4 ms·

It's because shortening doctors' shifts does not result in better patient outcomes: https://psnet.ahrq.gov/primers/primer/19/Duty-Hours-and-Patient-Safety http

It's because shortening doctors' shifts does not result in better patient outcomes: https://psnet.ahrq.gov/primers/primer/19/Duty-Hours-and-Patient-Safety https://psnet.ahrq.gov/primers/primer/19/Duty-Hours-and-Pati... The reduced fatigue is offset by the increased number of patient handoffs between shifts. More handoffs = more errors. My wife is a resident and she supports 30 hour shifts for residents for this reason. The difference between your judgement at hour 23 and hour 29 is less of a risk factor than a 20% increase in handoffs.

taeric 8y ago
This seems to imply that we can't improve hand off procedures. That is, what about the hand off is so risky? Are we fixing that?
- rscho 8y ago
  No, we're not. Clinical software is the 9th circle of hell for every healthcare worker, as it was made not for patients, but for billing efficiency. As a result, handoffs get no benefits from tech. Computers make it even worse by increasing administrative loads.
- bookofjoe 8y ago
  Retired neurosurgical anesthesiologist here (38 years experience in top-tier academic centers: UCLA/USC/UVA). It was always my practice — from my first year of residency in 1977 until I retired as an attending faculty anesthesiologist in 2015 — to refuse relief at 3-4 pm when the night crew came on duty IF I had a patient who'd been unstable or whose management was especially complex. I didn't believe I could fully read the relieving anesthesiologist into the case because so much of what I'd been through with the patient had been idiosyncratic and it had taken me hours to get a "feel" for what worked and what their safe parameters were. Relief teams loved me, you can be sure. Oh, one more thing: when I did hand over a relatively straightforward case, rather than quickly dress and get the heck out of Dodge, I'd go down to the cafeteria, get a snack, come back to the anesthesia ready room, shmooze a bit, and then go back to the OR I'd left and quietly enter, saying to the relief doc, "Did I leave anything out, or is there anything on the anesthesia record that isn't clear?" Quite often there WAS a question or three whose answer(s) helped orient the relief anesthesiologist.
  - nradov 8y ago
    Thank you for going the extra mile for your patients. I'm sure you saved lives that way. Too many hackers make the mistake of thinking that human bodies are deterministic machines that respond consistently to inputs just like computers. The reality of medical care is far more complex.
  - taeric 8y ago
    I'm assuming "quite often" was not 90% of the time. To that end, was there a common theme to the times this happened? Not necessarily automatic automation obvious, but some learning. And, I do want to echo the sibling post. Just because I am asking questions like this does not mean that I am not thankful for the job you do. Thanks!
    - bookofjoe 8y ago
      No common themes: a million and one different loose ends, many that never came up before and never would again, having to do with inability to read lab values written in tiny spaces, shorthand to myself not known by anyone else, etc.
      - taeric 8y ago
        This feels unlikely. I get that there would be a lot. But the number is almost certainly smaller than you are giving credit. Just with the two you gave, automating the recording of lab results should help. Such that you never have to copy from a readout to a record. That should just happen. (Consider how many shipments and other things are transferred with minimal errors every day.) Similarly, why have a shorthand that is known only to you? We have plenty of things like that in computers. Not everyone has my macros. That said, they expand to known things. Or contract from them. So, I would hope we have folks working to help in this area. I'm not convinced we don't. I do think it gets presented adversarially all too often. There is no need for that.
- boldslogan 8y ago
  This sounds like a start up idea? Doctor notes app to synch with other doctors?
  - rscho 8y ago
    For multiple reasons, it is unfortunately a very poor basis for a startup. No one would ever buy that.
  - taeric 8y ago
    I would guess you need more than just a notes syncing app. Triaging will likely depend on sensors as it does handover of notes. That said, I meant the question genuinely. I suspect that we aren't just ignoring the problems.
RHSeeger 8y ago
> The difference between your judgement at hour 23 and hour 29 is less of a risk factor than a 20% increase in handoffs. What about the difference in judgement at hour 8 vs hour 29? Honestly, the fact that there's not much benefit for working "way too many hours" vs "way too many hours + 6" doesn't surprise me. It's the difference between "a reasonable number of hours" and "way too many hours" that should have more of an impact. And how much would handoff issues be reduced if everyone was working an 8 hour shift?
- rscho 8y ago
  Do you have an idea of how much 8-hour shifts would cost? It is not pretty. It is doable, but has numerous drawbacks as well. The first being that you would not want to have surgery done by someone trained on 8-hour shifts without a complete rehaul of medical education that would also cost you an arm and a leg.

