One of more
depressing enlightening aspects of growing older is the recognition of the emergence of more subtle parental traits. Both of my parents did History degrees, and although I have always seen myself as a scientist, as my career has moved on, I have come fascinated by the place of ‘narrative’ in medical records.
As an advocate of openEHR and highly structured, coded interoperable records, I might be expected to shun all things ‘free text’ but as a GP I recognised that many things are not codable, and even when coding/structure was appropriate, there was, at least for me, a tangible, cognitive effort in translating my analog thoughts into digital code and structure.
There is another aspect of narrative in medical records, which is poorly understood and bedevils efforts to interoperate between different care settings: the meta-narrative or summary.
Whilst the prevailing view of the medical record is as a steady accumulation of facts, a journal or register of events, in fact it is a highly interpretive, and re-interpretive, set of documents. The patient’s medical story is continually revised and re-told as new facts emerge, old ‘facts’ are disproved or revised and the raw data is compressed and summarised to make it intelligible. Some years ago had an interesting debate with my colleague, Thomas Beale, about what represented the ‘truth’ of a medical record. As an engineer he favoured the raw factual, observations, whilst as a clinician I saw ‘truth’ in the most recent re-telling of the patient’s history.
This is what we do every time we transfer care to another clinician, or impart the story to a senior : ’55 yr old man, previous mild MI 2 years ago. non-smoker. Presented with 2hr central, crushing chest pain’. Concise, effective and rolls up a whole host of ‘facts’ into the salient information that make onward, human information processing possible. Some informaticians believe that effective clinical decision support is only possible if it computes primarily on this kind of meta-narrative, rather on ‘raw fact’.
One of the barriers to interoperability is that different care settings demand rather different approaches to meta-narrative and it occurs to me that a patient’s health story is rather like a soap opera such as EastEnders. There must be a whole world of unreported events that go on in Albert Square unreported by the patient or scriptwriter since they seem uninteresting, but other collections of events are brought together to form interesting storylines, some dramatic, others with ‘slow-burn’, others still are non-pathological.
The GP is rather like the dedicated EastEnders watcher, aware of subtly developing storylines, of background, of complex relationships. The GP record will try to reflect these nuances and will (should?) be regularly updated and revised to reflect the twists and turns of who is in-favour, out-of-favour, or about to be killed-off. The GP/regular needs to be able to pick up all these threads at the next episode so that the ongoing drama can be refreshed in his mind with minimal effort.
In contrast, the hospital clinician, is like the occasional viewer confined to EastEnders blockbusters- the Christmas specials. The story is dramatic, eventful, and to a large extent, self-encapsulated. At the start of the episode ,the specialist does need a concise, summary of relevent events and facts to orientate himself to the developing crisis, but the subtleties of relationships and interwoven story lines seem irrelevant when one character is about to throw another from a roof. The specialist’s meta-narrative needs to focus on the details of the episode, not the long-running story, and whole load of ‘stuff’ will happen in the episode, much of it critical to the immediate storyline, but irrelevant thereafter.
Where real-life differs, of course, is that our GP is not able or willing to watch the ‘Christmas special’ episode but expects the specialist to update him with the salient facts at conclusion of the the episode. The details of which roof, which door, the struggle etc are unwelcome , the GP just wants to reconcile the highlights back into his longitudinal view. Very strange things happen in ‘Christmas special’ episodes which are often best regarded as anomalies or ‘continuity lapses’ in the outside world.
In Beyond the Hawking Horizon, Ewan Davis nicely summarises issues around single, shared records which make safe, reliable use difficult in the absence of clear record governance. I think some of these issues may be a further example that, while the stream of facts might be usefully shared in a community setting e.g lab results, appointment schedules, medications, tasks due, we struggle to share coherently when we introduce interpretation of these events, particularly diagnostic labelling and attempts to construct meta-narratives of the patient story suitable for our own specific clinical setting.
The trick will be to find ways of sharing interoperable information seamlessly, where this can be done usefully and safely e.g. Community medication records, but with ‘border controls’ where automatic import would interfere with each care setting’s need to tell the patient story in a way that is fit for its own purpose and to support internal processing. I don’t believe that these meta-narratives can be algorithmically constructed but will require careful human thought and quality control. There is some interesting work emerging in the US, where they are looking at the ‘Reconciliation’ of problem lists, in the same way that we currently think of medicines reconciliation. In some respects this seems like a backward step, engendered by the US being some way behind the rest of the world in interoperability potential, but perhaps instead, it simply acknowledges the Hawking Horizon, and is developing useful and tractable solutions.