back to article Bad software destroyed my doctor's memory

"I have a problem," my medical specialist said – before he saw the fear on my face and quickly corrected himself. "No, not with you – you're fine! With this." Both hands swept past the shiny new convertible laptop-slash-tablet that sat on his desk. "It’s not working?" I asked. "Oh, it's working," he replied. "We spent six …

  1. My-Handle

    One of my best software development projects was also one of my earliest. Despite the fact that I was still green, I had the massive advantage that I was developing this tool while sitting with the team who would be (and latterly was) using it. It was originally designed to solve a couple of problems they had with their workflow, and any issues or complaints they had they could tell me directly. I could also observe them using it passively, over the course of the working day. I could see what worked, what didn't, what made sense to me but wasn't intuitive for them, and what else might be added to further streamline things.

    That project taught me an awful lot about UI.

    1. Doctor Syntax Silver badge

      This was also my experience. I also had the advantage of coming into IT as a user. If you were a scientist and wanted a program to do something you either wrote it yourself (my option) or asked somebody who could (me sometimes being the somebody). Back in those days the job title was very often "analyst/programmer". It seemed perfectly natural to sit with the users, listen to what they wanted, write up what I thought that was and then present it back. That rewriting in one's own words is an essential step in gaining mutual assurance that what was said has been understood correctly.

      1. LybsterRoy Silver badge

        We need a new role & title - User Programmer and ideally part of the analysis should be actually doing the job for at least a short while. I'm not saying you shouldn't challenge poor workflow but you should UNDERSTAND what the workflow is and why.

        I'm also reminded of my favourite definition of "user freindly" - its what they're used to!

        1. bpfh
          Coat

          I'm a previous job many moons ago...

          Agile was bandied about but it was more waterfall development...

          We had a functional project manager team who designed a PowerPoint of the general look and feel of the pages we needed to build. They would come to talk to the technical project managers - and sit with them as they explained their needs and goals, the TPM would explain what was feasible and what was not - or what would be over budget or over time. Once the FPM had an executable design, it went into development within the TPM's team and went to preproduction.

          The FPM would then test and check, then return to sit with the TPM and snag the design, if needed get the devs to come and change the design sometimes on the fly, the push to preprod. The FPM went and presented the changes to the stakeholders, got a sign off or another sit at the TPM's desk for adjustments. The FPM's guaranteed the UI/UX/expected functionality with the client, and made the bridge with the technical teams that made the magic happen behind the scenes. Everyone talked, everyone knew what was expected and what was feasible, problems were overcome in real time, clarifications could be asked in real time and projects ended up delivered globally on time and under budget. Having a team that knows ui and ux and able to guide the teams that actually delivered the behind the scenes technology, nobody evolved in a vacuum.

          Today I see too many designs that have a technical target to reach, but the ui/ux is forgotten, ending up with a lot of client complaints because what was fast to design is not alway usable or logical... and all it took were 2 empowered project managers out of a team of 30 engineers to make magic happen.

    2. Alan Brown Silver badge

      One of my constant complaints when trying to turn a project into a frontend for users is that it's virtually impossible to get people to test things (for various reasons, not least of which is "they're too busy" for manglement to spare them)

      The result is that devs work in a vacuum and it's extremely harmful, where weeks are spent polishing the graphics on a turd which is difficult to drive and use no matter how shiny the wrapper

  2. Benegesserict Cumbersomberbatch Silver badge
    Angel

    Have a hundred upvotes

    Abso-bloody-lutely.

    I daily use a digital medical record which mimics the layout of the old paper record, segregated into alerts / outpatient episodes / inpatient episodes / diagnostics / correspondence sections then partitioned according to patient contact episode, and internal sorting of those according to admission / progress / specialist inpatient (ICU/theatre/diagnostic procedure) / discharge by the hospital paper form MR type. Precisely what the paper record had evolved into after centuries of human experience showed us how to minimise the use of any sentence beginning with the words ”Where in this thing am I supposed to find...".

    This is a perfect exemplar of you don't know what you've got till it's gone. No-one gives this problem a second thought until some bright spark comes along and, presented with a problem that is already solved, says "no, we should do it this way..." and tries to implement some system that was abandoned the last time it was tried because it didn't work as well.

    1. tfewster
      Coat

      Re: Have a hundred upvotes

      Computers could certainly enhance a paper file, assuming the contents are filed properly. I would hope that system has a chronological summary of the patient history but also allows

      - Searching for any term in the file, e.g. "hip". Easy in a digitized file, not so easy in a paper file..

      - Linking an Episode, e.g. a broken hip treated in a hospital plus outpatient physiotherapy sessions plus disability assessment.

      - Selecting an item, then paging backwards/forwards through related OR unrelated pages. (e.g. You had a fall, but also have low blood sugar. Hmm, could they be related? The human brain is great at making correlations, computers not so good)

      - Side-by-side comparison of two pages, e.g. blood test results, to see what has changed. Maybe even highlight changes or anything out of the norm.

      That's just off the top of my head, and with little knowledge of current practices. Hopefully those involved are way ahead of me!

      (It occurs to me that any CRM system should have those features, so the wheel doesn't have to be reinvented for each use case)

      1. Doctor Syntax Silver badge

        Re: Have a hundred upvotes

        A CRM might but a medical application might also hooks to take various lab reports and be able to adapt as more tests are developed. It would need to have the features of a LIMS.

      2. Scoured Frisbee

        Re: Have a hundred upvotes

        Most likely every EMR system has those things already, but:

        - no doctor is going to receive more than minimal training, and nobody is going to pay them to fuss about with the system off hours, so they won't know how to get to these features until some new kid shows up and figures it out

        - paper records are scanned and - if you are lucky - OCR'd with Acrobat or some other barely-functional product, so searching keywords is at best 50/50

        - Handwriting OCR is... Well have you seen doctors' handwriting?

        - dates of paper records don't OCR better than anything else; especially if you have a pen-and-paper doctor the records are probably scanned and presented in the order received, muddy footprints and all

        - for a specialist in the US, lots of stuff - especially lab results - come in as PDF images or faxes, so they aren't really easily searchable even if the site has been using an EMR for years

        Digitizing paper is a pain. Once the patient records are entered digitally it gets easy enough to do all the things you mention and more. It is, however, very painful in the meantime, and having something transitional as mentioned in the article would at least keep down the error rate in the near term.

        1. Benegesserict Cumbersomberbatch Silver badge

          Re: Have a hundred upvotes

          Digitising paper records works.

          It is easier to make words on paper look like the thoughts in my head than it is to make words on a screen do the same thing. If you can't find a way to do that more efficiently than pen and paper, give me pen and paper and pay someone to scan it.

          Information handling in a medical context is so important that increasingly hospitals are hiring scribes to work with doctors as functionaries, to make the orders for tests, writing of notes, liaising with allied health and other practitioners just happen, to free up actual medicine. It's not just painful to reduce the quality and clarity of medical recording and communication, it's inefficient and dangerous.

          Doctors don't generally need text search. We need event search. If it's a familiar patient, we use our memory. If not, we use other doctors' and nurses' notes.

          OCR is only useful as much as things like free text search are useful - not as much as you would think. Anyone who contemplates OCRing a medical record hasn't seen one.

          Collating results from different providers and integrating them in a record is the back end - and that's IT's job. If it doesn't work, we'll call you.

        2. LybsterRoy Silver badge

          Re: Have a hundred upvotes

          You missed out one small fact - most doctors (in the UK at least) can write faster than they can type. The end result may be illegible or in some foreign version of English but its faster.

      3. Elongated Muskrat Silver badge

        Re: Have a hundred upvotes

        - Searching for any term in the file, e.g. "hip". Easy in a digitized file, not so easy in a paper file.

        Actually, not as simple as it sounds. The way we search digitally isn't the same as the way we search physically.

        For example, if searching through a paper medical record for things to do with hips, that would probably involve looking for things that clearly relate to the orthopaedic inpatient department of the local hospital, specific hand-written notes, and so on, and scanning them for things that mention hips or related terms ("pelvic", perhaps?). Your brain is doing a lot of different things seamlessly when you do a search like that.

