentations there was a question and answer session, where I was able to ask the speakers about the practical realities. I asked about the costs of collecting and categorising all this information. It may be cheap to manipulate once it is in a database, but, didn't every single piece of information have to originate from a human entry? I asked if the proposers of these schemes were taking into account that the nature of most patient's condition was continuously changing as their disease progressed or went into remission as they were put onto different treatment schedules? Could this highly volatile information be kept up to date on millions of different patient files?

I asked about the practicality of any organisation being able to judge the contents of 20,000 journals; differentiating between the merits of 2,000,000 articles every year: removing bias and selecting those to kite mark. There was a stunned silence as the reality of the situation dawned upon everyone at the meeting. They weren't dealing with a finite amount of stable data. All this data was highly volatile and even if it could be put into a database it was likely to be out of date by the time it came to be used. This would seriously compromise the value of any conclusions that could be drawn from any analysis.

I pointed out that what they were proposing as new British initiatives in the fight against cancer, appeared to be new initiatives only because there was little evidence that any of these proposals were already in existence. However, it was extremely unlikely that nobody had thought of these ideas before. And, if they had been proposed, surely the funding would have been available if the proposals had been viable? Certainly these ideas would have been an attractive proposition for Venture Capitalists if they were practical possibilities. The fact that these proposals weren't already in existence would seem to indicate that the cost and the practical problems involved had ruled them out as realistic approaches to improving the quality of patient care.

The inescapable conclusion must be that an organised, server side approach cannot provide a solution. The only possible way for this information to be organised and made useful would be if it were possible to engage the coordinated help of millions of people, sharing the work such that each could look after just one small section of the database to keep it up to date and make sure that it interacts appropriately with all the other information in the database.

Here was a real world example to explain to Dan Winchester the need for a client side solution. If every patient handled their own data and if each could individually control the database to get results specific to them, then it would be possible to deal with this vast amount of volatile information and use it to make a difference to patient care and treatment.

The trick would be to get such a system into operation. Clearly, trying to organise such huge numbers of people would be an administrative nightmare. The only possibility of success is to create a self organising system of self motivated people, where people take part because it is in their best interests to do so.