taeric 8y ago

This seems to imply that we can't improve hand off procedures. That is, what about the hand off is so risky? Are we fixing that?

rscho 8y ago
No, we're not. Clinical software is the 9th circle of hell for every healthcare worker, as it was made not for patients, but for billing efficiency. As a result, handoffs get no benefits from tech. Computers make it even worse by increasing administrative loads.
bookofjoe 8y ago
Retired neurosurgical anesthesiologist here (38 years experience in top-tier academic centers: UCLA/USC/UVA). It was always my practice — from my first year of residency in 1977 until I retired as an attending faculty anesthesiologist in 2015 — to refuse relief at 3-4 pm when the night crew came on duty IF I had a patient who'd been unstable or whose management was especially complex. I didn't believe I could fully read the relieving anesthesiologist into the case because so much of what I'd been through with the patient had been idiosyncratic and it had taken me hours to get a "feel" for what worked and what their safe parameters were. Relief teams loved me, you can be sure. Oh, one more thing: when I did hand over a relatively straightforward case, rather than quickly dress and get the heck out of Dodge, I'd go down to the cafeteria, get a snack, come back to the anesthesia ready room, shmooze a bit, and then go back to the OR I'd left and quietly enter, saying to the relief doc, "Did I leave anything out, or is there anything on the anesthesia record that isn't clear?" Quite often there WAS a question or three whose answer(s) helped orient the relief anesthesiologist.
- nradov 8y ago
  Thank you for going the extra mile for your patients. I'm sure you saved lives that way. Too many hackers make the mistake of thinking that human bodies are deterministic machines that respond consistently to inputs just like computers. The reality of medical care is far more complex.
- taeric 8y ago
  I'm assuming "quite often" was not 90% of the time. To that end, was there a common theme to the times this happened? Not necessarily automatic automation obvious, but some learning. And, I do want to echo the sibling post. Just because I am asking questions like this does not mean that I am not thankful for the job you do. Thanks!
  - bookofjoe 8y ago
    No common themes: a million and one different loose ends, many that never came up before and never would again, having to do with inability to read lab values written in tiny spaces, shorthand to myself not known by anyone else, etc.
    - taeric 8y ago
      This feels unlikely. I get that there would be a lot. But the number is almost certainly smaller than you are giving credit. Just with the two you gave, automating the recording of lab results should help. Such that you never have to copy from a readout to a record. That should just happen. (Consider how many shipments and other things are transferred with minimal errors every day.) Similarly, why have a shorthand that is known only to you? We have plenty of things like that in computers. Not everyone has my macros. That said, they expand to known things. Or contract from them. So, I would hope we have folks working to help in this area. I'm not convinced we don't. I do think it gets presented adversarially all too often. There is no need for that.
boldslogan 8y ago
This sounds like a start up idea? Doctor notes app to synch with other doctors?
- rscho 8y ago
  For multiple reasons, it is unfortunately a very poor basis for a startup. No one would ever buy that.
- taeric 8y ago
  I would guess you need more than just a notes syncing app. Triaging will likely depend on sensors as it does handover of notes. That said, I meant the question genuinely. I suspect that we aren't just ignoring the problems.

RHSeeger 8y ago

> The difference between your judgement at hour 23 and hour 29 is less of a risk factor than a 20% increase in handoffs. What about the difference in judgement at hour 8 vs hour 29? Honestly, the fact that there's not much benefit for working "way too many hours" vs "way too many hours + 6" doesn't surprise me. It's the difference between "a reasonable number of hours" and "way too many hours" that should have more of an impact. And how much would handoff issues be reduced if everyone was working an 8 hour shift?

rscho 8y ago
Do you have an idea of how much 8-hour shifts would cost? It is not pretty. It is doable, but has numerous drawbacks as well. The first being that you would not want to have surgery done by someone trained on 8-hour shifts without a complete rehaul of medical education that would also cost you an arm and a leg.