Coming Correct: Changing the Culture of Incident Reporting

Calef Letorney explains the importance of reporting incidents and how doing so can benefit other pilots.

By Calef Letorney
Originally published in USHPA Pilot, July/August 2018

Admitting mistakes is never fun. It’s embarrassing to think about screw ups, let alone discuss them. And so it’s only natural that most people opt to deal with mistakes quietly... like the time I tore the kayak off my roof rack while driving into the garage. No big deal, nothing a blowtorch couldn’t fix and not much to talk about… But when it comes to piloting, it is our responsibility to share incidents. Why? Because we need this information to improve the safety of our sports. The bonus is that incident reporting is the single biggest thing we (you, me, everybody) can do to help reduce our insurance premiums. Counter intuitive, I know, so stick with me.

Before we delve into why you should care, here’s why I care: I am committed to studying accidents and working to avoid them because of Max. A college kayaking buddy, Max was the first person to tell me about paragliding. He had just earned his P2 and he glowed while described flying. My mind was blown. I’d never heard anything like it. Seeing my interest, Max argued that my success in whitewater kayaking demonstrated excellent decision making and an expert understanding of fluid dynamics. I was a shoo-in. “But Max,” I said, “bad decisions and misreading whitewater are pretty much your MO. Your signature move is to take a beating and claim you meant to... So what does this mean for you?” Max fired off a grin and explained that he feared the consequences of flying more than kayaking. That fear kept him in line. Good enough for me.

I dove in head first. A month later I got a call from my instructor to tell me that Max had crashed. Details were slim, but apparently he had taken a road trip to the big mountains. On the ride to launch Max asked a mutual friend how to do wingovers. He received accurate, but, I would argue, inappropriate advice. Max proceeded to enjoy an epic, high glass off flight that lasted over 90 minutes; it would be his best and worst flight. Max was found deceased with his reserve lying next to him, out but not yet open.

Max’s accident has had two profound influences on me. First, I don’t hesitate in throwing my parachute. I’ve tossed twice and twice I’ve walked away with only mental injury (fear that lasted years) and flew again the next day. Yes, it’s scary and embarrassing, but as Rev. Santacroce says, “Those that toss, live.” Having dutifully run down that rabbit trail, the second impact of Max’s impact, was my enduring interest in accident analysis. Simply put, dissecting accidents and learning how to avoid those mistakes was just about the only way I could justify the absurdity and selfishness of sticking with this new drug that had just taken my friend.

You may have heard that accidents have multiple causes; we seek to identify and reduce those causes, as breaking just one link in the chain may prevent an accident. Looking at Max’s accident, it’s not hard to see the errors. Max was attempting to teach himself acro (strike one), over the ground (two), low (three), at high altitude (four), while amped up from an epic flight (five). When you think about it like that, Max’s accident hardly seems random. But do you think Max understood those risk factors?

Currently we rely mostly on anecdotal evidence and intuition for accident analysis; we see accidents and we make assertions like I just did. I’ve no proof and even less ability to understand the relationships between various factors. But if we had statistically significant data, smarter people than me could use regression analysis to really pick apart this problem. We could not only better understand the primary causes, but also tease out relationships between contributing factors.

What is regression analysis and how can it help make free flight safer? According to our very own data analyst and HG legend Dr. Felipe Amunategui (AKA Flipper): Our ability to predict events such as weather patterns or stock market changes is very limited, but it can be improved somewhat by looking for relationships between one or more variables and the chances of an event happening. Statistical methods allow scientists to measure the strength and relationship between variables and a specific outcome. These methods are usually referred to as correlation analysis or regression analysis. Regression analyses are used by insurance companies, for example, to determine the risk a given individual has of experiencing an adverse event such as death. Regression analyses have been used to determine that women live longer than men, that married men live longer than single ones, and that women’s life expectancy is not increased by marriage. Another statistical trivia from this type of analysis tells us that the lowest probability of getting a speeding violation on an interstate is between Wednesday night and Thursday morning in the middle of the month. By accumulating information on accidents and incidents over time, we will be able to create increasingly accurate risk estimates for a given pilot flying a specific site during a specific part of the year. Without the information you provide we will remain ignorant about what are real risks, for whom, and under what circumstances. All tidbits of information may be useful at this point as our accident-and-incident database is in its infancy. I have a dream that one day pilots may be able to accurately estimate risk and to identify specific variables to address in mitigating the risk.