        On the other hand, searching for occurrences of "hip" in a load of digital medical files will miss things which haven't been OCR'd or categorised correctly, and will probably pick up 300 instances of "medicine shipped" that you have to wade through.

        Someone has to be trained to use the second mode of searching, because the first (the system they have been using since birth) is not applicable.

        1. An_Old_Dog Silver badge

          Re: Have a hundred upvotes

          That's why I print (some) computer manuals.

        2. Kristian Walsh

          Re: Have a hundred upvotes

          100% agree. The golden rule for line-of-business apps is that you NEVER assume your user knows nothing, but consumer app design is based on the idea of a user who needs to be coaxed into choosing to use your app. This use of consumer design patterns in professional applications software is the biggest design flaw I’ve seen in this kind of software, and where it comes from highlights the root of the problem (hint: who does the development team liaise with? The users, or the management layer that is paying for the software?). These aren’t consumer products; the user of these system know a lot about their domain. They know a lot more about it than some database designer who is thinking only of how data elements relate to each other, or the “UX architect” who will inevitably try to impose “surprise and delight” and onboarding patterns onto something that its users already need to use.

          Users invariably know exactly what they want, but they’re just not great at articulating it. (e.g., Try to explain to someone what you need in a text editor.. I guarantee you’ll leave out something basic like search/replace). The existing system they have is the best requirement specification for any digitisation project: that works. For them, it’s accessible, and it is adaptable, and it is the accumulation of years of small process improvements. That system could be a collection of spreadsheets, a document pile or a paper file, but it is the most important thing to study, and not just its structure, but how it is used. It is your job as a development team to preserve that knowledge, not throw it all away in in favour of the “new flashy”. Your users don’t need onboarding or guidance on how to do their job, they know how to use the tool already; they don’t even need pretty visuals - for desktop/laptop systems, information density is often more important than something looking aesthetically pleasing.

          As an example, my GP uses the ugliest piece of software I’ve ever seen, but she swears by it.. they’ve it’s exactly like a paper case-file or an email client. Open my page, and you get a summary of my details, a list of active prescriptions, then the rest of the screen is basically like an inbox of all the notes and documents about me, in chronological order. It isn’t pretty, but it keeps the information organised in a way that makes sense to a doctor running a general practice, and reduces the amount of clicking and hunting needed to find things. The first few years I visited the practice, I kept thinking “that software is shit, surely I could design something better”, but after a while seeing how well it suited the workflow, I realised that I was confusing aesthetics with usability. An important lesson learned. (And my doctor’s been pretty good too).

          1. Doctor Syntax Silver badge

            Re: Have a hundred upvotes

            "the user of these system know a lot about their domain. They know a lot more about it than some database designer"

            They very likely know a lot more about it than their manager. If the manager wants to lay down how the system is to be used you need to get past them somehow. "Yes, I quite see your point but we have a set process. There's a 'Speak to user' task in this phase. More than my job's worth mumble mumble. I'm sure you'll agree I've got to be able to say I did it."

            I'd say the design is likely to evolve from two different starting points. One is how it's supposed to function - the UI, the other is the data. The data which might be dictated by the domain, by established standards such as RFCs. The balance between the two will vary from project to project - the UI might even dictate the data at one extreme.

      4. CowHorseFrog Silver badge

        Re: Have a hundred upvotes

        Most search is broken because different people write down or type different things for the same idea or concept. For medial matters, i would imagine those horrible product names for starters are probably writen ten different ways by ten different doctors and this means search doesnt work or is incomplete.

      5. Alan Hope

        You're kidding, yes?

        As a clinician using computers in a major hospital I have to smile at your naive optimism. You have no idea what we deal with daily, and I work with recently installed and expensive systems.

        In our prescribing system searches/filters only work from the beginning of a list of terms - you can't search on the second word/term. Wildcards are not implemented.

        Multiple scrolling windows within scrolling windows - even worse mostly negotiated from Laptops with dodgy touchpads.

        No confirmation messages ever: for example, order a blood test and click on "Submit", you go immediately back to the home page. No "order successfully sent" message. Did it go? Wait and see.

        A near random position of buttons in screens: "next" at the top, next screen "next" at the bottom, next screen "next" on the right.

        No "Print" buttons in one major package, all the print functions are called from buttons saying "Complete".

        Labels that become buttons then revert back to labels without changing their appearance in any way.

        An almost invisible very pale grey X in a white box as the close button in some windows.

        I could add a hundred other basic interface disasters to this list. Medical software interfaces are clearly written by disinterested 16-year-olds doing School projects.

        1. Doctor Syntax Silver badge

          Re: You're kidding, yes?

          A few years ago I got off a train and saw in the station a large advert for programmers - "No experience necessary". The company that writes the system my GP uses is based in that city and they were the ones who were advertising.

        2. FatGerman

          Re: You're kidding, yes?

          >>Medical software interfaces are clearly written by disinterested 16-year-olds doing School projects.

          I believe most of them are actually written by unqualified humanities graduates working for large consultancies and earning 90K a year.

        3. Groo The Wanderer Silver badge

          Re: You're kidding, yes?

          >Medical software interfaces are clearly written by disinterested 16-year-olds doing School projects.

          Not true! They're highly motivated $5/hr contract programmers overseas working in body shops, not one of which has ever delivered a properly designed system. But they're the lowest bidders, so...

        4. Alan Brown Silver badge

          Re: You're kidding, yes?

          "Medical software interfaces are clearly written by disinterested 16-year-olds doing School projects."

          "Kids in Bedrooms" is the term used by one old manager I worked with to characterise Open Source software

          He didn't like it when I co-opted the term for shitty UIs in "professional" applications and liked it even less when I used the term in a meeting with one vendor

          As I pointed out to him, when OpenSource has a shitty interface, someone always fixes it (and if it has a good interface, someone always comes along with an unusable variant), but for commercial wares you're stuck with whatever the vendor supplies

          Opensource frequently ends up with a plethora of ways to access the same data, tuned to the requirements of the people using any given interface. This is NOT a disadvantage

      6. Dr Dan Holdsworth

        Re: Have a hundred upvotes

        As long as you design the file so that the medical person has view of the notes, the drugs being taken and the flagged conditions on the page then you should be doing reasonably well. If you add to this a database of known drug interactions and known allergies that each patient has, and you allow the main system to add in hints that an allergy to antibiotic A might mean that the patient will be also allergic to closely related antibiotic B, then you also have a useful system.

        Similarly, if you have in the notes details of the patient's age, biological sex, weight etc. then you can also put in a checking system at the prescription level, which checks that the amounts of drugs being prescribed are inside the guidelines for those drugs and drug interactions and which also flags up potential for abuse of these drugs.

        That is adding value to the system.

        Before this, you have to duplicate the existing functionality and add in a little extra robustness such as off-site backups with encrypted transport.

        1. Alan Brown Silver badge

          Re: Have a hundred upvotes

          All of what you've said is valid - and very rarely happens

          I have to remind my GP that I'm critically allergic to penicillin every time they want to prescribe antibiotics

    2. very angry man

      Re: Have a hundred upvotes

      I am doing testing on a travel app,

      I am the worst person for this, and possibly the best.

      A borderline technophobic turrets sufferer.

      I'm testing on a motorcycle, so has to work with one or two touch's.

      And when it doesn't, I try to convey my problem and frustration, with at least some words describing the details, amongst the cussing and threats of dismemberment or worse.

      That's why they use me, "the lowest possible common denominator" .

      But when All work's well, peace quite pay raises and the occasional lobster dinner.

      So far we have managed to not redecorate in the style of the Texas chainsaw massacre, but they do get more productive if I spend a little while sharpening to chainsaw chain

      1. Alan Brown Silver badge

        Re: Have a hundred upvotes

        "I am doing testing on a travel app"

        The best tester is the one who constantly breaks things

        This is usually the tester that the devs hate the most and don't want to deal with

        Given a hard choice of jettisoning the tester or the devs, I'll keep the tester

    3. Anonymous Coward
      Anonymous Coward

      Re: Have a hundred upvotes

      I've noted a problem, when looking into what are blithely called "knowledge management systems" by organisations, that those overseeing development want to see structure - and a structure they can understand and document. When actually getting down to usability (at the user level) I found that a wiki approach was far better: easier to use, both in adding information and retrieving it. But, even showing such system pilots to manglement, they still wanted a system that looked like the one they were replacing because it wasn't't working.

