As promised sometime ago, I have summarised parts from the Checklist Manifesto by Atul Gawande
Gawande traces the use of checklists back to 1935 when the US Army Air Corps were looking for a long-range bomber. Boeing developed the model 299, which was significantly faster and had a greater capacity than anything offered by other companies and was nick-names the 'flying fortress'. However, at a demonstration flight it crashed shortly after take-off. No technical failure was identified, and it was concluded that it was caused by the pilot forgetting to release a locking mechanism on the elevator and rudder controls. Some people concluded that the plane was too complicated, and would never be flyable by humans. Douglas won the contract to supply their less able, but less complex plane.
Some in the Army were still keen to use the Boeing 299. They realised that the pilots on the plane that crashed were some of the most experienced pilots in the business, so more training could not be the solution. Instead they came up with the idea of checklists for take-off, flight and landing. These lists were simple, brief and to the point. They were short enough to fit on an index card, but they worked. The Army went on to order nearly 30,000 of the planes, which were dubbed the 'B-17.'
Gawande explains that in complex environments, experts are up against two main difficulties:
1. Fallibility of human memory and attention - especially for mundane routine matters that are easily overlooked when there appears to be more important things to attend to;
2. The tendency to skip checks, even when remembered, because they don't always matter - things that could cause a problem but have never done so in the past.
Checklists help to overcome these difficulties because they provide a reminder of what needs to be done and they instil a "kind of discipline" that makes people do things they may otherwise skip over.
In 2001 a critical case specialist at John Hopkins Hospital names Peter Pronovost developed a checklist with the aim of reducing infections by making sure key steps were carried out when inserting a central line (tube inserted into a vein). At first nurses were asked to observe doctors and record whether the steps were performed. The result was that in a third of cases, at least one step was missed. Nurses were then authorised to stop doctors if a step was being missed. This was seen as 'revolutionary' because it gave nurses some power over doctors.
After a year of using the checklist the results were so spectacular Pronovost was not sure whether to believe them or not. So they kept monitoring. It was calculated that after a little over two years the checklist had prevented 43 infections, eight deaths and saved £2 million in costs.
Pronovost observed that checklists helped memory recall and clearly set out the minimum steps in a process. He was surprised with the results because he had not realised how often even very experienced people did not grasp the importance of certain precautions. His results were impressive, and he put a lot of effort into spreading the word across the country (speaking in an average of seven cities per month). But others were very reluctant to take up the idea, either because they did not believe the results or they simply thought that would not need a checklist.
A checklist can be useful, not just if there are lots of steps that can be forgotten or intentionally skipped, but if there are lots of people involved in the task. Taking construction as an example, Gawande explains that in the past the 'Master Builder' designed, engineered and oversaw all aspects of construction for a building. By the middle of the twentieth century this did not work any more, and instead a team of experts and specialists are required. This is where checklists start to become necessary.
The trouble is that some professions have been slow to realise that the nature of the job has changed and become more complex. In medicine doctors don't seem to realise that most patients receive attention from many different specialists. If a checklist or equivalent is not used the result is duplicated, flawed and sometimes completely uncoordinated care.
Sunday, June 06, 2010
Subscribe to: Post Comments (Atom)
Post a Comment