Shared Clinical Records Fallacies and its waste of time
I am clearing out paperwork at our old office. This pile is some of the work I did for NPIT and then CfH, Connecting for Health, the UK health service plan for a single NHS record. It cost billions. It failed.
Although I got reimbursed for some of my time on committees this was such a waste of time and I knew it. I was not popular as never liked the concept of a single health records anyway, and detested the idea that Government should write software. The Government’s agenda and civil service procurement contracts were so precise that it was as if they were writing code.
The main effect, indeed aim at the time, was to kill off multiple medical UK IT businesses. Only four primary care system survived. NHS bureaucracy believed that it would be so much better if there was only one system, one supplier in the NHS. The next best option, as far as they were concerned, was that systems should all look and behave in the same way. The result is that the fast and innovative clinical systems we had have become stuck in aspic. The drive to innovate, to compete was killed off by the pressure and time needed for government conformance. Little attention paid to what doctors, nurses or for that matter what a patient might need. The systems in use now are slower and much less ergonomic than the medical system in place in the late 1990s. Most UK GPs had systems on their desks by then because they saved time and created legible scripts and records. Now systems slow down consultations and act as a barrier between doctor and patient. The doctor has to glare at the screen point a mouse at a tiny icon.
Fallacies of a single record
1 “It saves so much money and time, and improves patient safety if there was a single system and NHS record” Not true. A single record lead to sloppy thinking and lazy doctors. The patient is a day older. The past record is just as likely to bias you in the wrong direction by what was written the past. There is likely to be something new. Take a history and examine the patient. It works every time! This is especially true in an emergency situation. Junior doctors should be forbidden to look at old records for 48 hours.
In the UK GPs can have access to records (paper for the most part) going back to 1927 as the records follow you around. When working in Australia there is no such thing, and lo! to my surprise I did not miss the records clinically. As a matter of prurience I might have. When I told and Australian patient of GP UK record access she said that would a be gross invasion of her privacy. She had a point.
2 “But we must know if a patient is allergic” . If only recording allergy was that simple. Most people who report that they are penicillin allergic are not. A patient may report that she felt upset on that green pill, so not unreasonably the GP puts on his record allergy to green pill, so now the system will almost block the GP annoying the patient when attempting to prescribe green and related hues in the future. Alas a hospital will interpret that to mean the patient will collapse with anaphalaxis to green pills. To avoid this the GP now needs to record green pill allergy in greater detail, more than the patient nor he needs in his practice. Detailing what was reaction, what type, intolerance of allergy, rash, urticaria, wheezing and so on and on. It may need a tree of many codes to describe the reaction and how serious. It is so tedious that it is left. But this coding dilemma occurs with recording of almost any disease. Heath Data has fractual properties, like the length of the coast of Britain, seemingly approaching infinity once you draw round each pebble or then grain of sand. The level of detail you need to record depends on your need in front of the patient, its use to you, the patient and locality.
3 That leads onto the provenance of data. Where is comes from has a huge effect on its meaning. A nurse may find a patient confused, so she might record dementia on her system (it might even be recorded as Alzheimer’s). For her work that all she needs to do to remind her. However this could be any different type or dementia, or even psychological, dementia-precox (psychosis). Who and when was the data entered changes its meaning, and then who can correct it should it turned out to be a toxic confusional state. Correcting data entered by others is a minefield in shared records. A GP may record heart failure, but a Cardiologist may need to know what type, or worse it has since resolved and is still there on the record.
4 “Clinical systems can improve safety by warning doctors of interactions of patients other disease and between drugs”. Unfortunately the fear of being sued by suppliers for failure to provide warnings, means that the most idiotic warnings pop up. Such as in treating blood pressure: “Adding X with Y will cause BP to fall”…. YES that is what I want! No intelligence in offering up these warnings which are so frequent, so minor, or that the risks are already fully understood beforehand, that doctors simply flash past all warnings as they are wasting time, in doing so increasing the risk a genuine high risk warning will be missed. This is true for current systems and would be worse if there was a central record as that has to be designed for the most junior nurse doctor or admin clerk.
Solution. Competing Systems: Healthcare Secure Search Engine.
Keep encouraging different systems. The can use common coding systems as they do now. They can work differently, have different interfaces functions and innovate.
When a patient is seen at a different place, such as at hospital, an authorised doctor should be able to do a search of all NHS records within date ranges or selected places, to create a new record. Preferably a day later. This NHS search engine would be secured with two factor authentication or more, with swipe card; the patient notified that such a search was made and by whom and when.
That way provenance of each item is clear. There would be no need to create a single record. There is the NHS spine record, but it is only a summary with few details, and keeping it in sync with GP records is not easy.
Connecting for Heath was doomed. Its premise was wrong, but those concepts and wish for a single patient record still come up. It still a dream of civil servants everywhere. It is not needed and it kills innovation in healthcare computing. Without NHS meddling we would now have slicker, faster, even tablet based clinical systems by now. More patients could be seen in less time.
PS if you invent such a search engine I will need 5% stake in it.