      And talking about the wiki (freeform) approach, can anyone remember the database app from WordPerfect, sometime back in the 1990s, that allowed you to link items of information in an open structure? I'm struggling to remember what it was called...

      1. midgepad

        Re: Have a hundred upvotes

        Idealist, originally by Blackwell the publisher was one of them.

        Bought by a seller of databases to libraries, and suppressed so as not to compete.

        Nice bit of kit.

  3. Ian Johnston Silver badge

    The whole history of IT and medicine has been littered with failed digitization projects. Almost always that boils down to IT professionals trying to radically alter the workflow of medical professionals, without their input.

    Not just the medical world. Every field is bedevilled with IT geeks who think they know better than actual practitioners how those practitioners should do their jobs. Or who claim to know better as a cover for not taking the time or trouble to produce a usable system.

    The GNOME projects bizarre "usability" guidelines are just one small example of this; the problem is pervasive. Ever tried using SAP Concur?

    1. yoganmahew

      When I started, I was in Marketing Automation or Agency Automation (in the airline/travel industry). The idea was the existing paper processes were good and existed for a reason, and the IT challenge was to automate them. That was it. Not to reinvent them, not to put a one-size-fits-all solution in.

      1. Doctor Syntax Silver badge

        Yes, the time for reinvention is when what's in use doesn't work.

        1. Alan Brown Silver badge

          OR if what's in use is cumbersome and what replaces it REDUCES effort

          One of my pet bugbears has been when I've brought in a system which has XYZ all built in and manglement insist on tasking people to go develop XYZ all over again (usually badly, for all the usual reason expounded here - users are utterly crap at giving a full list of requirements)

    2. Greybearded old scrote Silver badge

      99 times out of 10, yes. On rare occasions an IT expert can see an improvement that the user can't imagine or understand until they see it. That's the origin story of Dan Bricklin creating visicalc after all. (Can't find a reference to the initial resistance he encountered.)

      The few occasions when I've seen realisation dawn on a colleagues face are some of the prouder memories from my career.

      Of course you absolutely must understand what they need to do and why first.

      1. Ian Johnston Silver badge

        Isn't that more a case of a developer offering the users the opportunity to do something additional, rather than trying to modify their current practice? Innovation is absolutely fine; fucking up working practices you don't understand in fields you don't know about because you think you know better is not.

      2. NeilPost

        I think you are mangling BA - Business Analyst who should assess and understand the User Requirements and define them for the IT Expert (as in programmers and system implementers) to build.

        In all of the above lamenting and belly-aching - Business Analysts, User Requirements are not really mentioned at all.

        Hardly surprising you end up with a dogs dinner.

        Far too much blind faith/belief that existing manual processes are always good - if you digitise and automate a shit manual process/cottage industry it becomes a shit digital process, and often expensive custom dev.

        Also no mention of off-shoring to least xommon denominators.

        1. LybsterRoy Silver badge

          I think you are overly hopeful that another "expert" will solve the problem. You almost get there but then you use should when the correct word is MUST and forget a critical step so ...

          who must assess and understand the User Requirements and demonstrate they do understand them

          Existing manual processes may well be shit but they're what the user knows. Trying to force a change because the "expert" knows better is a sure way to problems.

        2. ChoHag Silver badge

          > Far too much blind faith/belief that existing manual processes are always good

          They don't have to be good, they have to work. Usually --- surprise! --- they already do.

          > if you digitise and automate a shit manual process/cottage industry it becomes a shit digital process, and often expensive custom dev.

          And chances are, the cottage industry's owner will have exactly what they want rather than what a developer tells them they want.

          1. Anonymous Coward
            Anonymous Coward

            Management is key

            The only thing which really matters is the change management process, which nobody wants to pay extra for.

            Take GNOME 3: To this day, people can still use the old UI style, so no change was ever actually forced, but people remember how broken and half-arsed the original releases were. Had Flashback (later called Classic) been a prerequisite for releasing the new major version, and a default for in-place upgrades, nobody would have cared about the new UI being pushed,for new installs.

        3. Alan Brown Silver badge

          "if you digitise and automate a shit manual process/cottage industry it becomes a shit digital process"

          It's ALWAYS worth analying every step along the way anf find out the hostory of why something that doesn't make sense is included in the process.

          It's usually a hangover from kludging around something broken that was fixed decades ago. Once you can point that out it's easier to get buy in on changes

          Another thing to bear in mind is that apart from management's near patholigical need to see people "busy"(*), people themselves don't like twiddling their thumbs and anything which halves their time to do a task may result in pushback unless there are other ways to keep them occupied

          (*) This applies to maintenance staff in partucular. The guys you see sitting around playing cards are "on call" waiting for something to break. If they are able to play cards all shift then something is going right(**) The guys you can't find when you go to their office may be down a hole somewhere clearing out drains or in a roof inspecting for rot. They're seldom skiving off in the bogs as many micromanagers seem to believe

          (**) This applies to things like backup systems too. My robots had 6 tape drives and a hundred slots, not to speed up backups but to ensure rapid restores when the shit did hit the fan. The mangler who only sees 2 drives and 8 slots in use on a daily basis overlooks that requirement

    3. anothercynic Silver badge

      *makes the famous Lucille Ball expression*

      If I could upvote this more than once, I would!

      (ICYDKI: https://giphy.com/gifs/1172G00AYkL9ra)

    4. Anonymous Coward
      Anonymous Coward

      "Ever tried using SAP Concur?"

      Concur is absolutely fantastic compared to the homegrown spreadsheet I had to use at $dayjob[$current-1].

      Of course, it sometimes thinks my airline's logo is my actual receipt when I forward the email.

    5. ChoHag Silver badge

      I knew a developer once who was trying to make an interface for a local delivery firm's drivers to scan packages. He was struggling with some design aspect or other and I asked him if he'd been out with a driver for a day to experience first hand how the work is actually done. He looked at me like I'd escaped the asylum.

  4. Pascal Monett Silver badge
    FAIL

    "radically alter the workflow of medical professionals, without their input"

    And there is the problem. You do not design an interface without spending some time with the people who are supposed to use it and learning about how they work.

    Designing a product without ever talking to its users is a high sign of arrogance, these days. That only works when you are creating something that never existed before.

    The word processor is a perfect example. When the first word processors came out, nobody had ever even dreamt of it. There was no one to talk to, because it was an entirely new product.

    Today, if you embark on making a new word processor, you hardly need to talk to anybody because if it doesn't work like Word, you're up the proverbial creek without a paddle.

    So you're designing a product to "help" doctors ? Go talk to a doctor. Find out what they need and how they work, and then go back to the office and wrack your brain to find out how you can replicate that in your UI.

    Not doing that just makes your product useless, and at the mercy of anybody who understands the issue and gives two hoots about it.

    1. Fogcat

      Re: "radically alter the workflow of medical professionals, without their input"

      I first made a comment about this 8 years ago.

      https://forums.theregister.com/forum/all/2015/09/02/nhs_to_go_paperless_by_2020_says_info_director/#c_2620739

      System designers absolutely must talk talk to the users. And not just talk, watch them do their job, try and spot the things they do so instinctively that they would think to tell you about them.

      1. Hans Neeson-Bumpsadese Silver badge

        Re: "radically alter the workflow of medical professionals, without their input"

        System designers absolutely must talk talk to the users. And not just talk, watch them do their job, try and spot the things they do so instinctively that they would think to tell you about them.

        I agree with that in principle, but the key verb here is "listen to" not "talk to"...I've seen too many cases where someone has been sat down with a user and after less than a minute of observation is telling them that they're doing their job wrong (thus demonstrating lack of domain knowledge/understanding, and setting up barriers between the user community and the IT bods before we've even started)

      2. J.G.Harston Silver badge

        Re: "radically alter the workflow of medical professionals, without their input"

        "System designers absolutely must talk to the users. And not just talk, watch them do their job, try and spot the things they do so instinctively that they would think to tell you about them."

