There's a lot of talk about air traffic controllers falling asleep on the job these days. First at Reagan in Washington DC, then apparently at other airports - including Reno and Boeing Field here in Seattle. In each case, the dozing employee was suspended. And today, we find out that the Chief Operating Officer of Air Traffic Organization Hank Krakowski has resigned from his post. FAA Administrator Randy Babbitt had this to say:
"Over the last few weeks, we have seen examples of unprofessional conduct on the part of a few individuals that have rightly caused the traveling public to question our ability to ensure their safety. This conduct must stop immediately."
The FAA is also adding air traffic controllers to midnight shifts at 27 towers in different cities around the US. But as a frequent flyer, I have to ask myself if I feel safer as a result of these actions and tough rhetoric. Any of you who have been through our root cause analysis training know that disciplinary action employed as a solution strategy raises red flags because it is seldom effective at actually reducing risk. Were errors made in these cases? Undoubtedly - but it doesn't take a professional investigator to figure that out.
I'm a fan of James Reason and his approach to understanding and managing human error. While I don't subscribe to his categorization model for accidents, I do find his analysis of how errors come to pass very useful. My understanding of it goes like this:
You can't commit an error without intent. In other words, you need to have intent as a motivator for your actions in order to commit an error - otherwise it's just an involuntary response. Drooling on your pillow while sleeping, for example, is not an error because there was no prior intent. In this case, the air traffic controllers did have an intention - which was to safely manage the air traffic in the airspace for which they were responsible.
The next question is whether the actions proceeded as planned. From the articles I've read, it would appear that in most of these cases, the answer to this would be no - the air traffic controllers did not plan to fall asleep on the job. One exception to this was in Knoxville TN where the controller apparently made a bed on the floor of the control tower out of pillows and slept for 5 hours. That's another story, but one that likely still shares many causes.
When actions proceed as planned, and they achieve the desired results, then there's no problem. In fact - this is what happens most of the time. But in the case of the air traffic controllers, we see that this wasn't the case. So now what?
Reason has developed a decision tree to assess culpability, which can be found in it's original form on page 209 of his book Managing the Risks of Organizational Accidents. There are various customized forms of this model, but they each subject the error to a series of tests to determine whether disciplinary action would effectively reduce the risk of recurrence. The following is based on a modified version of Reason's decision tree.
First is the Deliberate Harm test. Were the actions intended to cause harm? If so, then disciplinary action is absolutely appropriate. Did the air traffic controllers intend deliberate harm? I think the answer is no - not in any of the cases.
The next is the Health test, which examines whether there were underlying medical conditions that resulted in the error. I would say that the air traffic controllers were not fit for duty due to the fact that they could not stay awake during their shifts. Therefore disciplinary action is not likely to be an effective solution strategy. Effective solutions are effective only because they control one or more causes of the problem. And I don't think you can control a health problem with punishment.
The next test examines the Procedures that govern behavior. Did the employee violate established procedures? And if so, were the procedures clear, correct, safe, and being followed by others? In the case of the air traffic controllers, I think the answer to the first part of this test is yes - each of them violated established rules against sleeping on shift. The second part of the test though is harder to answer accurately without in-depth analysis and interviews, particularly whether or not the procedures were being followed by others. I suspect these procedures were violated as a matter of course, although I can't back it up.
The final test examines whether or not someone with similar training/position would have likely acted in the same way in a similar situation. Given the fact that so many instances have come to light, I think the answer to this question is yes - they have acted in a similar fashion in the past, and will likely do so again in the future if no changes are made.
Last night I saw a segment on the PBS News Hour that included Alan Levin, who covers aviation for USA Today. In this interview, and in his April 14, 2011 article in USA Today, Levin discusses systemic fatigue as a problem throughout the air traffic control group at the FAA. The FAA has apparently known about this problem for some time now, according to this article. That shouldn't be news to anyone familiar with the patterns of undesirable events. For every incident that comes to light, there are several that no one found out about. I think it's safe to say that air traffic controllers frequently fall asleep on the job during this midnight shift and that they don't post it as a status update on their Facebook page:
In the News Hour interview, Levin goes on to discuss how structured napping has been proven to help alleviate fatigue but that no one will propose this as a solution because it won't pass the "Jay Leno" test... in other words, it would be ridiculed non-stop by late night comedians as well as politicians.
At the end of the day, this drama shows just how far we have to go with respect to managing human error. The solutions proposed include the resignation of a high level manager, the suspension and likely dismissal of 6 air traffic controllers, and the expressed outrage of FAA Administrator Randy Babbitt. I hope they don't stop here, but if the story dies that's likely going to happen.
Whew... I feel like they've really gotten to the bottom of this one.