Ten years after the creation of the International Helicopter Safety Team, a lack of detailed, reliable information on accident rates and causes continues to hold back safety progress.
By Elan Head | February 16, 2016
Estimated reading time 28 minutes, 8 seconds.
If you’ve spent a lot of time flying in helicopters, you’ve probably also spent time wondering about how dangerous that is. You likely take it for granted that it’s statistically more dangerous than flying, say, Delta Air Lines, but how much more? Ten times? A hundred? Or is it astronomically more dangerous, as suggested by a widely quoted statistic that helicopter air ambulances in the United States have a fatal accident rate that is 6,000 times that of commercial airliners?
The fact is, you don’t know. That’s not because you haven’t spent enough time researching or worrying about it. It’s because no one really knows, at least not to a high degree of certainty. The estimates in the United States are getting better, but in most parts of the world, and most sectors of the helicopter industry, all we can do is guess. Accidents are relatively easy to count, but accurate flight hour estimates are harder to come by–and while that poses a number of problems for the industry, it poses a special problem for an organization that is committed to reducing the worldwide helicopter accident rate by 80 percent.
Ten years ago, the International Helicopter Safety Team (IHST) was founded with just that goal: to reduce the helicopter accident rate to one-fifth of its 2001-2005 baseline level by 2016. Now that that horizon is upon us, the only sure thing we can say about the IHST’s progress is that it did not achieve its target. In the U.S., one of the few countries that calculate a helicopter accident rate, the IHST estimates the 2014 rate at 4.26 per 100,000 flight hours, which is still well above its goal of 1.59. Moreover, even the apparent 47 percent reduction compared to the baseline rate of 7.97 accidents per 100,000 flight hours is mostly illusory; almost all of the improvement occurred in the first year after the IHST was founded, and was largely the result of using better flight hour estimates.
There are a few possible reasons why the IHST fell short of its goal. One is that the organization was simply too ambitious: in aiming to bring all helicopter operations, not just commercial ones, to a commercial level of safety, the IHST sought to do something that fixed-wing general aviation hasn’t even attempted. Indeed, the U.S. Federal Aviation Administration (FAA) is now seven years into a 10-year campaign to reduce the general aviation fatal accident rate, but it’s only targeting a modest reduction of 10 percent.
The IHST’s broad focus on all helicopter operations–from private owners flying their friends to air medical operators transporting patients; from crop-dusting Bell 47s to offshore Sikorsky S-92s–also shaped the IHST’s analysis of helicopter accidents in important ways, giving more weight to generalized human factors and less to underlying structural ones. As a result, the team has concentrated its efforts on relatively weak forms of pilot outreach, such as safety bulletins, while doing less to tackle big, expensive issues, such as infrastructure and crashworthiness.
So did the IHST simply make the wrong choices at the outset? Maybe; maybe not. The helicopter industry might have been better served by establishing different targets for various types of operations–just as Delta Air Lines is held to a higher standard than a weekend warrior in his Piper Cub–but the lack of accurate flight hour data by sector would have made that tough. And, while the IHST could have focused from the start on reducing the fatal accident rate rather than the total accident rate, the industry has largely ignored one of the team’s most important recommendations: the installation of cockpit recording devices. That means there’s still a lot we don’t know about how and why people die in helicopter crashes, particularly catastrophic ones that leave no survivors.
By most standards, including its own, the IHST’s first decade has been a disappointment. But it doesn’t have to be a failure. There’s still plenty of passion for improving safety in the helicopter industry, and the IHST’s experience to date can tell us a great deal about how to move forward. More and better data would be a good place to start.
What was missing
Ironically, the IHST was founded with the explicit goal of bringing a rigorous, data-driven approach to identifying and addressing the causes of helicopter accidents. The organization grew directly out of the first International Helicopter Safety Symposium in Montreal, Que., in 2005, where industry and government representatives convened to discuss the growing number of helicopter accidents worldwide. A key lesson of the conference was that new analysis methods were available to identify safety issues and guide interventions, and the IHST was created to put those methods into practice.
