From the ‘Patient Copy’ of a printed “General Discharge Summary” handed to a family member yesterday.
Discharge Medications
“No Medications Entered” - in fact a script was handed to the patient containing 3 new medications.
Miscellaneous
“No Pharmacy medicine” - really???
“No Medical Certificate” - incorrect - one was, in fact, provided
Clinical Information sub-headings
What on earth do the following abbreviations mean to a patient…
PC, (politically correct?) BG (good band) HPC (high-performance computing - hardly), O/E (bit old for that :)) and Ix (sent home in the latest electronic BMW?).
Then, there’s the host of clinical shorthand in the details. Clue to those creating artefacts for patients - the vast majority of us haven’t attended medical school.
Of course, nothing that cannot be translated in 30 minutes with Dr Google or generative AI. If the damn thing was in atomic, electronic format that could be achieved almost instantaneously. It might also be transferrable to a patient record!
Until that sunny day, we continue to suffer the consequences of the failure to progress the successful Electronic Discharge Summary Pilot Project led by Patients First back in 2011. IMHO, the single biggest mistake that the our Sector has made in my time. I still recall the experience related by @Dr_Will_Reedy at the HiNZ Conference one year about the American visitor who photographed a printed discharge summary that he’d spent 30 minutes typing and then ripped it up and threw it in the bin!
I could also rant on, at some length, about sub-optimal clinical decision making, impaired by inaccurate paper records and manual processes dating from Victorian England, but will limit this post to the discharge summary. Suffice to say that, ultimately, my family member’s experiences over the past 3 weeks are a sad reflection in general on digital health “capability” at Waikato Hospital that doesn’t appear to have changed in the past 30-add years!
This level of dysfunction at the patient interface is pretty much endemic across our systems. My own personal experiences as a patient down in the South completely concur - and it makes my blood boil.
I dream of a time and place where positive improvements to this are encouraged, supported, and celebrated. And there are patches of NZ that are doing much better on these fronts. Currently, it feels more like a stagnant wasteland.
I do wonder if we as a digital health community could affect this sort of positive improvement despite the milieu we find ourselves in. I certainly don’t have much faith in our existing structures to deliver!!!
This is an interesting discussion. Especially seeing as a lot of work has already been done.
If you go to Google Scholar and use this as your search phrase ‘discharge summary adnan’ you’ll find some great work done as a PhD thesis over 10 years ago by Mehnaz Adnan who now works in ESR. Might be a good idea to bring her into this discussion.
The major barrier is (still) the overall IT capability (or lack thereof) at NZ hospitals. Poor and sub-optimal hardware, outdated software and versions thereof (e.g. CSC logos on IPM Logon Screens at Waikato Hospital). The failure of the National Infrastructure Project (NIP) 10 years ago was a disaster for the Sector. We can design new Discharge Summary specifications - using international standards - but we don’t have the software, hardware or even change management capability at the hospitals to deploy them on a national scale. No point in building sports cars to run on unsealed roads!
Your report is symptomatic of an entire system of dysfunction—as Deming or Balestracci would eagerly have pointed out:
A failure to create a system that makes it easy to do the right thing and difficult to do what here is obviously the wrong thing (Safety II)
Creation of invidious software that doesn’t build around the problems the patient actually has (a failure to listen to the good bits of Larry Weed, back in 1968)
Failure at a deeper level to create data structures that support joined-up, normalised problem-centred models of thought and execution (failure to listen to Ted Codd, among others)
Failure to invest in teaching—especially teaching junior staff to think like human beings and not automatons, steeped in medical jargon (A failure that dates back to Osler and before)
A failure to listen to patients, which is perennial
A failure to listen to criticism, which dates back at least to Semmelweis (1847)
Failure to integrate evidence, causal inference and Bayesian thinking into the curriculum—and failure to support this in software design (Judea Pearl pointed out the causal stuff in the '90s—but most people still don’t get it; we’re only starting to embrace Bayesian now that the Frequentist statisticians are dying off)
Failure to identify where dysfunction is occurring and remediate it (Shewhart worked out most of this in 1931).
Failure to learn from aviation—especially failure to invest huge amounts of money into checklist design. Medicine is far more complex than aviation, but decision support here is inept, cheap, prolix and dumb.
And above all, a lamentable failure of successive governments to invest appropriately—skimping on both money and (more important) simple, orthogonal system design based on a robust strategic model, rather than randomly throwing money at stuff based on poor tactics and who shouts most, and then in turn shouting in response to common-cause variation in their targets, traffic lights and league tables.
Now that we’ve both got this off our chests, what’s the solution?
It’s not more of the same. It’s also not less of the same, as our current government (who seem to never have even heard of Goodhart’s law) seems to think!
As a clinician of 40 years who has also done his 20,000 hours of programming, and who has tried for several decades to get reluctant managers to invest in quality, my greatest lament is still that software designers don’t have a clue about most of the above, and create products that support none of it.
Do you have a solution Peter, or is this just a piteous lament?