        I've just come off a project and sent a shouty email to the project designers saying exactly that about their process documentation. They MUST go out on site and observe the users attempting to use the atrocious rubbish you've written. This in response to an email from *them* complaining "why oh why oh why are the on-site techs getting everything wrong, we've added even more clarification to the documentation, how on earth can we make it clearer?"

        Some examples:

        * Half the documentation is in the email booking, half the documentation is in the attached process documentation.

        * The process skips between different processes, mentions two options to follow when there are three, and doesn't explain why to use which.

        * Pictures with red text on black background. I don't know what model eyeballs they use, but they're not human.

    2. Doctor Syntax Silver badge

      Re: "radically alter the workflow of medical professionals, without their input"

      "Go talk to a doctor."

      Not just one, several. Several in each of the specialities you're supposed be covering. Sketch something out* and then take it back. Only when they've something in front of them will they remember some of what they missed out. What they tell you is likely to be what they were doing last, that's why you might need to keep going back.

      * Lot's of ways for this.

      -Some sort of RAD system might well serve for this even if it's not what you propose to use.

      -Enterprise Architect has an diagram option that uses common UI elements you can arrange, then add what happens for each click or change. It can give you a sketch of what the screen will look like and a narrative of how it works.

      - With Informix I could draft out what I thought the main tables the database was like and generate a default Perform screen on that. It was often enough to discuss with the users to get their agreement or refine as needed.

      1. anothercynic Silver badge

        Re: "radically alter the workflow of medical professionals, without their input"

        *standing applause* This. That's what RAD was all about. Build prototypes to test things. Then take the prototypes away and build the real thing. Then run A/B testing on it with the users. Does it work? Does it do what they want? Iterative design is so important. But on cheap-on-paper quote-and-build T&M contracts that's not possible.

        1. Doctor Syntax Silver badge

          Re: "radically alter the workflow of medical professionals, without their input"

          "Then run A/B testing on it with the users."

          It's my belief that systems in general - but particularly web sites - should be tested by a team consisting of a user, a developer and an invigilator. The user has a series of tasks to carry out. The rules are the only question allowed to be asked by the user and answered by the developer is of the form "Where does it tell me how to ... ?". Officially the invigilator is there to enforce that rule and maybe score the performance. In reality the invigilator is there to stop the other two coming to blows.

          1. LybsterRoy Silver badge

            Re: "radically alter the workflow of medical professionals, without their input"

            If we can add that the developer must have a machine at least two generations older than the ones their target market will use and internet that runs at dial up speeds I'll go along with you.

          2. anothercynic Silver badge

            Re: "radically alter the workflow of medical professionals, without their input"

            The developer shouldn't even be in the room. User with invigilator, yes, user alone with invigilator watching by video, maybe. That's what good UX testing is like. The more the user struggles to find what they need, the more the case is made that the UI is suboptimal and needs changing.

            I worked with an absolutely amazing UX designer, her work was stellar. The eye tracking software they used was fascinating.

            1. Doctor Syntax Silver badge

              Re: "radically alter the workflow of medical professionals, without their input"

              If the developer isn't in the room they learn nothing.

              1. Anonymous Coward
                Anonymous Coward

                Re: "radically alter the workflow of medical professionals, without their input"

                Not true.

                Accepted best practice is for the developer to be watching (one-way mirror or recording), but not in the room. This prevents them from influencing what the user is doing (by "trying to help", making frustrated noises, etc).

              2. Alan Brown Silver badge

                Re: "radically alter the workflow of medical professionals, without their input"

                In my expereience: They frequently learn nothing anyway, unless threatened with pain (financial is almost as effective as physical, but emotional works best. People HATE being ridiculed in front of their peers but if someone's constantly ignoring what the users want then it's an effective tool)

                I think whiffle bats should be mandatory in any development environment

        2. NeilPost

          Re: "radically alter the workflow of medical professionals, without their input"

          RAD … Pft. Junk that, we need Agile. Innit.

          1. Toni the terrible

            Re: "radically alter the workflow of medical professionals, without their input"

            Nah, junk that use AI to sort it all out

    3. Elongated Muskrat Silver badge

      Re: "radically alter the workflow of medical professionals, without their input"

      The impossibility of doing this properly becomes apparent when you consider that every single general practitioner has their own way of working and organising materials. There is no standardised working practice for a GP's practice, let alone record-keeping or IT standards.

    4. J. Cook Silver badge
      Go

      Re: "radically alter the workflow of medical professionals, without their input"

      This is true for any application.

      We have a system in use that captures certain physical security events from various sensors placed around the facility. accessing these events is done through a web page, which from what I understand is a horror show of un-usability, to the point where we are putting in an adjustment to have the system send emails to people instead with the event data, mainly so it can be searched easier.

      UX and UI design is something that should be near the very top of the list for designing functional applications, be they on-premise or cloud based.

      1. LybsterRoy Silver badge

        Re: "radically alter the workflow of medical professionals, without their input"

        You have identified part of the problem "UX and UI design" - stop using buzzwords / acronyms. They teach those in uni these days and you get light gray text on a bit lighter gray screen, boxes / controls without borders so you can't see what belongs where.

        Our local council are proposing a change to road layout to "promote active travel to school" - they actually want the kids to walk or cycle to school.

        Jargon, overthinking and a retreat from first principles of usability are not going to produce good systems.

      2. anothercynic Silver badge

        Re: "radically alter the workflow of medical professionals, without their input"

        The sad thing is that UX is usually the last thing looked at, by which time a) the application is too far along to do something substantial, and b) the developers don't like what the UX/UI expert mandates.

        Often things are designed as a stop gap, which then becomes too entrenched to be removed, and that artefact then eventually is a problem. But because everyone's used to it, well... you know.

    5. herman Silver badge

      Re: "radically alter the workflow of medical professionals, without their input"

      Wordprocessor - new??? In the beginning there was ed. Adam used ed on his abacus to write love letters to Eve.

      1. CowHorseFrog Silver badge

        Re: "radically alter the workflow of medical professionals, without their input"

        Adam and eve would no have known how to read or write. Education costs a lot of time, and people back then could barely get enough food too eat let alone have kids at school.

        1. LybsterRoy Silver badge

          Re: "radically alter the workflow of medical professionals, without their input"

          But they did manage to have kids <G>

          1. CowHorseFrog Silver badge

            Re: "radically alter the workflow of medical professionals, without their input"

            Cockeroaches also have kids and they cant read or write.

        2. Anonymous Coward
          Anonymous Coward

          Re: "radically alter the workflow of medical professionals, without their input"

          Who wrote Genesis then?

          (and if anyone says "Phil Collins" they shalt get smited)

          1. Toni the terrible

            Re: "radically alter the workflow of medical professionals, without their input"

            It wasnt written down until millenium after they were informed about it and they had got as far as cuniform

          2. alisonken1

            Re: "radically alter the workflow of medical professionals, without their input"

            Who wrote Genesis then?

            Unfortunately, Genesis (the band) was already active with several albums before Phil moved from drummer to lead.

            The sorry part, a lot of people won't remember that part.

          3. Norman Nescio

            Re: "radically alter the workflow of medical professionals, without their input"

            Who wrote Genesis then?l

            Jewish tradition has it that Moses wrote down the Pentateuch, of which Genesis is the first book.

            Current secular analysis tends to the idea that Genesis was written down in about 500 BC, but that the content came from several earlier sources, some of which date back to about 1000 BC. In addition, Genesis displays evidence (in textual style) to have been constructed from orally-transmitted sources, so the content pre-dates the written-down form. The text is consistent with the idea of there being no single author, but of at least two different groups at different times collecting and editing sources to generate the version we are familiar with today.

            So if you believe the creation myth, it is entirely possible that Adam and Eve told the story to their descendants, who re-transmitted it to their descendants until people decided to write it down.

            Secular non-believers can trace elements of Genesis that are common with Mesopotamian/Sumerian creation myths and other stories, and one can argue that Genesis is a retelling of older stories with editing to reflect local (Jewish) sensibilities; or one can argue that Genesis uses the literary form or style of the Mesopotamian/Sumerian stories while having independent content. I'm not an expert, and experts in this area fight like cats in a bag, so finding the Truth or the truth of the might might be impossible.