The team modeled itself after the U.S. Commercial Aviation Safety Team (CAST), which was founded in 1998 as a collaborative effort between the commercial airline industry and the FAA. The CAST used a data-driven strategy involving three types of teams: Joint Safety Analysis Teams (JSATs) that performed in-depth analyses of accidents and developed intervention recommendations; Joint Safety Implementation Teams (JSITs) that determined the feasibility of those recommendations and created action plans to implement them; and a Joint Implementation Measurement Data Analysis Team (JIMDAT) that measured the effectiveness of implementations and identified future areas for study. By all accounts, the CAST has been a success; in its first decade, it saw the fatality rate of commercial air travel in the U.S. decline by 83 percent.
From the beginning, however, there were some significant differences between the IHST and the CAST. Whereas the CAST focused its efforts on a narrowly defined, highly regulated sector of the aviation industry in the U.S., the IHST sought to impact all civil helicopter operations worldwide. This was less of a grand, inclusive gesture than a calculated commercial decision; although members of the general public draw a clear distinction between Boeing 737s and Cessna 172s, to most of them, a helicopter is just a helicopter. That means that any helicopter accident tends to negatively impact public perception of the industry as a whole, and airframe manufacturers have a vested financial interest in reducing negative perceptions.
However, there were also practical reasons for focusing on a single helicopter accident rate, since no one had helicopter flight hour data by sector. A helicopter was just a helicopter as far as our statistics were concerned, too.
Another difference between the teams was that the CAST targeted a reduction in the fatality rate of commercial air travel, while the IHST chose to focus on cutting the total helicopter accident rate. The IHST’s founding members only made this decision after considerable debate, as many of them thought that fatal accidents should be the priority.
“In the beginning, I favored a focus on fatal accidents, but was willing to support the broader goal,” recalled Bob Sheffield, a former managing director of Shell Aircraft who was a founding member of the IHST’s executive committee. He explained, “Everyone intuitively had the hope that if you focus on all accidents, fatal accidents will go down as well.” Again, there were some practical reasons underlying the IHST’s decision.
According to Roy Fox, a former chief of flight safety at Bell Helicopter who argued in favor of focusing on the total accident rate, one of those was the relatively low number of fatal accidents, while another was the shortage of good evidence in them. As he has often pointed out, without cockpit recording devices or survivors who can provide reliable accounts of what went wrong, much of the evidence in fatal helicopter crashes is purely circumstantial.
That lack of solid evidence may have been the biggest difference between the CAST and the IHST. Most transport-category aircraft–the CAST’s focus–are mandated to have cockpit voice and flight data recorders, which are rich sources of information in the event of an accident. Moreover, accidents involving commercial airliners in the U.S. are generally high profile and trigger major National Transportation Safety Board (NTSB) investigations that dive deep into systemic and organizational issues. Consequently, CAST members had a wealth of detailed, substantiated information available for their analyses, which helped guide them toward salient recommendations and effective implementations.
By contrast, very few helicopters are required to have cockpit or flight data recorders, and the technology wasn’t even available for light helicopters until relatively recently. Not only that, rarely do helicopter accidents receive thorough investigations by the NTSB or other transportation safety bodies around the world. When the U.S. Joint Helicopter Safety Analysis Team (JHSAT) performed its initial analyses of helicopter accidents in the U.S., it found incomplete data issues in around 80 percent of them–the same proportion of accidents the IHST was trying to prevent.
In a report analyzing calendar year 2006 accidents, the U.S. JHSAT noted many of these incomplete data issues reflected “information unavailable to the investigators,” and could have been resolved by the presence of flight data monitoring equipment. However, it also identified issues related to “incomplete accident investigation” and “inadequate documentation of the accident circumstances.” That’s because, when it comes to helicopter accidents, many NTSB investigators literally phone it in.
