Login

Push vs. Pull in the ER

Looking at Interruptions as Broken Setups that Cause Human Changeovers

June 16, 2010

The last time that I took my daughter to the emergency room, I marveled at how the ER staff continued to turn on a dime to address the priority of the moment, for every moment of the five hours that we were there. It made my head spin to think about how frequently and how quickly the nurses and docs had to shift gears and re‐focus their attention. Granted, I’m a person who by nature prefers concentrative tasks with few interruptions. I’d probably be on the ER admissions list if I ever had to juggle all the incoming stimuli of an ER attending on a busy night.

Recent research findings revealed that regardless of preferred work style, multitasking takes its toll on productivity. A study published last month by Australian researchers found that interruptions result in ER physicians spending less time on their assigned tasks and leaving almost a fifth of their tasks incomplete. Read a summary of the study published in Quality and Safety in Health Care.

An earlier article written by Dr. Robert Wears, University of Florida, also focused on the impact of interruptions and multitasking on emergency department staff. Dr. Wears' article reported the findings of a root cause analysis of a "near miss" in which the incorrect medication was dispensed to a patient. Read "Caution, Interrupted".

The Lean Perspective — Push vs. Pull

In a transactional setting such as a hospital emergency department, being “demand driven” is inherent in the environment. It is not a case of the organization adopting pull scheduling to manage the workflow. Emergency department personnel do nothing until there is a customer order. As soon as there is demand, all activities are focused on meeting that demand in a pull execution strategy.

Rather, the challenge is to manage the 'push' from staff members who tend to push information and tasks along the process, irrespective of the recipient's readiness to handle the work. That push translates to interruptions on the receiving party's part and can jeopardize patient safety.

Looking at Interruptions as Broken Setups

Regardless of how convincingly our teenagers attempt to argue that they can effectively multitask and manage homework amongst text and IM messages with their latest iTunes playlist blaring in the background, humans are serial communicators. We transmit and process data one bit at a time. Any appearance of so‐called parallel processing is really a speedy switch from one data stream to another.

In Lean terminology, an interruption in work is a broken setup, just the same as if you had stopped a machine on the shop floor, mid‐batch, to make another part, explains MoreSteam faculty member Dr. Lars Maaseidvaag. Consider a typical manufacturing scenario: a machine that is producing good parts is interrupted and stopped to produce a different item, then turned back on after the setup and begins producing good parts of the new item. Machines have a very binary existence, on or off, and the amount of 'off' time is very easy to calculate and see.

The assumption is that humans can be more effective than machines in switching between tasks with little or no 'off' time. However, studies such as the two cited above find that humans cannot instantaneously switch from task to task and maintain full productivity. Humans don't immediately make good parts. People have to re‐orient themselves and get their heads around the topic each time that they switch to another data stream.

We have the illusion that we are able to immediately switch between tasks because we are probably getting a little bit of productive work done when we get interrupted. The reality is that we are not anywhere near full productivity for a period of time after the interruption and it takes us time to return back to our original task. Our productivity may not drop to zero as in the case of the machine, but we definitely experience our own, human version of a setup.

These considerations are critical in emergency departments where interruptions cannot be avoided. Interruptions cause "human changeovers" that damage productivity. Therefore, the authors of the two studies suggest ways to manage interruptions — such as task prioritization, workspace design, procedural and communication standards — and, thus, to protect patient safety. These same suggestions can be applied to a number of other transactional settings where it is impossible to simply post "no interruptions, please".


Ellen Milnes