            1. Alan Brown Silver badge

              Re: "radically alter the workflow of medical professionals, without their input"

              "trace elements of Genesis that are common with Mesopotamian/Sumerian creation myths"

              Erm, like large chunks lifted wholesale from the Gilgamesh Epics (themselves having been written down about 1000 years earlier but dating back at least another 1500 years) and tweaked only slightly from the original

              You can see the garden of Eden/apple/serpent story, Flood, Exodus and others in the early poems

    6. Boris the Cockroach Silver badge
      Thumb Up

      Re: "radically alter the workflow of medical professionals, without their input"

      Quote:

      " You do not design an interface without spending some time with the people who are supposed to use it and learning about how they work."

      If I could upvote this 200 times, I damn well would.

      But then this was jammed into my head by university while we were learning howto do application dev work (waterfall/iterive design process)

      See the end users.

      Ask what they want.

      Build interfaces for them to try and comment on.

      Then take that away and build the UI (flashy/handy stuff added at this stage)

      Take it back to the endusers and get more comments.

      Hooray... we have a custom comms application UI consisting of press button(1 of 7) for where the file is going, a file selector dialog, and a big button marked 'transmit'

    7. Benegesserict Cumbersomberbatch Silver badge

      Re: "radically alter the workflow of medical professionals, without their input"

      The human interface problem of medical record keeping is essentially solved, decades ago by the hundreds of people that hospitals used to employ in their medical records department. The basic idea has been converged upon out of painful experience.

      I switch back and forth between workplaces between a full paper system and a digital system. (Yes, full paper systems do still exist, sometimes across entire networks or, in my case an whole state public health system.) The digital systems that actually work are either paper forms that are scanned, or skeuomorphic digital versions of what the paper forms used to look like. Say what you like about an interface having to look like Word, but there's nothing so intuitive as a flat piece of paper and a pen.

      There's a form and a MR form ID for everything. The important things (patient alerts, allergies, demographic identifiers) have low numbers. The forms for outpatient contacts come next, then inpatient events. The sequence of each follows the patient through admission, progress, procedures and discharge, with form numbers increasing as you go along. Sort by form type number, then chronologically. When advance care directives were introduced, they became an MR0, because they had priority over MR1, the alerts page.

      A useful medical note is a mind-map, narrative where something's being narrated, bullet point or enumerated lists or sketch drawings otherwise (past medical history, family history, medications, allergies; history, examination, investigations; assessment and plan). Because it comes out of a human conversation, it's not dictated but constructed in a nonlinear way that's really hard to replicate without the freedom of pen and paper. I haven't seen a capable digital system that isn't awkward, and therefore less efficient that the paper one.

      There's a fundamental inefficiency in shifting work away from the medical information department and on to the heads of the medical and nursing staff, who have enough to do already,

    8. An_Old_Dog Silver badge

      Re: "radically alter the workflow of medical professionals, without their input"

      if it doesn't work like Word, you're up the proverbial creek without a paddle.

      Geh. Just because a thing is common, does not mean that it is necessarily good.

      WordPerfect (version 4.2 for DOS). PC-Write (for DOS). QEdit (aka The SemWare Editor). All of these were great programs, and all of them had a different UI than did MS Word. All three of them made heavy use of function keys.

      1. LybsterRoy Silver badge

        Re: "radically alter the workflow of medical professionals, without their input"

        I still love WordPerfect (currently have X5 installed) but regrettably Word won

        1. Toni the terrible

          Re: "radically alter the workflow of medical professionals, without their input"

          I recently deletd Microsoft Office in favour of Ashampoo Office - cos it is simpler and cleaner (Timotei!)

          1. teebie

            Re: "radically alter the workflow of medical professionals, without their input"

            "I recently deletd Microsoft Office in favour of Ashampoo Office - cos it is simpler and cleaner (Timotei!)"

            You went from a real poo to a sham poo?

      2. Alan Brown Silver badge

        Re: "radically alter the workflow of medical professionals, without their input"

        Wordstar predates all of those and its keybindings are STILL in common use today, nearly 50 years later

    9. Anonymous Coward
      Anonymous Coward

      Re: "radically alter the workflow of medical professionals, without their input"

      There's a pretty good article about the medical record system that was developed at the VA:

      https://www.politico.com/agenda/story/2017/03/vista-computer-history-va-conspiracy-000367/

      Doctors consistently liked this system more than the alternatives, and one key point was the constant interaction between the developers and users.

      1. Norman Nescio

        Re: "radically alter the workflow of medical professionals, without their input"

        Wikipedia has an article on VistA, which contains this paragraph, which tells you what happened after the politico article was published.:

        Wikipedia:VistA

        In May, 2018, the VA awarded a contract to modernize VistA by implementing a commercial EHR. The projected completion for implementing the commercial EHR was by 2028. By March 2023 - half way through the program - only 5 the total of 150 VA medical centers (3%) had piloted the new system. Numerous reports of safety and reliability had emerged at the commercial EHR sites, and four veterans had suffered premature death. As a result, in April 2023 the House Veterans Affairs Committee for Health IT issued a bill to terminate the commercial EHR contract.

    10. ColinPa Silver badge

      Re: "radically alter the workflow of medical professionals, without their input"

      My sister was a nurse. She said the "new IT team" came round to see how people used the system. Because the wards were very busy, the new IT teams had more time with the back office team.

      The result was the system was great for the back office team - but useless for the ward staff. As my sister said having the patients address on the front screen was fine for the back office - but useless on the wards. She had to page down 3 (slow) pages to get to the "on the ward" page. 10 seconds for every access to a patient's details soon adds up.

      NEW EMAIL SYSTEM

      She said they were also going to change the email system to a different system, with different format email addresses. When they asked how they converted their address books, they were told you contact all the people and ask for their details... then you update your address book with the new info "overnight". My sister said I have nearly 1000 contacts in my address book. I use 200 regularly. The management decided to put the project on hold - then cancelled it.

      1. Ken Hagan Gold badge

        Re: "radically alter the workflow of medical professionals, without their input"

        "Because the wards were very busy, the new IT teams had more time with the back office team."

        I think you've hit an important nail squarely on the head.

        The people that programmers can most easily speak to and spend time with are (approximately in order): each other, their friends in IT admin, their own company's sales force, the customer's management, light users of the eventual system (your back-office team) and lastly the heavy users of the eventual system.

        My guess is that unless you make a conscious effort to address this bias, you'll end up with a system that reflects it.

  5. ectel

    Sometimes the inverse works

    I work in NHS IT looking after a system that is written by a Clinician (from a different trust) . It seems to have hit a sweet spot, works really well for the users, adaptions that are needed are quickly understood and implemented as all of us have worked in type of department it serves. From an IT point of view it runs on Ruby/Postgresql and is open source, the tables are well organised to serve the UI so report writing is easy.

    Sometimes things can be done well

    1. Flocke Kroes Silver badge

      Re: Sometimes the inverse works

      Quiet, you might put a stop to hundreds of millions going to 'digitise the NHS' pork projects that get cancelled and restarted every five years or so.

      1. Adrian 4

        Re: Sometimes the inverse works

        No, you won't.

        Government procedure would be to take that working application and have it reimplemented by their favourite vendor. The result will be completely useless, if it works at all.

        rinse and repeat.

    2. LybsterRoy Silver badge

      Re: Sometimes the inverse works

      A good example of a User Programmer <G>

  6. darklord

    that's ok if its one user or one specific group of users. but what happens when there are multiple branches of users all wanting different things from one system. what youve all described are simple silo databases and useful to that specific person or element.

    When multiples all want different things and IT spend goes out the window cause billing need something in blue or red and clinicians want in green cause they dont like red, everything starts costing as work is rehashed and redone on the fly.

    Medical and specifically the NHS system was destined to fail no matter how much money was thrown at it. as a single system that does appointments, Medical records, Stores, Billing, triage, Xray access, diagnosis and then prescribing and reporting as well as being user friendly to every user it will ever encounter (including the 80 year old receptionist whos been there since 1972 and only works on Tuesday mornings)and god knows what else the NHS do that i don't know about.

