Flying while intoxicated doesn't happen often, but when it does the results are usually tragic. A single DWI may point to trouble ahead in airplanes.
April 27, 2003
This article appeared in the October 2002 edition of Aviation Safety and is reprinted here by permission.
Recent high-profile incidents involving allegedly drunk airline pilots reporting for duty have renewed the call for more random drug/alcohol testing of flight crews. Public hypersensitivity over anything aviation even led CNN to host an online poll that asked if all pilots should be tested for alcohol -- to which 89 percent replied yes.
Pilots, meanwhile, point to a relative dearth of alcohol-induced crashes and wonder what all the ruckus is about. Good sense, it seems, keeps alcohol-impaired pilots out of the cockpit the vast majority of the time, because the pilot's sitting in the seat that frequently hits first.
The FAA has been trying for years to get a handle on alcohol use by pilots, both through random testing and driving records. In fact, for thousands of pilots, box 18.v. of FAA Form 8500-8 represents either a dilemma or a cruel reminder. The form is the FAA's application for the medical certificate. The box: Have you been nailed for driving while intoxicated?
While drunk driving has been under relentless scrutiny for the last two decades, alcohol use by pilots has been a much less visible affair, aided by the fact that a study of U.S. airline accidents up to 1990 failed to turn up a single accident ever in which a pilot was found to be drunk at the time of the crash.
While general aviation has not compiled such a spotless record, pilots do a much better job than drivers at avoiding bent metal. While the rate of alcohol-related accidents is low, they are frequently deadly. A review of accidents the NTSB determined to be alcohol-related found 33 such accidents in 1996-2001. About 70 percent of those accidents involved fatalities, and in only two did any occupant escape unharmed. Thirty-six people died.
The accidents generally fall into four categories: buzzing, skill impairment, drunk pilots trying to function normally, and what can only be classified as general irresponsibility. Granted, there's some overlap here, but we'll get to that.
||(Click graphs for larger versions)
Not surprisingly, buzzing is the most common. Just as beer and joyriding go together in the eyes of many newly licensed teenage drivers, the record is populated with pilots who have a few and decide to take their friends flying. It is during just such an activity that the pilot is vulnerable to errors in judgment, reaction time and motor skills.
The speed rush involved with flying low seems particularly appealing. One pilot even taunted a passenger who suggested they fly higher, saying "Why, you ain't scared, are you?"
One pilot flew underneath a bridge before hitting wires. One pilot skimmed a golf-course pond, skipping the wheels off the water before one caught enough to make the airplane nose over.
Eleven of the 33 accidents fell into the buzzing category. In several cases, there were other factors that showed the pilots' general disdain for playing by the rules. In two the airplanes were stolen. In another the pilot had no pilot certificate and the airplane was several years out of annual.
Motor Skills? No Thanks
Skill-impairment accidents are those in which the pilot appears to be trying to fly responsibly (if that can be done while intoxicated), but just isn't up to the task. We'd equate this roughly with driving home slowly and carefully after imbibing a few too many at a party.
Eight of the accidents primarily fell into the category of skill impairment. In these we include accidents prompted by the lack of a preflight, such as the one in which the control lock was still installed. It includes the one where the pilot reported being out of gas, with the fuel selector on both, when in fact the fuel selector didn't have a both, one tank was empty and the other was nearly full. We include the unexplained loss of control shortly after takeoff (aided by a .285 blood alcohol level) and the three in which a safe landing was just too much to ask.
Finally, we'd include the accident in which the pilot, who had been drinking at a bar with a friend until 3 a.m., decided to fly home at 4 a.m. so he could have the rental returned by 8 a.m. His wingtip caught the water during an abrupt turn.
Welcome to Reality
Five accidents we classified as drunk pilots trying to function normally, although the alcohol may have made it more difficult to make appropriate judgments. These don't necessarily happen late at night or after a few hours of barhopping. In fact, they tended to occur during the day, often when the pilot was working.
There didn't appear to be any measure of buzzing or showing off or joyriding, but in other respects they tend to be similar to the skill-impairment accidents outlined earlier.
They include the pilot at a ranch who was helping riders on the ground round up stray cattle. He flew the airplane into a cliff after encountering downdrafts. It was 9:30 in the morning and the pilot had not been partying. However, his driving record included several DUIs and he had not gotten a medical or annualed the airplane in more than five years. An associate said he treated his airplane like he did his tractors.
Another accident involved a pilot who supplied charter flights to a local business on occasion. On one occasion, he took off with the de-ice equipment inoperative on one engine and entered icing conditions. The airplane crashed, killing all three aboard.
I'll Do it My Way
Another category of accidents can only be described as irresponsibility -- or perhaps even more flagrant irresponsibility than the pilots in the other categories. Here's where some of the seven accidents got weird.
A couple of pilots embarked on low-level aerobatics while intoxicated. One accident involved a pilot whose license had been revoked seven years earlier for stealing an airplane. He and a friend stole another one and overran the runway trying to take off. Seems the airplane they appropriated was being refurbished and was missing its elevator.
One pilot crashed his airplane off the end of the runway. A sheriff's deputy saw the crash and went to the scene to render aid. The man at the wreckage claimed it was his airplane, that it had been stolen, and he was investigating where it had crashed. The deputy gave him a ride back to the hangar. The pilot got his truck to tow the crashed airplane out of the swamp, but on the way to the airplane was arrested for DUI by another deputy.
