Complexity of patient records poses huge challenge for data integration programme

Slow progress in computerising patient records has been highlighted by a BCS debate, which found that IT specialists often underestimate the complexity of the task.

Slow progress in computerising patient records has been highlighted by a BCS debate, which found that IT specialists often underestimate the complexity of the task.

The challenge is that patient records are far removed from bank statements or business records, said the report on the debate, which brought together 45 health and health IT specialists under a rule of anonymity.

"Patient records combine data, opinion and observations from a wide range of medical professionals and for a wide range of purposes," the report said.

"Information that is hugely important in one context, such as blood pressure readings taken after heart surgery, may be of little significance to a GP during an appointment weeks later, whereas a single routine reading can be highly significant years later in the event of unexplained heart failure.

"In addition, although basic data such as blood pressure readings is precise and easily categorised it might be linked to non-unique opinions such as 'it is low' or 'it is due to inflammatory bowel disease'.

"Indeed, medical records can contain seemingly contradictory statements, but there might be no contradiction if the statements are by different observers at different times."

Such interpretations and opinions cannot be recorded in a structured form, the report said, and this creates huge complexity, which is rarely appreciated.

"Systems specialists new to the field fail to spot the complexity, because health care is unlike any other record system they have encountered. If they do glimpse it they attribute it to doctors being unable to agree on anything."

There have been two typical responses, both leading to failure, the report said. "The first shies away from the complexity and asserts that everything will come together through 'enterprise interoperability'. This means the complexity is pushed on to the design of the interoperability, which never appears.

"The second denies the complexity, which gets pushed on to the requirements, specification and design of the 'solution', and the whole thing crashes.

"There sometimes seems to be a blind faith in health care that top-down IT solutions, addressing untested requirements and loosely assembled specifications, can cure problems without the growing pains associated with such disruptive technology and without commensurate and complementary focus on bottom-up investment in know-how, disciplined experimentation, capacity development and organisational change."

All this adds extra challenges to the usual IT project stages of requirements, specifications and implementation, and raises questions for the NHS Connecting for Health programme.

"We will wait a long time if we seek perfection, but what we do must be good enough," the debate concluded. "But do we have a concept yet, and are we prepared to accept 'good enough for now'?

"Do we need health targets for Connecting for Health: for example, aiming to cure 75% of people by implementing IT, and if not, why are we pursuing the programme? Is IT to be just a substitute for paper or do we expect it to bring something more? Are we entirely clear about what the programme is intended to support?"

Full report:

Read more on IT architecture