    No matter how good you are postgresql or other freebee database wont do that any time soon without massive support programs.

    It would be nice and simple to think thats what can be achieved on the cheap but not to run a hospital

    1. James Anderson

      Said it before but I will say it again.

      What the NHS and medical profession needs is a well designed set of messages and protocols to enable systems to exchange information.

      This would allow specialist applications or even ordinary applications like appointments to be developed independently and replaced or upgraded independently.

      The financial services industries have used this model for decades with great success, think SWIFT, chaps, SEPA, ISO8583, etc. etc.

      1. ectel

        We have HL7 and FRIE https://www.hl7.org/fhir/overview.html

        The devil is in the detail, particularly when you have multiple systems. EG think of a large trust that has 2,3 or more pathology systems and patient administration systems. what identifiers are you going to use? what to do when the Date of Birth is different on one system to another? all sorts of problems and compromises.

        That's where the difficulty and therefore cost is

        1. Elongated Muskrat Silver badge

          I have experience of using HL7 to communicate between different IT systems. Highly specific messages that don't match the use case, so you end up misusing the fields in them to hold information that is subtly different to the message design spec. The other 95% of the fields are left blank, and you have to count the delimiters when reading them to work out whether you have got your data in the right field.

          Eugh. Give me some use-case-specific JSON instead.

        2. James Anderson

          Well take the history of ISO8583.

          It started off as a protocol for ATMs. The main design goal was to minimise the amount of data transmitted over very slow networks. Which was spectacularly successful but resulted in the hardest to parse message format you will ever come across.

          However over the years the standard evolved to encompass credit card transactions, point of sale, web based commerce, chip and pin, NFT, phone purchases. Most of the worlds retail transactions involve messages using this format.

          The point is having a good standards body with the all the right actors involved and a flexible approach makes this stuff work. Even something as godawfull as ISO8583 is still in use 35 years on and is spectacularly successful.

      2. herman Silver badge

        Yup - Divide and conquer.

      3. CowHorseFrog Silver badge

        Banking is quite simple compared to medical records.

        A bank transfer is pretty basic, date time, some account numbers and then the monetry amount. Medical reports are a bit more complicated than that.

      4. LybsterRoy Silver badge
    2. Greybearded old scrote Silver badge

      One view for everybody is rarely an appropriate design.

      1. Alan Brown Silver badge

        I can't upvote this enough

        Furthermore, if users can't change/customise their view AND SET IT AS DEFAULT, the system is already unfit for purpose

    3. Doctor Syntax Silver badge

      "that's ok"

      Which that is that?

    4. An_Old_Dog Silver badge

      The Old(er) Way of Doing Things

      Was that the same "80 year old receptionist whos been there since 1972" who threatened me with a ruler when I came to take away her green-screen IBM 3278 terminal and replace it with a shiny-new IBM PS/2 with all our hospital's standard apps (including a terminal emulator to connect to her beloved mainframe)?

      I did the sensible thing -- I backed off and told my supervisor what had happened. After presumably some higher-level intervention, some other unlucky sod got the job of replacing her terminal with a networked personal computer.

      1. Ochib

        Re: The Old(er) Way of Doing Things

        Normally the way to do this was to wait until they went on holiday and replace it then.

    5. LybsterRoy Silver badge

      I may be misinterpreting but you seem to be imagining one massive program that does everything. If so - yup guaranteed fail. On the other hand several smaller programs that can talk to each other as needed - perfectly doable.

  7. that one in the corner Silver badge

    10 comments in: UI and free databases have been discused

    But so far, no mention of something that we used to call:

    Systems Analysis

    Is this just a completely forgotten art, or is is it being done piecemeal by whoever feels like doing a bit of it today? A bit here, to fill in a piece of UI work, a separate bit over here to decide what to put in another database column.

    And, no, it isn't something that should be done, once, by a separate consultant and then dumped onto the dev team as a badly organised pseudo-Requirements Spec, but has to be open for updates as implementation discovers gaps (just like DB design and UI progress hand in hand).

    Oh gawd, I'm starting to sound like Agile! But without codifying that to the point of uselessness as well!

  8. Anonymous Coward
    Anonymous Coward

    System designers vs interface designers

    I somewhat doubt it's complete arrogance by the system designers, it's usually that the end user isn't the target market. The specification should include that a given sample of end users should be part of system acceptance, especially when it's something as important as this. It also means that the working system isn't replaced until its successor is bedded in and equally functional. Anyone who specifies otherwise is not doing their job.

    System delivered to store and query data at ridiculous expense to enrich large consultancy firm that's bribed (oh, sorry, I mean 'lobbied') to get through tendering process that favours exactly those firms : check

    System not actually completely suitable, as per user requirements, but technically matching specification : check

    Cost to change anything outside implemented specification : oooh, that's painful, what a shock. check.

    It can also be that some firms don't adequately fund UI development : doing it really well is difficult. I'd rate the systems I deal with as adequate, but barely more than that. There has never been the spare time or money to change this. Customers are often not prepared to pay the development time for an outstanding interface, so they get something 'good enough'. That's not to say thought did not go into products, especially the most used parts, and that it wasn't tested, and user feedback considered. However, as time goes on the priority is often matching an updated look and feel, and compatibility, rather than attempting to evolve the interface.

    I was involved in one project where I admit we delivered a barely adequate product (technically it worked, the interface was functional but a real pain). However, this was due to the project being assigned to a rogue developer with insufficient management oversight. When things went badly wrong myself and a colleague got pulled in to rescue it, and rescue it we did, within the constraints we were neither UX designers or usually on development. We did our best, but delivering an all singing and dancing wonderful product would have required months extra development time, when the priority was not losing the company money, rather than truly satisfying the customer. Fortunately this was a once off, as the rogue developer had been swiftly departed, and more oversight occurred afterwards.

    1. Doctor Syntax Silver badge

      Re: System designers vs interface designers

      "The specification should include that a given sample of end users should be part of system acceptance"

      If you don't bring users into the process before acceptance testing then you're doing so far too late for their input to be of use. It may well be so far askew from what they need that either it will have to be scrapped or, more likely., forced on them regardless of whether they accept it.

  9. Bebu
    Headmaster

    Skeuomorphia - sounds serious

    Something incurable surely ;)

    I think the acid test after a manual system has been converted into an electronic skeuomorphic system is whether it can print a faithful replica of the original records.

    When I think about how I locate information, in books I own and have read, or in a filing cabinet, or from under the piles of paper on my (typical BOFH) desk, the physicality of these records seems really important for efficiently locating the records sought. The psychologists might be able to shine a light of why this might be so.

    An recent example was when trying to recall the differences between ipsec's transport and tunnel modes - I knew exactly where in a text to find a coherent and cogent explanation but unfortunately my copy of the text has been in storage since covid-19 so I was reduced to trawling through the semi-illiterate drivel that infests the internet. (The text I was thinking of were the very literate Radia Perlman's "Network Security" or "Interconnections.") The RFCs are rather dry and a bit on the terse side.

    I recently updated my detail at the storage facility - not a computer to be seen. An index card with the mutable details recorded in pencil. A few minutes and rubber eraser later address and contact details updated. The irony is they almost certainly would not get a cybersecurity (or ISO27K) certification.

  10. Mike 137 Silver badge

    "This sort of "design blindness" happens whenever system designers prioritize the needs of the IT system, rather than its users."

    Been there more than once. As an application designer in a specialised niche (experiment automation) I've had to work with developers who think the code is more important than the functionality of the product, and instead of respecting user priorities and advancing the product by including additional necessary features, insist on endlessly tweaking the existing code without providing any discernable benefit to users of the product. On other occasions, developers have hijacked the product design and refused to adhere to the official design spec. In one case, due to both these snags, a potentially useful product has been frozen at beta for several years, and has therefore probably lost any customer base it might have had.

    1. Alan Brown Silver badge

      " I've had to work with developers who think the code is more important than the functionality of the product"

      Plus more than a few developers whose response is "Why would you want to do that?" and refusing to consider it

      That kind of attitude resulted in several crashed Airbusses.