I think @DrJo is right about teaching - I’m sure digital services can help but I imagine some of these mistakes would relate to the working conditions of/communication with junior doctors - also an issue that I can’t see being resolved anytime soon
More a combination of frustration and disappointment, @DrJo.
While it would be somewhat hubristic of me to offer an overall solution, I can offer a starting point for present and future publicly-funded NZ digital health projects which is to achieve some form of consensus about what we are attempting to achieve and the resources (particularly financial) that we are prepared to commit to those outcomes. In fact, that might be applied to the entire healthcare system itself.
Unfortunately, as in many other domains, we appear to be drifting apart and our current state is simply a reflection of that. Going back a decade, the Ministry of Health was informed by a number of expert advisory groups (e.g., Sector Architects, HIGEAG, Patients First, HISO, etc.) One by one these have been disestablished and/or subsumed by the Ministry and now Health NZ. Consequently the tension between the government and external parties has just increased and - without delving too deeply into politics & conspiracy theories - has lead to the undermining of initially the National Health IT Plan and now Hira and its various predecessors.
Furthermore, at present the NZ Health IT Sector is operating in a low income, low profit margin, and chronically underfunded environment. Expecting quality software to be produced in those circumstances is possibly wishful thinking.
I spent the best part of 10 years (on and off) working for QLD Health and can advise that there is a vast difference in the amount of funding QH is putting into digital vs. Health NZ. Having worked for HNZ for the last three years it is clear to see the difference the lack of digital funding is making.
With enough time and money, you can achieve most things. Sadly, HNZ appears to be short of money and at least 10 years (and counting) behind Australia in digital health systems.
Hi Inga. Are there artifacts from the work you refer to that could be updated (probably not much has changed) that I could use in engaging HNZ? Waikato don’t use the Orion Te Manawa Taki regional Clinical Portal as you probably know (instead using the software purchased from CSC back in about 2014) so I can’t influence that, but I can raise this (as many of us can) via the NCPUG in order to advocate for prioritising action.
Governance from HNZ will of course have to agree that the current situation is unacceptable and that something should therefore be done. As we move into 2025 we will be actively working with HNZ on this process and support will be essential to defining our work plan. @NathanK based on your experience in the South and yours @MValentine in Te Manawa Taki I’m sure you have constructive feedback for how the Orion portal generated discharge summaries in your regions could be improved (separating software capability from the other issues incl teaching etc further down the thread).
I’m all ears!
I believe the remedy is with private organisations outside of our public health system pushing forward with new innovations. Flawed discharge information has been discussed for the past 30 plus years. It shouldn’t be that hard!!
Yeah, it’s a bit like going round and round the mulberry bush like in the nursery rhyme, looking like we’re doing something and getting nothing done. We should all know how a discharge summary should work, who the audience is, and how communication skills are most important for good outcomes. A quick n dirty literature search on Google Scholar shows that back in the 90s there were articles that established what quality discharge summaries consist of.
Van Walraven C, Weinberg AL. Quality assessment of a discharge summary system. CMAJ: Canadian Medical Association Journal. 1995 May 5;152(9):1437.
In the 20-teens there was enough literature for a systematic literature review:
Wimsett J, Harper A, Jones P. Components of a good quality discharge summary: a systematic review. Emergency Medicine Australasia. 2014 Oct;26(5):430-8.
Thing is, things have changed a lot. The issue of understaffing and time-sensitive work and the general environment of hospital clinicians means that a good quality discharge summary isn’t only the result of a good template and good training, it’s also the result of environmental impacts on clinical work, and the psychological impact of interruptions, being time-poor, and having a bigger workload that previous generations have had.
So how do we fix this? We can’t only rely on great tech. AI might show promise, but who checks the content for accuracy and completeness, and who gets into trouble when errors cause harm?
A national standard for discharge summaries developed collaboratively would certainly help, but would need to be bolstered by a robust change / support process or we’ll have the same sort of dysfunction continuing (just standardised instead!).
At present, I have little confidence that could possibly be delivered by our existing MoH / HealthNZ structures.
It would be an interesting exercise for us (in this community) to develop this standard, and then trumpet loudly about it until people / the system is forced to take notice.
Certainly the status quo is intollerably awful and downright dangerous. We’ve got to do something! @ciln-advisory - would this be a useful thing to plant a flag on?
I agree, a national standard would be helpful so that anyone who learns how to do discharge summaries can know that they don’t have to tweak what they do when moving to another unit/organisation, or learn to do it from scratch over and over again. This assumes that the same software and template are used (with, at the very least, robust interoperability) regardless of who does the summary and how it’s completed.
I recently saw a really good discharge summary when my painting buddy’s 92 year old wife (who is digitally literate but not medically literate) asked me to interpret the 3-page summary for her. Right at the end, all that she needed to know was written (on the third page where she didn’t see it) where there was a designated space for instructions to the patient. But she wanted to understand and check the accuracy of the whole document.
My questions: How bad and big is this problem? What needs to be fixed? What research has already been done to guide us (I’m channeling Jo Schalkwyk’s post in this conversation)? What is required to fix the problem? What does the solution/success look like? How do we know we’ve fixed the problem sustainably over time?
