19 February 2008

Diurnal variation

I'm always fascinated by the changes that occur in the course of the daily rhythm. It comes, I suppose, from working in an industry that is staffed 24/7. The daily life of an ER is predictable and cyclical, illustrated above in an old graph (2004) of daily registrations at one of our ERs. 7AM: the place is pretty quiet; clean up a bit and roust out the drunks. 10AM: the ambulances start rolling in. 2PM: getting overwhelmed. 8PM: fully overwhelmed Midnight: starting to dig out. 4AM: dead.

One thing I had never thought to look at is the likelihood of surviving cardiac arrest, by time of day. Apparently, according to a study released today in JAMA, patients who have an arrest are more likely to die if it occurs on the night shift or weekend. That makes sense, I suppose. Since there are fewer scheduled admissions during those hours, the staffing levels are lower; the experience levels of the nursing staff tends to be lower, since more senior nurses tend to grab the day shifts; and the normal human functional nadir occurs around 0300. So the likelihood of noticing a patient going into arrest is slightly lower, the delay in calling the code is longer, and the time required to assemble a full team is longer. I haven't read the full study, but on quick review it seems very reasonable to me.

But that's only true for inpatients. Apparently, in the ER, your survival rate is the same regardless of time of day. That probably has to do with the ER's ability to rapidly respond to, well, emergencies. The entire team is there, including the doctor, there may be some advance warning from EMS, and while we too are not at our best at 0300, the workload is lower and we are able to quickly jump on emerging situations.

Interesting study. I'll have to read it later, but for now I'm off to my own night shift.


  1. That busy time? From 10 am to midnight? That's pretty much my shift. I come in at 11:30 am, start running, and don't stop til I leave at 12 midnight.

    Ow, my feet!

    The only good thing about it is I get to sleep while it's dark out and I never have to get up at 5:30 am ever again.

  2. I had a friend who did root cause analysis for nuclear power plants. He said it is remarkable how many major problems occur at the mid-shift low around 2 a.m. And, even more interesting, how many minor problems happen and the similar midday low around 2 p.m.

    It is just interesting how predictable human beings are.

  3. WRT the Jama study, the differences were trivial in all comparisons; a negative study overall.


Note: Only a member of this blog may post a comment.