      Pilots want to do "that" to test the systems are working properly when things go pearshaped. Users frequently want to do "that" to ease their workflow or test worst case scenarios

  11. martinusher Silver badge

    This should be straightforward enough

    My medical provider in the US -- Kaiser Permanente -- has been fully digitized for years. The software isn't perfect -- anything that involves Windows and HIPAA has to be klunky almost by definition -- but by and large it not only works but works very well. In fact it works so well that its only when you step outside the envelope to a world of paper and (believe it or not) fax that you really comprehend just how good it is.

    The patient / member portal is the first point of contact. Here interactions with doctors and specialists use secure email, appointments are made as necessary, tests ordered and managed and prescriptions ordered and refilled. There's also opportunity to consult online with a doctor either by phone or video, the latter being particularly useful (when it works -- caveat is "the software" as usual)(its not Linux friendly but give it an Apple and its as happy as a pig in slop)(Windows......meh.......). The information is available to all providers no matter where their offices are located -- all offices and consulting rooms have systems in them (can't really use wireless/tablets because of security/confidentially) with individual logins handled by RFID from the worker's ID card.

    You have to experience this to believe that it can actually work. (...and don't confuse this with the billing systems endemic to US medical providers; being an HMO KP works more like a NHS than a traditional fee-for-service operation)

    PS -- As a bonus KP is also metric although it doesn't talk about this much to the general public, I suppose to avoid spooking them (this being the US and all).

  12. Anonymous Coward
    Anonymous Coward

    Grumbling oldie

    Going back a few decades, one of the most successful things I worked on, well we had analyst/programmers and we had expert users. We didn't have business analysts or product delivery managers. People spoke to each other all the time - not at some daily stand-up. We had expert users sitting with the dev team ready to answer questions, give opinions on UIs and all that. We had a QA guy who was a total pain in the arse in that he would run code coverage analysis over his regression tests and be really picky about being able to make sure every build was sound.

    Nowadays, I find myself in meetings where one person knows about tech, one person knows about the business problem and the other ten are just using up time and oxygen.

  13. Pete Sdev Bronze badge
    Boffin

    To produce good software, developers need to talk (and listen! ) to the intended end-users.

    In practice this seldom happens, which is why so much software sucks.

    In my work, the best I can get is my manager talks to the contacts at the client company. These contacts are also usually managers or similar and not the end users, or at least not the primary ones.

    1. veti Silver badge

      Trouble is, the "intended end users" aren't particularly interested in talking to you. They've got jobs to do. The only time that's not likely to be the case is when someone, probably their boss, has already committed to spending a crapton of money on you.

      Which puts the focus squarely back on sales, which is probably where the mischief starts anyway.

      1. martinusher Silver badge

        In my last job hiding contacts and information kept marketing in a job. They could be guaranteed to screw just about anything up because they just didn't have the depth of technical knowledge to work with complex systems. Its not that marketing people can't do this -- when you work with someone who doesn't fancy themselves as a know it all its like striking gold -- but the norm is to issue directives and maintain a stranglehold on information.

        Eventually natural selection deals with the problem. Engineers who won't suffer this regime move to companies that offer them better prospects so the wall of mediocrity eventually sets the tone for the company's product line.

  14. John Miles

    Handwritten v typed

    A couple of years ago there was a study suggesting students handwriting notes had better memory retention than those who typed notes - I wonder if we would see similar for the doctors and they are more likely remember more about a patient when they hand write it on paper than type it into a computer

    1. NeilPost

      Re: Handwritten v typed

      I doubt it - far too many patients, presenting with many different conditions and short consults. Without looking at the Patient Administration System (PAS) .. I doubt your Doctor even recalls your name.

      A student is only doing their own thing - hence perhaps a glimmer of retention. The typed notes are the very least key-word searchable and when scaled up into the real world by many. Think Confluence, a Wiki. Microsoft KB or Unix/Linux ‘man’ v’s a printed technical reference manual.

      1. ectel

        Re: Handwritten v typed

        I knew a Chest Consultant that didn't recognise the patients face, but did recognise them from their chest x-ray

        1. Toni the terrible

          Re: Handwritten v typed

          Recognising a patient by an x-ray or their condition is not at all unusual

        2. Richard Pennington 1

          Re: Handwritten v typed

          Could be worse. In the 2011 film "What's Your Number?", the protagonist is surprised by how her obstetrician recognises her.

      2. John Miles

        Re: Handwritten v typed

        For a patient like myself who hasn't been to the Dr for 5 years, yes however those who have to go more frequently I'm not so sure and I suspect they may remember more about the problems than the name purely because what they talk/write about

  15. nightflier

    Completely portable laptop-slash-tablet

    I sure hope that little device device is locked down with a strong password and encrypted drive, and that it is never left unattended when unlocked. HIPAA is picky about protecting patient information.

    1. AVR Bronze badge

      Re: Completely portable laptop-slash-tablet

      I think the writer's an Aussie. Whatever legislation they have over there it's not HIPAA.

      1. nightflier

        Re: Completely portable laptop-slash-tablet

        Fair point. I plead guilty to US-centrism. Still, if it was my personal data on that easy to lose or steal device, I'd want it to be safeguarded.

  16. Fruit and Nutcase Silver badge
    Facepalm

    Sounds like

    An MBA driven design than one driven by what the MBBS wants

  17. IceC0ld

    those of a certain age, and UK based, may well recall an NHS push to get ALL medical practices attached to ONE single piece of software, so that the GP could, for ANY patient, arrange the date and time of any operation, and included the surgeon of preference - the name of this :- CHOOSE AND BOOK :o)

    ANYHOW

    I was involved in the install phase at various GP surgeries around the UK, and at most, I was treated to a mostly unpleasant shouting at game from all the doctors, and at the hospitals, even the bloody surgeons decided to get stuck in too

    it appears that although the general outline of the app was seen as a good thing, the coders involved had missed one or two, may seven various DB's they could of / SHOULD have linked to

    for example, as one surgeon was rather keen to emphasise, they have added me to this bloody operation, although my schedule shows I am ALREADY set to operate on someone ELSE, at a different hospital on that time .......

    and one other pointed out that the bloody app had him set to operate on several people, when he had his holidays set to go ...............

    in short, great idea, VERY poorly planned

    and if that wasn't enough, this was when MOORE'S law kicked in big time, we could not possibly install the entire UK that fast, and so both the expectations of the app developers grew, as did the capability of the PC

    so we NOW had to RE VISIT surgeries to upgrade the Choose and bloody Book install, instead of increasing the number of surgeries on the go

    as far as I am aware it may still be going on :o)

    for ME, it was a year of reasonably fun times, a good bunch of guys, a VERY understanding Co, but the scale and scope of it all had been VASTLY under estimated :o(

    sound familiar :o)

    1. Alan Brown Silver badge

      speaking of software

      My GP has just switched to the THIRD vendor of front-facing customer software

      It's still using google-analytics and various other 3rd party javascripty things that send data overseas, without disclosing it

      THEY may not be leaking data, but that's no guarantee whatsoever that 3rd parties is not - and we know that Google and FaceAche both aggressively harvest whatever they can get their grubby mitts on

  18. raving angry loony

    Idiots in charge

    Sadly, it's rarely (if ever) the people who actually will have to use the interface who are consulted. It's their managers. Or even their directors. People who don't have a CLUE what the actual job is and how it might work.

    So many industries, so much bad design, and so few people giving a fuck because hey, it's cheaper to do it wrong, so it can't be wrong!

    Idiots. The lot of them.

