We're doing it again! This time we are biking from Petra in Jordan up to Nazareth in Israel, to raise money for the Nazareth Hospital Dialysis Unit. Last year we raised over £50,000 for the Children's Unit! Nazareth is the largest Arab town in Israel; the people are lovely, and the kids are awesome. Nazareth also treats kids in the West Bank of Palestine who have very limited access to healthcare. They need your help! Go to my sponsorship page to find out more and see what you can do! Maybe even join us..? http://justgiving.com/shanenaz

17 August 2016

How the Clinical Note has to change

How not to do it.
Everything is wrong about this image of clinical noting in the digital age. White coats are an infection hazard. He's not bare below the elbow, presumably in a clinical area. The device is a big clunky useless piece of crap that's a pain to type on using this stylus. Why is he typing anyway?

So here is the problem. When Electronic Health Records were first designed, their job was to replace and carefully emulate what we had been doing with paper records. The concept was that the paper record was something of a Gold Standard, and our objective was to leverage the benefits of electronic onto this gold standard.

However, as Larry Weed pointed out FORTY FIVE YEARS AGO, there is a deep problem with the way we structure clinical records, and I'll be honest here - in my journeys around hospitals and clinics, and even in my own note-taking (even? Good grief!), our notes are not fit for purpose.

Here's a typical example of a note on a baby on day 3 in the Neonatal Unit (I've made this up, but if anything it's better than most): "Thanks for referral; Hx noted. D3. b38/40 twin 1. Other twin OK. bwt 3.21kg. Meconium at delivery, req resus. Resp now OK; off vent. NG feeds. Dysm features: hypertelorism, small mouth, crumpled ears, clinodactyly V, abn palmar creases, hypospadias. Suggestive of genetic syndrome. DNA for array. Will RV. Pls get clinical images & skel surv."

Now in the context of a busy NICU that's going to get buried in loads more clinical notes very quickly. If do that in an electronic system it's going to take me longer to type in (I could have scribbled that in half the time it took me to type it, even with the abbrevs), BUT again it's going to be lost in the load of other observations, consultations and notes that get added in.

As if that wasn't bad enough, just look at it - there's little structure to it. A computer is going to have to be pretty smart to parse even that highly lucid (in my opinion) text into something it can analyse or search on. It's free text. Furthermore there's a load of duplication there - much of that info is recorded elsewhere, similarly in free text. Maybe I've just jotted it down to persuade some lawyer some day that I've actually read the record (not necessarily understood the clinical case - those are different concepts).

So let's say we ditch the paper (YES!) and go digital - how do we change our practice and train doctors, nurses and AHPs to bring the clinical note up to date?

It's not an impossible task. For one thing, if we can crack the login/ID problem it should become easy to see who has made a note, and when (yes, we're supposed to sign and date/time all notes, but that is often missed, or people miss their IDs eg GMC number). It should also become easy to contact that individual through the secure EHR system. So in that area the clinical note is a good "stamp" to focus at least some clinical care around.

We surely don't need to repeat the basics - that should all be in a summary box every time we open that patient's EHR. But who curates that? How do we turn the mass of data that we generate into a coherent story that outlines the scenario relating to that patient, and that all the professionals AND the patient/family can group around and agree? Moreover, how do we turn that summary into something we can perhaps share with tertiary or supra-regional professionals outside our local (or in Northern Ireland's case, we hope) regional EHR?

And (critically) how do we ensure that electronic notes actually bring benefit, not just to the patient, but to the staff using the system? We need it to free up time. We need it to be a pleasure to use. But I feel that our approach to noting has shackled us to the past in such a way that we have lost sight of the purpose and function of The Clinical Record.

In the digital era we need to actually enter things manually to a computer as LITTLE AS POSSIBLE. Voice recognition is still pretty damn basic, but it's making headway. But my clinical note should be short and to the point. If I need to enter something quantitative I should be able to do that, but I'm not writing a legal document here. I want to construct a digital record that is dynamic and positively contributes to good clinical management and outcomes for my patient.

So like clunky tabs and white coats, perhaps the clinical note itself needs a major overhaul before we start replicating in electronic form the mistakes that Larry Weed pointed out to us all those years ago, but nevertheless persist in our training and practice.


  1. It would probably help to know more about what you think would actually work in terms of cognitive process / workflow. For example, in the example are a bunch of abnormal items that would presumably go on some kind of curated 'problem list', maybe one dedicated to post-partum. Some obs are just normal or typical minor birth related events e.g. respirator / resus etc; probably part of a set of post-partum notes? Somewhere prior in the mix must have been the obstetric notes for the delivery itself. How should the EHR be broadly structured from the point of view of childbirth + post-partum (+ pre-natal as well)?

    Would it help if there was a visual guide (maybe like a mindmap visualisation) to quickly recording birth defects? E.g. if you clicked on 'physical abnormalities' you might very quickly get to 'clinodactyly' where you can then choose 'V' for the finger etc? Similar for hypospadias; clearly selecting some of these could then bring up appropriate forms for relevant further exams or even interventions.

  2. Hi Thomas, yes - a dynamic ontology browser is a great thing. For instance the HPO (human phenotype ontology) allows you to do exactly what you describe. Then the data can be queried and analysed, which opens up lots of avenues.

  3. Maybe we need to break down clinical notes into a 'What-does-this-contribute' hierarchy. So what's the aim of the note & it's subsections? From this we should only type/capture that which is needed for wider electronic record/sharing/archiving. Paper that covers prose regarding ward level care aspects probably has limited use beyond the immediate environment. Now I realise we can all think of examples of wider use of the latter but the question is- Are these examples common enough & important enough to invest in what is expensive and excessive electronic record keeping? So much time is spent documenting/duplicating the mundane. Changing that culture is as big a challenge as every thing else. Finally we should not assume that every system developed over decades in the NHS can be made better just because it's on a database and website. An arranged marriage is not the way!!

  4. Not an arranged marriage, Damian, but paper records as we have them at the moment pose considerable risk, and can't be reliably shared or analysed (without getting some poor sod to enter them into a computer). We need to find ways to capture and analyse data in natively digital ways, so we're not tied to the computer, but rather the digital properly helps us. Perhaps the most exciting prospect (and you guys know this!) is opening up the full record to the patients...

  5. Voice transcription and direct messaging prior to storage in ECR ?
    Some work to do to staandardise a) between specialties b) between primary and secondary care
    Is electronic record keeping expensive compared to paper?

  6. I think we need t remember that vast majority of noting is carried out by non-consultant juniors. If we go to dictation we loose the cognitive support given by structured clinical documents. Structured documents are widespread in organisational efforts to improve care (we can debate the efficacy of that another time). I think we need to go further an understand how clinical information is processed at that near-synchronous interaction of encounter & recording. This process will be different according to setting, speciality and grade as PK outlined, but thats why we need a UI that supports a blend of authoring modes. To do that we need to waste less on semantic duplication & vendor lock in and put more effort into developing the front end.....