09 January 2006

May I take your order?

For some time, long ago, I worked as a waiter. It was a fun and challenging job and I was pleased that I got pretty good at it. At the time, I was vaguely embarrassed that I took such pride in being good at a generally menial job. But there was a real learning curve, and not every person who worked there seemed to "get it." The difficult thing to learn was not the menu or anything like that, but the mental organization to keep a running, prioritized "to-do" list in your head. e.g., "Table 31 needs salt, 32 needs to be greeted, 33's order should be up any minute and I have to go to the kitchen and look for it, and 34 is going to want their check soon and I need to go ask if they want dessert." It sounds pretty simple, but when you consider that items are constantly being added to or dropped from the list, and they carry different priorities, and some are quick and easy while some tasks are more involved, and some need to be done immediately, while others are less urgent, you can see that it's a pretty dynamic process requiring a sophisticated decision-making algorithm.

I forgot about this for a while, until I started working in the ER. Then I went through the same learning curve, as I again learned to manage my live to-do lists. It did not take long for me to recognize the similarity between my new and former jobs. I was now, I joked, a "medical waiter."

What I am interested in is the heuristic model with which I (and other successful waiters/docs) keep our time effectively managed. To formalize the model a little: view each patient as a project, for which my goal is to move them to discharge or admission in order to free up the bed for the next patient languishing in the waiting room. The enviroment is such that time is the most valuable commodity. I am always managing multiple projects in parallel, with no commonality between projects. For each patient I can pretty quickly define the things I will need to do to accomplish this disposition: initial greeting; data collection; therapy; communication and disposition.

In order to do this, I create and continually update a multivariable ranking “to-do” list. The most critical variables for each task seem to be:

1. Clinical Urgency: how sick is the patient and how urgently do they need intervention? This variable, fortunately, comes into play only a few times in the course of a given shift, even in a busy ED, and, of course will propel the affected item to the top of the to-do list.
2. Benefit/Importance: How much will accomplishing this task move this project forward? Conversely, what are the consequences if I do not get this item done immediately? This has more to do with how long they have been waiting, how pissed off they are, and how much not doing the task will delay disposition
3. Time consumed: How long will this take to get done?

So a general ranking order might be:

Highly Clinically Urgent items
Highly beneficial, quick items
Moderately Clinically urgent items
Highly beneficial, time-consuming items
Poorly beneficial, quick items
Poorly beneficial, time-consuming items

Or, if we give each category a symbolic value:
Urgent (H-M-L) (High, Med, Low)
Benefit (H-L)
Time (Q-S) (Quick, Slow)


This will let me rank the order in which I intend to do my tasks. Of course, the value of a given task will change with time – patients’ clinical conditions evolve, the low-benefit stull become more important the longer the patient waits (e.g. putting in a couple of stitches may be the least important thing you need to do for the trauma patient, but you can never discharge them till it’s done, and as all the other tasks are completed on that patient, the stitches become the only thing keeping them there), etc. I generally am ready to pick up a new patient when all my to-dos are U(L)B(L)T(Q) or better. The worst thing is when I get pulled out of the work cycle for a long time (doing a procedure or tending to someone really sick) and everything is at the same rank because it’s been neglected for so long.

It’s really amazing, when you think about it, that we all do this more or less unconsciously, without a lot of formal thought. Some are better than others – I’ve seen a lot of folks who are very slow and carry a small patient load because they couldn’t effectively organize their workload. I like to think I am pretty good at it.

Of course, this whole discussion neglects the external interruptions which plague every ED doc through the day. One study suggested that the average ED doc gets interrupted 20 times ah hour.

It’s no wonder I have ADHD!


  1. This is why I could never be a waiter (or ER doc apparently). I have a hard time juggling projects without a serious time crunch.

  2. You haven't posted in awhile, but the quality is high. I have a hard time sitting down and doing really meaty posts like this.

    This is a gift. I'm only moderatly good at it. That's why I was marginal as a waiter and a failure as a short order cook, where I would argue the problem is worse than for the waiter.

    Also, while the range of problems are fewer and the stakes much lower than the ER, being a Sys Admin is often very much like this, in the abstract. There are constant fires, there are the fires that interupt you working on fires, and somewhere in there is the preventative maintenece you should be doing. And finally, there's actually trying new stuff. Now I'm also managing Sys Admins and am supposed to be not only prioritizing for myself, but helping others prioritize. I need your skill set :-)

  3. You haven't posted in awhile, but the quality is high. I have a hard time sitting down and doing really meaty posts like this.

    I have the advantage of working overnight shifts and get occasional slow bits when I can muse and knock up a longish draft in the middle of the night when there is nothing else to do. I try not to post at 0400 because my thought processes may be amusing at that time but they are rarely clear or comprehensible.

    And I wonder whether my little algorithm might be something you can apply to almost any situation. Seems like it might be almost universal.

  4. And I wonder whether my little algorithm might be something you can apply to almost any situation. Seems like it might be almost universal.

    I think its a case of Gentry's Anvil. Yes, the algorithm is generally applicable, but it is a simple algorithm that is not hard to formulate. I have formulated something similar for my job. The difficulty is in the execution, especially consistant execution. At least for me, that is where the problem lies.

  5. So you have ADHD, yet you are able to post a nice long diatribe about how you have no attention span.


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