  19. Prichy

    In my experience marketing can also screw it all up

    I used to work in retail for a large chain of convenience stores. When I joined, many years ago, they still had non-GUI character terminals and keyboards. The keyboards were a custom design with some dedicated buttons for very frequent actions (.e.g Total, select payment type, open cash drawer, void item etc) and a numeric keypad to enter the six digit item code for anything that you couldn't scan with the barcode scanner. After only a couple of weeks a new staff member was incredibly quick on these things, knew the most-used codes off by heart (coffee, fresh items etc) and could conduct a conversation eye-to-eye with the customer without ever looking down at the keyboard. Then a new era arrived: a brand new system with touch-screen graphical terminals. Marketing went crazy! Great, they thought, now we can push the current special offers onto the "front-screen" with lots of other specials and other functions buried on sub-screens via menus. They would change the layout on a whim, as they implemented each new dazzling offer and pushed that front-and-centre on the terminal, completely re-arranging the GUI. It was a disaster of course; now all the 'buttons' changed weekly and the poor cashiers had to look to see which menu etc they were in. They spent ages searching for the items they needed and even when they found them, the position would change the next week. Moreover the touch screen wasn't anywhere near as responsive as it needed to be and of course it didn't buffer keystrokes like a keyboard does, requiring constant vigilance to ensure every 'button' press worked. It was never replaced of course, because my company had invested a fortune in these things and the incompetent management consultants (beginning with "A" and ending with "ccenture") who had led the project and, as usual, insisted the long queues now forming at every store were the fault of our staff and not the appalling system they'd implemented.

  20. Evil Scot Bronze badge
    Pirate

    This is Comp Sci 101 FFS

    We were literally taught this in the first months of my degree course. "Why do projects fail?"

    College tutors who taught this also branched into "How to milk Government Contracts." Bid low, they have no idea what they want and milk them for every change.

    Actual Uni tutors on this course had as much interest in teaching as their Discworld counterparts.

  21. Anonymous Coward
    Anonymous Coward

    In a previous life in a bank, I had the pleasure of looking after a system that all post related to a specific customer account would be scanned in and logged against that customer's record.

    It made it very easy for anyone that needed the permissions to look back and forward through that file to see the full history, without reference to towering paper archives that would be otherwise impossible on that scale.

    That was well over 20 years ago, and so getting these things wrong now... Well, that's a startling level of incompetence.

  22. Mike Hayes

    VV&T

    Oh how often do we see it. The management want the stats yet the users aren't considered. We used to have VV&T - Verification Validation and Testing. Or to put it another way, Are we building the right thing, are we building it right. Getting the test community involved at the design stage. Issues found before a line of ode is written pays so many dividends.

    1. Alan Brown Silver badge

      Re: VV&T

      "are we building it right" - should include "are we leaking personal data to 3rd parties?"

      EVERY piece of 3rd party javascript is a risk and remains a risk even after being audited, because it may change without notice (as with android APPs being sold off to malware authors, this happens at times to js occasionally)

  23. J.G.Harston Silver badge

    Exactly! EMIS was written by two doctors, from the point of view of the needs of doctors, for doctors. I don't have the history, but I believe SystmOne has a similar history. Other systems have been clearly written by programmers for programmers.

  24. ChaosFreak

    EHRs are billing systems

    Sadly, this is another case of "the user is not the customer." Why does Amazon and Google search suck? Because the user isn't the customer, the advertisers are. Google and Amazon search don't return the best results, they return the results that maximize sales for advertisers and ad revenue for Google and Amazon.

    Why do EHRs suck for doctors? Because they're not the customer, the accounts are. EHRs are designed primarily to maximize revenue and reimbursements, and only secondarily to serve doctors and patients. At least in the USA....

  25. Richard Pennington 1

    Going through the records

    Just before COVID, I asked my doctor a query which required a search way back in the records. I asked whether I was eligible to give blood.

    The issue was that, as a child, I had had hepatitis. Some forms of hepatitis would disqualify me as a blood donor. So I gave my GP as much detail as I could remember, and she took it as a challenge.

    And sure enough, she found it: a brief entry: "Infectious hepatitis", in December 1965. It was before they defined Hepatitis A, B, C, D or E. They would probably call it Hepatitis A these days (certainly the theory at the time was that it came from a stream which ran through the grounds of my infant/junior schools; Hepatitis A is usually water-borne). And Hepatitis A is not a disqualifying factor, especially with over 50 years of time passed since the infection.

  26. Anonymous Coward
    Anonymous Coward

    Digital vs paper processes

    The number of times I've heard,

    Dr : "IT is making my life harder"

    IT : "Sorry about that, what's the problem? "

    Dr : "you should have got a Clinician to help you design this".

    IT : "Ah yes we did, and told them their design was not good but they forced it through anyway"

  27. FatGerman

    Was it just me?

    Did I imagine this or did there used to be a mantra, even a rule, that the job of the UI is to convert computer-speak into human-speak? Even if it means having an entire translation layer so your old database can be mangled into something people understand. Even if it means the backend developers mouth-vomit at the sight of what you're doing to their beautifully organised schemas. I've seen it so many times - UIs with one checkbox for every single flag the backend knows about, with names that mean everything to the backend devs and nothing to the user, UIs where the controls the users need most are in 7 point light grey text on a white background and you have to scroll to get to them.

    In one of my earliest projects I was asked to give the team a "Big Red Button" that they could hit and then go home for the weekend. And that's essentially what I think all UIs should be.

  28. Tron Silver badge

    Good luck having a patient record.

    For digitisation, emulate what is being used and add easy to learn/access extras that users ask for. Ignore all proprietary file formats and what nots, and stick with the most generic stuff you can. Digitising the handwriting of the medical fraternity requires expertise. OCR or cheap labour abroad is next to useless.

    I used to have a GP who I'd known for years, knew everything that was wrong with me, and noted it all down in a folder.

    A couple of years ago I'd see a different locum GP each visit. They would ask me what was wrong and only used their PC for ordering referrals and printing prescriptions.

    Now I stand no chance of seeing a GP. For some things, the receptionist will tell me that it sounds serious and I should go to A&E. For others, the receptionist will tell me to dial 111. If 111 say I need to see a GP and I try again, the receptionist will tell me that they are too busy, and I should go to the urgent care clinic at the local hospital.

    I'm surprised that I might even have a patient record somewhere. If Google are buying access to these, they may be in for a shock.

    The government repatriated much of the NHS for Covid and kicked out as many more as they could over Brexit. I think the NHS is dying on its feet - GPs going the way of dentists. The Tories have finally managed to get rid of it. We may need to rely on OTC painkillers and YouTube from here on in.

  29. Nonymous Crowd Nerd

    "Well, I don't even know where it is," he lamented.

    This says it all about so many new systems.

    And it's not just problems with poorly designed User Interface. It's not just someone who - with good intentions but lacking skill and experience - screws up the design...

    There can't be any doubt that there are those whose actually intention is to force you to follow their preferred path through any given process and prevent people from working in any other way.. They sometimes deliberately work to hide the data structure - or to prevent us from performing free searches.

    These people want to curate our entire experience - it's like being forced to follow the preferred path through IKEA without skipping sections or going directly to the department you came for!

  30. JimC

    All these comments are interesting in the light of

    The often heard mantra that business processes should be adapted to fit off the shelf packages rather than vice versa.

  31. s. pam
    FAIL

    How about something more simple?

    NHS trusts in England do NOT share any patient data between counties.

    Case in point: I was in hospital in Berkshire and I reside in Buckinghamshire. When I was discharged, the Berkshire hospital gave me a PRINTED copy of my records with verbal instructions "Take this to your GP surgery in Buckinghamshire so they can, wait for it, -- SCAN IT IN!!

    In the name of all things digital/holy, how the fsck can we be in the 2020s and this simple bit of "push a button, send to the patients' GP surgery be so damn hard? We're not talking about sending data to Scotland, Whales or NI -- we're talking about INSIDE England.

    FFS, you just cannot make this shit up.

  32. Muscleguy
    Boffin

    Its possible

    My now ex wife works in a university. They got a student records package system put in. It took quite a bit of interface design and with a BA in Maths and a BSc in CompSci she pitched in. She even did some coding for them to ensure stuff worked properly. Why it pays to have a computer savvy person around for these things.

    I used to be a biomedical researcher. I have built all manner of databases in my time. So when it came to my last science job where they were getting a desktop robot to aliquot samples on plates for analysis and sensibly wanted to barcode it I built them a relational FileMaker DB which read and printed barcode labels. The techs were taking in bags of blood and processing them down to DNA and RNA in tubes.

    I used the robot myself as it was more accurate at gridding than anything else. So it gridded my real time RT-PCR plates for me. It had a nice macro type interface making routine building easy even for non specialists.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like