These “doctor’s notes” or “sick leave” are the attestations we do for a patient when they’re ill and they need a proof of that for work and/or insurance. The contents will differ greatly in different countries, but the ones I describe here are similar to the Swedish ones1.
We can also greatly enhance and streamline the entry of running notes in the medical record. These notes are usually structured as a list of “items”, where each item corresponds to a type of data, or a clinical sign or symptom. The actual selection of which items to use depends on why we’re seeing the patient, as expressed through the “type of contact” (or “problem”). The content (value) of the items, however, is free text, but usually limited to a few variations only. These, the system can learn and present.
While creating referrals in a record system, the workflow is fairly predictable. For any particular kind of problem, there’s only a relatively limited range of referrals you are going to write, so we can let the system record which ones we use and pop up a list of last used referrals the next time we see a patient with the same problem (“type of contact”).
When you’re seeing a patient, it’s very often useful to do a search for diagnostic and therapeutic guidelines, medical articles, regional or institutional recommendations and so on. These searches need to be restricted to appropriate sources, not just the wild internet. Once you’ve done a search and found some useful information, you’d probably want to save a reference to it and be able to locate it again the next time you see this patient, or another patient with a similar problem.
As I see the structure of medical knowledge and its application to patients, there are three levels:
- Biological science, pathology, EBM, epidemiology, etc. In other words, everything we know about human biology and pathology in the large, not at the individual level.
- Applications and methods that apply biological science to the individual patient, and methods using the history of the patient to search for applicable science.
- Knowledge about a particular patient, signs, symptoms, treatments and diagnostics that have already been performed. In short, the individual patient history.
Each of these three levels correspond to particular processes and methods, and computer applications also fit one or more of these levels. For instance, IBM Watson sits squarely in level 1, while current Electronic Healthcare Record (EHR) systems are fully in level 31.
I think the major problem is that buyers specify domain functionality, but not the huge list of “non-functional requirements”. So anyone fulfilling the functional requirements can sell their piece of crap as lowest bidder.
Looking at a modern application, non-functional requirements are stuff like resilience, redundancy, load management, the whole security thing, but also cut-and-paste in a myriad of formats, a number of import and export data formats, ability to quick switch between users, ability to save state and transfer user state from machine to machine, undo/redo, accessibility, error logging and fault management, adaptive user interface layouts, and on and on.
I’d estimate that all these non-functional requirements can easily be the largest part of the design and development of a modern application, but since medical apps are, apparantly, never specified with any of that, they’re artificially cheap, and, not to mince words, a huge pile of stinking crap.
It’s really easy to write an app that does one thing, but it’s much harder and more expensive to write an app that actually works in real environments and in conjunction with other applications. So, this is on the purchasers’ heads. Mainly.
This article is an excellent description of some of the serious problems related to IT security in healthcare.
Even though medical staff actively circumvent “security” in a myriad inventive ways, it’s pretty clear that 99% of the blame lies with IT staff and vendors being completely out of touch with the actual institutional mission. To be able to create working and useable systems, you *must* understand and be part of the medical work. So far, I’ve met very few technologists even remotely interested in learning more about the profession they’re ostensibly meant to be serving. It boggles the mind, but not in a good way.
“Unfortunately, all too often, with these tools, clinicians cannot do their job—and the medical mission trumps the security mission.”
“During a 14-hour day, the clinician estimated he spent almost 1.5 hours merely logging in.”
“…where clinicians view cyber security as an annoyance rather than as an essential part of patient safety and organizational mission.”
“A nurse reports that one hospital’s EMR prevented users from logging in if they were already logged in somewhere else, although it would not meaningfully identify where the offending session was.”
This one, I’ve personally experienced when visiting another clinic. Time and time again. You then have to call back to the office and ask someone to reboot or even unplug the office computer, since it’s locked to my account and noone at the office is trusted with an admin password… Yes, I could have logged out before leaving, assuming I even knew I was going to be called elsewhere then. Yes, I could log out every time I left the office, but logging in took 5-10 minutes. So screen lock was the only viable solution.
“Many workarounds occur because the health IT itself can undermine the central mission of the clinician: serving patients.”
“As in other domains, clinicians would also create shadow systems operating in parallel to the health IT.”
Over here, patients are given full access to medical records over the ‘net, which leads physicians to write down less in the records. Think this through to its logical conclusion…
Caspar Bowden spoke at the 31c3 conference. Snippets:
I told my technology officers at MicroSoft that if you sell cloud computing services to your own governments, this means that the NSA can do unlimited surveillance on that data. […] two months later they did fire me.
“Technology officers” represent MicroSoft in their respective countries.
On the “FISA Amendment Act of 2008 (Sec 702)”:
This means if you are not American, you cannot trust U.S. software services!!
The US congress was laughing, laughing at the idea that you have privacy rights. That is the climate of the US privacy debate.
“You”, in that sentence, refers to non-US persons outside the US.
FISAAA offers zero protection to foreigner’s data in US clouds.
US is “exceptionally exceptional”: The number of references in surveillance law that discriminate by citizenship/nationality (NOT geography of communication path), per country:
US: 40, UK: zero, Germany: 1, Canada: 2, New Zeeland: 2, Australia: 2. No others.
We need to give them watertight asylum, and probably some incentives, some rewards. I actually proposed to the parliament [EU parliament] that the whistleblower should get 25% of any fines subsequently exacted.
Big applause from the audience…
How do people know politicians and officials aren’t influenced by fear of NSA spying in their own private life? […] this is highly corrosive to democracy!
The thoughts that Edward Snowden has put in the minds of people cannot now be unthought.
What this all means, in practice, relating back to medical applications, is that we (Europeans) can’t use US software or services, which includes medical records such as EPIC, data analysis services such as IMS Health, data storage such as Amazon, Azure, iCloud, backup solutions (unless encrypted client side), or even US operating systems such as Android, iOS, OSX, Windows, a series of embedded OS, etc. At least not if we care about our patient’s right to privacy.
My TEDx talk is now public. It’s all about how bad current EHR systems are at helping us do our jobs as doctors. The title is “Rethinking clinical thinking: your doctor’s tools are blunt”. Enjoy it at:
Check out this article on “Improbable Research”. In short, it’s an application that can take raw data and write an article around it. Personally, I think it’s a good thing if the result is more objective and complete than most journalistic writing we see today. Can’t be less researched, at least.
But it also goes to show the opposite, which has a bearing on medical publications. In medicine, we have a huge problem with the sheer amount of articles published. If you want to find out the state of art in some particular disease or treatment, you have to collect a number of articles, skim through them, try to get at the original data that was used (very hard) and make up your mind. There’s not much guarantee of objectivity in selection or interpretation of the articles, and very little objective data on how reliable the articles are. If you can find a (reliable) meta study, it’s easier.
If a machine can produce medical articles based on study data, and those articles look like the real thing, this proves that the prose in the article is not a real value add. In other words, nothing in the text adds information beyond what the raw data already contains. And if it does, it’s probably misleading and wrong, anyway.
In conclusion, this only goes to show that what we need is more studies and less articles. What we need is immediate access to the raw data of all relevant studies and a desktop application that lets us view and manipulate the total of that data according to our needs, without going through the complications of reading articles and reverse-engineer the texts down to the objective facts hiding behind them.
Maybe this heralds the death of medical publishing as it looks today, and if so, good riddance.