Electronic records ‘a first step’ to knowledge-based care

Electronic records are just the first step in harnessing information and communication technology to improve clinical care – the real prize will come when clinicians can be provided with knowledge and the information tools to support their decisions.

The huge challenge and potential gains of connecting clinical records with knowledge bases, was the theme taken up by John Chelsom, chief executive of CSW Health, at the recent XML [eXtensible Markup Language] Summer School in Oxford.

He told the audience that introducing knowledge management and, ultimately, clinical decision support systems will be an order of magnitude more complicated than current efforts to introduce electronic patient records systems, and require the use of new data technologies. To deliver benefits knowledge bases will need to be coded to a far greater extent than those currently available.

Chelsom argued that only now is the NHS beginning to reach a position where it will be possible to provide clinicians with really useful knowledge resources and decision support to clinicians. "Only when we have in place the NHS Care Records System and web-based access will we have the components for reasoning-based systems."

The CSW boss said that using knowledge management and clinical decision support tools would soon no longer be an optional extra. He cited the experience in the US where he said clinicians have already been sued for not using an available knowledge management system.

Better integration and use of clinical knowledge is vital if the quality and efficiency of healthcare services are to be improved, said Chelsom: "Currently some 10% of hospital in patient admissions result in some level of adverse event".

While current efforts in England are focusing on the development of NHS CRS, which will be a 10 to 15 year programme, he said the overarching aim must also be to connect records to knowledge bases. This was, he acknowledged, a tough proposition, "an order of magnitude more complex than electronic records".

He said that practical difficulties faced include developing electronic records which follow complex integrated patient care pathways, and determining where on each pathway you actually incorporate knowledge.

Chelsom added that while computer scientists often sought to provide knowledge support with diagnosis, “in reality diagnosis tends to be an area where the least help is generally needed". He said that the part of the integrated care pathway that would benefit most from better knowledge support was prescribing.

The CSW boss also stressed the need for there to be more uniformity and systematic use of common data standards – specifically XML – among the diverse range of knowledge bases. "There is also considerable overlap and differences of opinion in the available knowledge bases, which is a problem when incorporating them into reasoning systems."

The basic components of a knowledge system were electronic records and knowledge base, he said. "A reasoning system can start off as not very sophisticated, even a search engine would get us a very long way from where we are now."

But to deliver widespread benefits in clinical care, knowledge was needed as well: "Reasoning engines are quite easy to build, but they don’t work unless they have a body of knowledge to work with. At the moment [we’ve] got a situation where most of the main knowledge sources are not yet in XML or clinically coded and that’s a major task ahead."

The value of XML comes when you place an XML-based electronic care record in the same environment as your XML-based knowledge base, Chelsom argued. "You can then extract the context from both the clinical knowledge and the session, the glue that makes that happen is clinical coding of the knowledge, of the clinical pathway and the patient record." On top of this you can then start to add in roles and business processes.

Because the National Programme for IT had XML as a core standard and was providing all applications and services through a web-browser with a single log-on it had great potential for applying knowledge and clinical decision support drawing on the care records themselves:

"In 10 years’ time we will be able to build quite sophisticated reasoning engines as knowledge begins to build up in CRS, and after that we will be able to develop real clinical decision support systems."

Developing such systems though will be a long-term objective, on which work must begin now, concluded Chelsom. "Are we trying to run when we haven’t even stood up yet?"