Sometimes a drunk pilot meant a drunk passenger. In two cases the passenger passed out, jamming the controls and inducing the pilot to lose control.
But perhaps the worst drunk passenger is the kind who, as the drunk pilot completed the landing roll and started turning off the runway, shouted, "Let's do it again!" and firewalled the throttles for another takeoff.
If you're counting that leaves two of the accidents unclassified. They need their own classification, with an asterisk. Both involved suicidal pilots. Both left notes. One was a young adult, drunk and despondent about his future. The other was a convicted arsonist who was about to be arrested again for arson. He'd had several DUIs and, at the time of the accident, had also ingested cocaine and Valium.
These two tragic events are not aviation accidents so much as individuals crashing against an impossible reality. If not an airplane, they would have used something else; they just happened to have access to wings.
The Regulatory Approach
The FAA has invested a great deal of effort and money into addressing the problem of flying while intoxicated. The fact that alcohol was cited as a factor in only 33 accidents over six years -- less than 3/10ths of 1 percent of the 13,000 accidents during that period -- shows either that it's working or that pilots really are too smart to drink and fly.
The FAA's primary methods of deterring flying while intoxicated are random testing of commercial pilots and inducing pilots to report being busted for driving an automobile while intoxicated. The DWI reporting, because it affects all pilots, remains the cornerstone of the effort.
The question remains, then, whether the emphasis on DWI during the medical application is screening the problem drinkers out of aviation or whether the drinkers are either keeping themselves on the ground or somehow managing to fly without accidents.
Attempting to make the connection between a pilot who gets caught driving while intoxicated and one who has an aircraft accident, Kathleen McFadden, a researcher at Northern Illinois University, studied the driving and flying records of more than 70,000 airline pilots over a seven-year period. The results were sobering.
She compared records from the National Driver Register database with the accident/incident database. The good news is that 98 percent of airline pilots had never received a DWI conviction. The bad news is that those who had received a DWI were more likely to have had an aviation accident.
And it gets worse from there. A history of one DWI conviction was associated with double the risk of a pilot-error accident. Two or more convictions more than quadrupled the risk -- this despite the fact that no pilot tested positive for alcohol during any post-accident investigation during the study period.
Interestingly, in the 13 cases in which the airline pilot had both a DWI conviction and a pilot-error accident, the DWI preceded the airplane crash 10 times.
Although these numbers are small, the researcher concluded it was reasonable to use a generous statistical threshold for significance because of the public-safety aspect involved.
The study raises an important issue with respect to the FAA's request for information regarding DWI convictions on the airman medical certificate application. Any DWI should be taken as a significant red flag -- not just a second in three years or a total of three or more, which is the criteria AMEs use in determining whether to issue the medical.
At the same time, the study also showed there was no evidence to support the notion that random alcohol testing would have prevented any airline accidents, except insofar as it served as a deterrent.
The same researcher followed up that study with another that looked at whether the FAA's program of penalizing pilots with DWI convictions has made any difference. Prior to 1990, the FAA did not take any action against pilots with DWIs unless they failed to disclose them on Form 8500-8.
Beginning in 1990, the FAA changed its tune and began to suspend or revoke certificates for pilots who had more than two DWIs in three years. The medical application got a provision giving the FAA permission to search the driver registry for convictions and began requiring pilots to disclose DWI convictions not only on the medical application but within 60 days of the conviction as well.
However, the program is not without exceptions, and some pilots who have shown remarkably poor judgment about driving continued to fly for airlines. In McFadden's follow-up study, she found 29 airline pilots with four DWI convictions, six with five convictions and three with six.
The researcher concluded the FAA's reporting requirements and sanctions against pilots with DWI convictions was ineffective. The number of airline pilots with DWIs increased between 1986 and 1997, going from 0.97 percent of the airline pilot population to 1.62 percent, while the number of drivers arrested for DWI during the same period dropped nearly 20 percent. The percentage of pilots with more than one DWI also increased, going from 0.08 percent to 0.13 percent.
Despite this, the fact is that airlines have an enviable record with respect to alcohol-related accidents when compared to transportation industries such as rail, bus and trucking. While medical-certificate screening may be part of the answer, it may also ride on the fact that there are two or more flight crewmembers on each flight, and they act as checks on each other's judgment and skills.
This is where the difference between the airline pilots in McFadden's study and general aviation pilots becomes apparent.
Flying is a complex task, far more complex than driving, and is susceptible to alcohol-induced errors to a far greater degree.
The first cognitive function lost to the effect of alcohol is judgment, which disappears long before the hand/eye coordination goes down the tubes. But GA pilots typically don't have the backup of sophisticated flight-management systems and another highly experienced pilot watching from across the throttle quadrant.
Drinking and flying can be an egregious affair. Getting into an airplane after a few drinks, even if you've flown the same airplane 1,000 hours in the last five years, is a far cry from getting into your car and driving home from a party. The car is stupid, the airplane insane. Same goes for trying to fly "the morning after."
There may not be any cops patrolling in the air, and there may not be much other traffic to dodge or curbs to avoid, but you'll still have to deal with the dynamics of moving air and stationary ground -- which can make bed-spins feel like a picnic.
Besides, make a rough landing and you may spill your strawberry daiquiri. And how will you get that stain out of your Hawaiian shirt?