Cross Talk on observational medicine

Mike Cross reflects on how cheap and portable devices are enabling public service workers to capture useful, real-time data in some unlikely places.

On the streets of Westminster, traffic wardens capture digital photographs of illegally parked cars. In the medical assessment unit of Portsmouth Hospitals trust, nurses capture patients' vital signs electronically as they go about their ward rounds.

Those are just two examples of public service workers collecting real-time evidence of what's happening on the front line thanks to cheap, commercial, portable IT.

The evidence base created on the streets of Westminster persuades more motorists not to challenge their parking tickets. The evidence base created at Portsmouth (and a handful of other NHS sites) will have more profound consequences.

"Since the days of Florence Nightingale, more or less, our main tool for doing this has been the observations chart at the end of the bed."

The observation chart

It's more than a millennium since the Persian physician and philosopher Muhammad ibn Zakariya al-Razi asserted that truth in medicine must be underpinned by observation. "We shall not accept any property as authentic unless it has been examined and tried," he wrote.

It's also more than 100 years since the German physician Carl Wunderlich realised the importance of recording a series of observations to plot the progress of a disease.

The theory may be well established, but we're not very good at doing it. That's the opinion of Professor Gary Smith, a consultant in intensive care medicine at Queen Alexandra hospital, Portsmouth.

He's dedicated much of his career to trying to improve our ability to predict when patients are going to get very sick. Since the days of Florence Nightingale, more or less, our main tool for doing this has been the observations chart at the end of the bed.

But even if it is conscientiously filled in, what use does this information serve? Professor Smith is only half joking when he says that the only people who read the chart are visiting relatives.

"One result is a leap of accuracy in the data collected. In one study, the number of incorrect items recorded in early warning systems dropped by two thirds."

The chart updated for the 21st century

One use for the information is to plot an "early warning score" to alert doctors that something is going wrong. Such scores are usually calculated by adding up the number of deviations from five or six normal parameters.

But what is normal? And how do we know that adding gives the best result? We don't - largely because of the difficulty of collecting and disseminating accurate information during the course of hospital treatment.

Enter the personal digital assistant (PDA). At Portsmouth, in a brilliant example of an IT project driven by clinical need, nurses have been issued with standard PDAs loaded with graphic intuitive software which enables them to record vital signs while going about their normal rounds and maintaining contact with the patients.

Observation recordings from the device are uploaded to a hospital-wide database via the Wi-fi network (with backups of the past 24 hours' readings kept on the device in case the main network goes down). Ninety eight per cent of observations are on the database within two seconds.

The database is linked in real-time to the patient administration system and the pathology laboratory database.

One result is a leap of accuracy in the data collected. In one study, the number of incorrect items recorded in early warning systems dropped by two thirds. Yet more evidence to support the theory that the accuracy of healthcare data increases in direct proportion to its value to the person collecting it.

Better data, better care

At Portsmouth, nurses in the medical assessment unit have been recording patients' vital signs electronically for some 18 months, now. One outcome is the creation of an accurate, real-time evidence base of what actually takes place during the course of a clinical intervention.

Professor Smith says this is big news. "We now have the sort of data that's not been available before to link the severity of illness with the outcomes. That might allow us to improve the outcome; it might tell us that whatever we do is not going to change the outcome."

As al-Razi knew, the consequences of evidence-based medicine can be discomfiting. Even more discomfiting than receiving a photograph of your car, parked on a double yellow line.

About the author: Michael Cross is a freelance journalist specialising in healthcare informatics and e-government. He is a member of the British Computer Society.