Flipper fantasizes that with enough data, one day we could go so far as to create an app that could use the various factors of the flight (weather, site, pilot, equipment, etc.) to give a pre-flight risk assessment with probability of incident. We’re talking about statistically-driven risk forecasting with suggestions of hazards to watch out for. To be sure, this is a long way off, but these things are not impossible.

Closer to reality, if we had accident information we could study it and perhaps learn more about how to mitigate the risk factors, then integrate these concepts into our instruction pedagogy. But first we need the quality data. You may have heard the expression “garbage in, garbage out.” Our hush-hush culture yields a dismally low rate of incident reporting: garbage data. Have you reported all your incidents? Yes, I am asking you. Until recently, neither had I, so it’s no wonder we can’t yet do meaningful data analysis. We need the data first, and in order to get that, we need to change the culture around incident reporting. Incident reporting needs to be commonplace, or better yet, expected.

What should be reported? For the best analysis we need data on ALL incidents and accidents. What’s an incident? An incident is anytime a flight plan deviated from normal and had the potential to become dangerous. Blown launch that resulted in a paraglider in the bushes or trees? Incident. Broken weak link towing? Incident. Broken HG tube on landing? Incident. Scraped knee while kiting? Incident. Think about the above scenarios—they can be commonplace and forgettable, or fatal; the difference is luck (and maybe quick thinking and skill) so we seek to study all incidents. Anything that required medical attention (the definition of an accident) should absolutely be reported. If you had to ask yourself “was that an incident?” the answer is almost definitely yes. Forget the excuses and give us your data!

"Have you reported all your incidents? Yes, I am asking you. Until recently, neither had I, so it’s no wonder we can’t yet do meaningful data analysis."

Now you’re probably thinking “ARE YOU SERIOUS?! That’s a lot of logging!!!” Yes, yes it is. Fortunately it’s quick and easy. I’ve reported many incidents (both my students’, friends’, and my own) and it only takes 5 to 10 minutes. With all the time and energy we put into flying, we can invest 10 minutes more to divulge our (hopefully) infrequent mistakes. Do your part.

HOW TO LOG: Go to www.USHPA.org > Pilot Resources > Safety and there’s a link.

Leading by example: We’re asking you to make a quick, private incident report, which to some still feels embarrassing and cumbersome. To show that talking about our mistakes is no big deal, Flipper and I are going to go way beyond what is being asked of you. Here are detailed accounts of our embarrassing screw-ups.

CALEF: At a closed ski area in Vermont on 11/18/2016 I blew a launch on a paragliding flight with a friend. I had just finished helping a few newer pilots launch and the conditions were getting worse. The ski-trail launch was a moderately narrow slot in the trees with a decent grass slope. The wind was calm, or up to 3 or 4 mph from 90 degrees cross. Occasionally the wind trickled downhill. Not ideal, but it was light and I’ve launched like this before (strike one). So we did a forward launch in a calm cycle. We ran, and ran, and ran down the ski slope. Actually, the video revealed it was just me running as my passenger sat down almost immediately. Still, my fault for putting us there and not stressing the running part enough during pre-flight (strike two). Finally just as we were lifting off, 70’ farther down the trail than normal, my left wingtip hooked a snow-gun boom sticking 20’ out into the ski trail (strike three). When I saw the snow gun hooked on my wingtip I immediately took a wrap and flared with all my might. We tetherballed around the snow gun. The boom rotated 90 degrees and took a lot of energy out of the equation. We left the ground for a second and came down in a tumbling tandem PLF. The wing fell onto the trees. We got lucky. No injuries to pilot, passenger, or wing. After a thorough inspection of the glider, two hours later I launched and had a nice sled ride.

How did this happen? The conditions were worse than marginal. I had taken similar risks before and always done well, so I had normalized that risk. This is especially not cool when flying tandem. Then the snow gun surprised me. It was over the rise, not really visible from launch. In late April at the start of the flying season I had moved that gun completely out of the way. I never thought to look again, but the snowmaking crew had recently (in the last week) put it back into position (sticking way out into the trail) in preparation for the snowmaking season. Surprise! In the heat of charging down launch, I just didn’t see the narrow silver gun sticking way out of the tree line. From this incident I have learned several things: First, I am even more conservative and respectful of launch conditions, especially on tandem. Second I now check to make sure the snow guns are out of the way every time we launch.