In 2006, the team found, NTSB investigators traveled to the site of just 14.5 percent of U.S. helicopter accidents. Meanwhile, fully 29 percent of those accidents received only “data collection” reports (while other, so-called “limited” investigations were delegated to FAA representatives). Data collection reports document accidents based solely on information provided by the operator; they assume that the operator is completely forthcoming, and has a full understanding of what went wrong. But the U.S. JHSAT found that “data collection reports frequently contained misleading information when an operator statement suggested conclusions that were not consistent with easily obtained evidence.” For example, one such report accepted the operator’s suggestion that the cause of the accident was “dynamic rollover,” when in fact the pilot ran out of fuel, entered autorotation, and made a forced landing on unsuitable terrain.
These incomplete data issues were reflected in the U.S. JHSAT’s top two intervention recommendations: “install cockpit recording devices” (which was relevant to 52.8 percent of analyzed accidents) and “improve the quality and depth of NTSB investigation and reporting” (relevant to 35.9 percent). Despite obvious problems with the quality of the data, however, the JHSAT proceeded with a CAST-like analysis of helicopter accident causes. And this is where the IHST became less data-driven than data-driving, as its conclusions were profoundly shaped by how it chose to handle the dearth of solid information available to it.
Take, for example, the IHST’s decision to use a single helicopter accident rate, rather than setting different targets for various types of commercial and private operations. Helicopter operations are extraordinarily diverse, and helicopter models vary tremendously in their capabilities. When you look for commonalities across the full spectrum of aircraft and operations, fallible human pilots may be the only constant you find. That’s exactly what the U.S. JHSAT discovered, and why most of its remaining “Top 20” intervention recommendations–ordered by the percentage of analyzed accidents to which they bear some relevance–relate to pilot training and judgment.
The team’s decision to focus on all accidents rather than fatal accidents also had an impact on its conclusions. For example, the U.S. JHSAT observed that, while only nine percent of the accidents it analyzed involved a post-crash fire, these accidents accounted for 48 percent of all fatalities. But in its matrix of “standard problem statements” (SPSs) versus intervention recommendations, “post-crash survival” scored as a low priority.
The IHST had compelling reasons to believe it was on the right track, despite the limitations of the data. According to Fox, after identifying SPSs and recommended interventions, the U.S. JHSAT found that “only 20 percent of their recommendations were actually mission-specific, whereas 80 percent were applicable for the entire fleet, regardless of mission.” That would seem to support the idea that the industry’s safety issues can be addressed most efficiently through broad educational campaigns of the type the IHST has pursued, rather than structural reforms narrowly targeted by sector.
But that may be an illusion sustained only by the absence of detailed accident data. Fallible human pilots are common to most transport-category airplane accidents, too, yet the CAST’s wealth of high-quality information allowed it to identify and implement dozens of highly specific, system-wide safety enhancements, rather than simply calling on pilots to follow the “I’M SAFE” checklist (number six on the U.S. JHSAT’s list of intervention recommendations). In citing its accomplishments, the CAST can point to new equipment and operational requirements backed up by regulations, revisions to training programs and standard operating procedures (SOPs), and changes to aircraft design. By contrast, the IHST’s achievements are mostly limited to safety “toolkits,” bulletins, and fact sheets–which may or may not have had much of an impact. In a report issued by the U.S. Joint Helicopter Implementation Measurement Data Analysis Team (JHIMDAT) in March 2014, it noted “stagnancy and regression” across most of the accident categories it examined. When it came to the tools and resources developed by the IHST, the JHIMDAT said, it was unable to determine whether they had been meaningfully implemented at all.
That’s not to say that the IHST hasn’t accomplished anything. Whether or not you agree with the priorities that inform their conclusions, the U.S. JHSAT’s accident analyses have made a tremendous contribution to the helicopter safety literature–underscoring, if nothing else, how much we don’t know. Other sub-groups within the IHST have also been extremely productive. One of those is the U.S. Helicopter Safety Team’s (USHST’s) Training Working Group, which has been responsible for many of the educational materials produced by the IHST in the U.S. Led by Bristow Academy’s Nick Mayhew, the team has created general educational products as well as products specifically targeted to the flight training sector, which consistently accounts for a high percentage of the industry’s total accidents.
The working group has drafted advisory circulars on autorotation training for the FAA, gotten involved with changes to the FAA’s practical test standards, and investigated topics such as night flying requirements and simulator training credits. Recently it launched a “special emphasis program” in cooperation with the FAA’s Orlando Flight Standards District Office, in which pilot examiners regularly speak to certified flight instructors (CFIs) about weaknesses they’ve observed on recent checkrides. “It closes the loop,” Mayhew said of the program. “My ambition is to work closely with the FAA to implement it [across] the U.S.”
Mayhew admitted that the all-volunteer nature of the IHST can be a challenge. He credited the working group’s success to the dedication of its members and their commitment to forward progress. “I think it has come down to having the right people in the working group and having regular meetings,” he said. “The secret I think is to keep people focused.”
IHST’s European arm, the European Helicopter Safety Team (EHEST), can also claim significant progress. The EHEST conducted its own detailed analyses on helicopter accidents in Europe, and has created additional safety guidance and toolkits, including separate safety management toolkits for complex and non-complex operators, and, more recently, toolkits for risk assessment and SOP development. It has published a comprehensive helicopter flight instructor manual, numerous safety leaflets, and a report on the safety benefits of technology, and it has integrated closely with the European Aviation Safety Agency and its European Aviation Safety Plan.
Although the IHST has gained less traction in other parts of the world, its global efforts are growing, notably in Brazil. There, a Brazilian Helicopter Safety Team (originally called IHST Brazil) has performed its own accident analysis, translated safety materials produced by the USHST and EHEST, and presented at a number of seminars and conferences.
Nevertheless, there’s a strong sense by many people both within the IHST and outside of it that more needs to be done–that the organization hasn’t come close to accomplishing what it was supposed to. In November of last year, the organization issued an upbeat press release pointing to a 54 percent reduction in the U.S. helicopter accident rate compared to the 2001-2005 baseline rate, but this claim doesn’t really hold up to scrutiny. In the U.S., accident rates are typically calculated using data from the FAA’s annual General Aviation and Part 135 Activity Survey; according to Roy Fox, Bell Helicopter knew from its own tracking of flight hours by individual aircraft that the survey historically underestimated helicopter flight hours. That resulted in an artificially high baseline helicopter accident rate, since the number of accidents was being divided by a lower-than-actual number of flight hours.
“We knew it was wrong, but that’s what public perception was, so the public perception rate was used as our IHST starting point/baseline to determine the 80 percent reduction goal rate,” Fox said.
After the creation of the IHST, helicopter manufacturers began supplying the team with more accurate flight hour data. Along with an actual decrease in the total number of accidents, this contributed to the nearly 45 percent drop in the U.S. accident rate reported for the first year of the IHST’s existence (from the baseline of 7.97 per 100,000 flight hours, to just 4.47 per 100,000 flight hours in 2006). Since then, the rate has gone up and down, without showing a sustained decline. The more optimistic members of IHST point to an overall decrease in the number of U.S. helicopter accidents as additional evidence of improving safety–the annual total dropped from a high of 205 in 2003, to 138 in 2014 (and, according to preliminary data from the NTSB’s Aviation Accident Database, just 119 in 2015).
But this argument does not fully consider two significant events that coincided with the IHST’s existence: the global financial crisis of 2007-2008, and a sustained plunge in the price of oil over the past year-and-a-half. Both events severely impacted industry activity, suggesting that our good luck in safety may have as much to do with bad economic luck as anything else.
More troublingly, the number of fatal accidents hasn’t seen the same reduction. Indeed, fatal helicopter accidents peaked in U.S. fiscal year 2013, when 37 fatal accidents resulted in 74 fatalities–the highest fatal accident count since 1994, according to the FAA. This trend prompted the FAA to refocus its own efforts on fatal helicopter accidents, and it pressured the IHST to follow suit.