FLIPPER: Turning from the downwind leg to base, the hang glider’s adverse yaw tracked me away from the narrow landing zone. With less than 100 feet from the ground, the path of the glider presented three options: a house, a large pile of bricks or a 60-foot-tall oak tree. It was 1996, the glider a Pacific Airwave Magic Kiss (of death), the site Italy Valley, NY. I had flown for 65 minutes in glassy air before heading to the LZ. While I had well over 1000 hours flying hang gliders, this was a new glider to me, and it had very different characteristics from anything I had flown up to that point. Also, I was about 50 to 75 pounds BELOW the recommended pilot weight.

I picked the oak, and did the best flare I could on the very top of the tree. Flying in the East one learns that there are two type of pilots: those who have landed on trees and those who will. I had just changed my status, but I was prepared. Upon “landing” (or arboring) I let go of the glider and grabbed tree limbs for dear life. Once the glider settled, I produced my tree-rescue kit, and proceeded to tie the glider to the tree. I had to wait an hour and 45 minutes hanging out waiting for a vintage fire truck with a telescoping ladder to arrive from a nearby town. I was fortunate to have a very understanding rescue squad. They agreed to use the ladder to pluck the glider from the oak, so it flew again the next day. There was no AIRS back then to report the incident, and I was terribly embarrassed by the entire event, but there were important lessons to gain from it. Like: Fly new gliders in places that do not require a precise approach. Carry a tree rescue kit with you, and know how to use it, and fly within the weight range of your glider.

But wait, there’s more: We’ll pay you cold, hard cash for reporting your incidents! Joking... sort of. You won’t get a check in the mail, but incident reporting is the number-one way to reduce USHPA’s insurance costs. Insurance is a leading expense for USHPA and if we can reduce the insurance costs, we can reduce membership dues. Most people would assume that reporting incidents makes insurance premiums go up, not down. Our situation is just the opposite. USHPA is the majority shareholder in our insurance company, but our rates and reserves must be approved by industry regulators so we don’t bankrupt our nascent insurance company. A major contributing factor to USHPA’s high insurance rates is IBNR, which stands for “Incurred but Not Reported.” Basically, unlike automobile crashes or home fires, USHPA has very, very little data on the number of incidents that happen in our sport and how likely they are to result in claims. As such, the insurance actuaries recognize that there is a high likelihood that people are getting hurt and not telling anybody, but these injuries could later result in insurance claims. We are exposed to liability from claims on accidents we don’t even know happened. This is the heart of IBNR. A major part of USHPA’s insurance premium goes into reserves that we are obligated to hold onto until the statute of limitations of potential IBNR accidents runs out. Yes, we get the IBNR money back eventually, but in the meantime our insurance rates are high, which directly contributes to why membership dues have increased from $99 to $150/yr. BUT, if the USHPA community can show success in logging all our incidents, we can convince the regulators to reduce the cost of our insurance due to IBNR. Confusing, yes. At the end of the day we need to trust when the smart folks at our insurance company tell us the best thing we can do to reduce insurance costs is to report all incidents. Crazy, I know.

It’s understood that culture is the hardest thing for an organization to change and the hush-hush culture around accidents is no different. This change isn’t going to be easy, and we need everybody‘s help. If you see or you are involved in anything that smells like an incident, PLEASE report it. If you’re an instructor, don’t forget that you are already obligated to report accidents, and failure to comply is grounds for revocation of instructor certification. Don’t be afraid that reporting incidents is going to get a school in trouble. On the contrary, PASA and the RRRG expect commercial operators to have incidents; it is an unavoidable part of our sports. So when a commercial operator has never reported, rather than looking squeaky clean, this appears dishonest.

To be clear, nobody’s asking you to publicly post so the misanthropic hyenas of the internet can pick the meat from your self-esteem. You just need to take five minutes to file a private report so we can all learn from your mistakes. And who knows, the knowledge gleaned from your incident report might save a life, perhaps even your own. So let’s report all incidents and work together to create a culture of incident reporting.

Your AIRS report is confidential, anonymous, and protected by a Federal Certificate of Confidentiality. This means that it is impossible for an attorney or judge to force the disclosure of information entered into our accident and incident reporting system (AIRS). This step was taken to protect reporters, the sensitive information shared by observers, and to protect access to sites. Once in AIRS, information cannot be removed, and it can be accessed by three individuals only. Access also requires having a decoding key. Even if the database is hacked, it is meaningless without the decoding key.

And if you witness or endure an event, do not assume someone else will report it! Go ahead and make your own report; we can easily handle multiple reports about a single incident.