In May 2014, the IHST announced that it was asking its worldwide partners to “establish an additional focus regarding the steps that can be taken to prevent fatalities in helicopter accidents.” The IHST said its future analysis would include the types of accidents that most frequently result in fatalities, and the most cost-effective measures to improve helicopter crash survivability.
It’s evident that the IHST would have prioritized different interventions if it had focused on fatal accidents from the get-go. For example, because flight training accounts for a high number of total accidents, the U.S. JHSAT made a specific recommendation for “CFI judgment and decision making training to follow student more closely.” But the training sector is typically not among the sectors that account for the most fatal accidents; according to the FAA, from 2009 to 2013 in the U.S., those were personal/private, commercial, helicopter emergency medical services (HEMS), law enforcement, and aerial application. None of the JHSAT’s “Top 20” implementation recommendations were specific to those sectors.
With respect to more general recommendations, “willful disregard for rules and SOPs” and “insufficient employee performance monitoring” were standard problem statements that the JHSAT identified much more frequently in fatal accidents than in non-fatal ones. Even more common in fatal accidents was the “post-crash fire” SPS. Although crash-resistant fuel systems have been mandated in all newly certified rotorcraft in the U.S. since 1994, many new-build helicopters with older type certificates still lack them. Likewise, many of these aircraft fail to comply with crashworthiness standards implemented in 1989 that were designed to reduce the risk of blunt force trauma–which an FAA study identified as the primary cause of helicopter fatalities, even in accidents with post-crash fires.
Although the FAA announced the creation of a Rotorcraft Occupant Protection Working Group in November 2015 to study occupant protection rulemaking, it will likely be years before its efforts lead to meaningful reform. If the IHST had identified post-crash survival as one of its early priorities, we might by now have been further down that road.
What we still don’t know
Perhaps the IHST’s new focus on reducing fatal accidents will finally yield a real decline in helicopter fatalities. But it’s not clear that, in the continued absence of better data, shuffling around standard problem statements from the JHSAT’s original analysis will do much to take safety to the next level. Although the FAA said last year that its Rotorcraft Standards staff is working on a methodology to accurately identify the rotorcraft accident and fatal rotorcraft accident rates, that’s just a piece of the information we’re missing. When it comes to the reasons why helicopters fall out of the sky, or fly into mountains, or get tangled up in power lines, we still have a long way to go.
On the night of Nov. 29, 2013, an Airbus Helicopters EC135 T2+ operated by Bond Air Services on behalf of Police Scotland crashed through the roof of the Clutha Vaults Bar in Glasgow’s city center. The pilot and two police observers on board the helicopter were fatally injured; so were seven people in the bar. Another 11 people in the bar were seriously injured. The tragedy left Scotland reeling, and the United Kingdom’s Air Accidents Investigation Branch (AAIB) immediately launched a major investigation.
A few months into the investigation, it emerged that the aircraft had crashed with 76 kilograms (167 pounds) of fuel in its main tank, but that the double-celled supply tank that fed the engines had run dry as a result of the fore and aft transfer pump switches being set to the OFF position. Meanwhile, investigators found the adjacent prime pump switches–which are normally used for starting–set to ON. However, the aircraft was not equipped with a flight recorder, and further investigation could not suggest why the pilot might have switched the transfer pumps off. Neither could it be determined why he acknowledged and ignored repeated low fuel visual and audio warnings (as recorded in the aircraft’s non-volatile memory) while he proceeded with routine surveillance tasks.
The investigation did discover that the left engine flamed out just 32 seconds after the right engine flamed out, rather than the three to four minutes predicted by the manufacturer. That element of surprise may help explain why the pilot had difficulty entering and sustaining autorotation, and why he did not transfer his SHED BUS switch from the normal to the emergency position. Consequently, he descended without a functional landing light or radar altimeter, both of which would have helped him judge the appropriate height at which to flare. Evidence indicated that he applied a large amount of collective control while the helicopter was still high above the ground, after which the main rotor r.p.m. became unrecoverable and the aircraft dropped like a rock.
Even so, why did he let it get that far? As the AAIB report put it: “The investigation could not establish why a pilot with over 5,500 hours flying experience in military and civil helicopters, who had been a qualified helicopter instructor and an instrument rating examiner, with previous assessments as an above average pilot, did not complete the actions detailed in the Pilot’s Checklist Emergency and Malfunction Procedures for the LOW FUEL 1 and LOW FUEL 2 warnings.”
The one consolation that many aircraft accidents offer is the hope of preventing similar accidents in the future. The Clutha crash is deeply unsatisfying in this respect. Would better training in aircraft systems, procedures, or crew resource management have prevented the crash? What about changes to the aircraft’s indicating system, electrical system, or switch positions? We can hazard some guesses, but we can’t say for sure–and this is after an investigation that commanded far more time and resources than the typical helicopter accident. The AAIB made seven safety recommendations as a result of the crash, and five of them concerned the need for flight recorders.
The IHST’s “data-driven” approach to reducing helicopter accidents sounds like a good idea, but without high-quality data, it can only take us so far. In looking ahead to the next 10 years, it seems unlikely that we will make much safety progress until we make significant headway with the U.S. JHSAT’s first two recommendations: installing cockpit recording devices in helicopters, and improving the quality and depth of helicopter accident investigations.
Likewise, while the IHST’s new focus on fatal accidents also sounds like a good idea, the underlying assumptions of that approach deserve closer examination. For example, Roy Fox has argued against tracking the fatal helicopter accident rate, because it gives the same weight to events with very different losses to society; a crash that kills one person makes the same contribution to the fatal accident rate as one that kills 16. A better measure of risk, he suggests, would be an occupant’s risk of fatal injury per 100,000 hours of exposure. And, what about serious injuries? Surely we care just as much about people who are horrifically burned or paralyzed in helicopter crashes, but the significance of their injuries gets lost if we only track fatalities.
Finally, perhaps it’s time to stop treating the helicopter industry as monolithic, and start making thoughtful choices about what safety enhancements are appropriate for each sector. No one deserves to die in a helicopter crash, but when a pilot steps into a helicopter, he or she is generally well informed about the risks. The same can’t be said for passengers without aviation experience, or, to an even greater extent, air medical patients who don’t have any choice in the matter. Aviation regulations generally recognize that these occupants deserve greater protections, and our accident reduction targets for passenger-carrying operations should arguably be higher. However, we can’t make much progress toward this goal until we can accurately track flight hours by sector.
In a 2015 audit report by the U.S. Office of Inspector General (OIG), the OIG noted that without comprehensive data for the helicopter emergency medical services sector, the “FAA cannot assess whether the accident rate is increasing or decreasing each year, develop HEMS-specific accident reduction efforts, or determine if its previous safety efforts have been successful.” Given that many sectors of the industry are fiercely competitive, it may take regulation to drive better data collection. In the HEMS sector specifically, Christopher Eastlee, president of the Air Medical Operators Association and USHST interim industry co-chair, said, “We feel that only through a broad federal requirement can we ensure accurate, reliable data that captures all aircraft transporting patients in the U.S.”
Despite the many ways in which the IHST has fallen short of its goals, it still has great potential as a broad-based, well-networked organization that has united hundreds of people who are passionate about helicopter safety. It may therefore remain an appropriate vehicle for advancing safety in the industry, if it can learn from its disappointments over the past decade.
“I remain optimistic,” said Tom Judge, executive director of LifeFlight of Maine, who through his activities in the USHST’s infrastructure working group is helping address some of the structural issues that were not prioritized in the original JHSAT analysis. “Every journey starts with intent and a single step forward. The other way to look at this is, if we hadn’t started this journey, what would [safety in the helicopter industry] look like? . . . I really, honest-to-God believe that if we hadn’t done this, it would have been worse.”
Notice a spelling mistake or typo?
Click on the button below to send an email to our team and we will get to it as soon as possible.