Two weeks after crashing his Piper Saratoga plane, Mark Williams left the hospital in a wheelchair and met with federal crash investigators for an interview that felt like an interrogation.
"They kept on berating me, 'How much fuel was in (the plane)? What was your fuel burn that day?' Their questions seemed so accusatory," recalled Williams, an amateur pilot and a dentist from Meridian, Miss.
Williams tried to describe how his six-seater had lost power 1,100 feet in the air. But the investigators had their own ideas. Because there was no fire after the crash, they suspected Williams had burned up his fuel and pressured him to take the blame.
Williams, who suffered a broken pelvis, broken shoulder and a severed ear, insisted there was no fire because the fuel drained out when a tree limb sheared off a wing.
A short time later, Williams' statement was confirmed by a third federal investigator, who found that a failed engine component — not pilot error — caused the June 7, 2002, crash, Williams said. Federal investigators would blame the same component for a New Jersey crash three months later that killed a Canadian couple and critically injured their two sons, ages 8 and 5.
Williams wonders to this day what investigators would have concluded had he not been alive to speak to them.
"Had I not survived, this accident would have been wrapped up as pilot error, fuel exhaustion, and nobody ever would know the difference," Williams said.
Federal accident investigators repeatedly overlooked defects and other dangers of private aviation as they blamed individual pilots for the overwhelming number of crashes of small airplanes and helicopters, a USA TODAY investigation has found. The failure of crash investigators to find defective parts, dangerous aircraft designs, inadequate safety features and weak government oversight helped allow hidden hazards to persist for decades, killing or injuring thousands of pilots and passengers, the investigation found.
The National Transportation Safety Board, which investigates accidents and urges safety improvements, has an international reputation for exhaustive investigations of commercial jet crashes. But its probes of "general aviation," which covers private planes and helicopters, are much more brief, often done by telephone and frequently reliant on manufacturers to find problems with their own aircraft parts.
NTSB investigators visited a crash scene or did "a significant amount of investigative work" in only 15% of the 64,000 general-aviation crashes from 1982 through 2013, according to a USA TODAY analysis. They did significant investigations of 32% of airline accidents, which often involve damage on a tarmac or injuries caused by turbulence. In the Williams case, the NTSB had Federal Aviation Administration safety inspectors conduct the initial interview with Williams.
"The NTSB could do more" to investigate general-aviation crashes, former NTSB chairwoman Deborah Hersman said in a recent interview, "but I think it's a question of asking what does the public want and what are they willing to pay for." The NTSB has given top priority to commercial airline crashes, which it has helped to virtually eliminate. There hasn't been a major airline crash since early 2009.
But the shortcomings in general-aviation investigations have compounded a death toll of nearly 45,000 since 1964 — roughly nine times more than were killed in passenger airlines.
An in-depth review of federal records shows:
• Nearly 1,200 people have been killed since 1985 in lightweight, home-built aircraft — an average of 40 a year — and their crash rate has been as much as four times the rate for other private aircraft. Yet the NTSB blamed 72% of the crashes on pilots and didn't discover until a 2012 special study — after 1,124 deaths — that many of the crashes resulted from engine failure, inadequate flight testing that did not uncover design problems and malfunctions, and flight manuals that failed to sufficiently explain the aircraft.
A series of civil lawsuits contradicted NTSB findings of pilot error and found manufacturers liable after juries or judges spent weeks reviewing records the NTSB didn't have. USA TODAY found 21 verdicts totaling nearly $1 billion against manufacturers that the NTSB exonerated. Several juries ordered companies to pay punitive damages after finding they failed to fix long-standing known defects in small planes or helicopters.
• At least 214 air-ambulance helicopters have crashed since 1982, killing 272 patients, pilots and medical personnel and injuring 175. The NTSB blamed 72% of the crashes on pilots, but a 2006 board study found that many of the crashes were caused by flight dispatchers failing to tell pilots about dangerous weather and hazardous landing conditions, or could have been prevented by safety systems that warn when an aircraft is too close to terrain.
• Crashes of skydiving airplanes have killed 306 people since 1964, including 10 crashes that killed 10 or more people. The NTSB blamed 62% of the crashes on pilots and didn't discover other problems until a 2008 study: Many skydiving airplanes were poorly maintained, pilots were ill-trained for the nuances of skydiving flights, and the FAA was failing to adequately oversee skydiving operators.
Isolated, minor incidents also showed the danger of shallow investigations. After a Cessna Skyhawk landed at Pilot Country Airport near Tampa in 1998 and plowed into a dirt mound at the end of the runway, the NTSB blamed the pilot and said nothing about the mound being so close to the runway.
A year and a half later, another pilot landed his Cessna at the same airport, collided with the same dirt mound, but his airplane flipped over leaving passenger George Gill a quadriplegic.
After Gill's lawyers presented evidence that the mound was closer to the runway than federal regulations allowed, airport operator Red Baron Aviation paid Gill $10 million, court records show. The company removed the mound.
"It's very easy to just plain blame the pilot and say 90% of crashes are caused by pilot error, so we don't really need to worry about other things," said Sue Baker, a public health professor at Johns Hopkins University who specializes in transportation injuries. "Pilot error is going to be part of almost any crash. But the reasons for pilot error may not be things under the pilot's control."
The NTSB says it has found pilots causing or contributing to 86% of general-aviation crashes. The board makes its determinations in one- or two-sentence probable-cause findings, which often cite a pilot's "failure" to do something.
John DeLisi, director of the NTSB's Office of Aviation Safety, said probable cause "is only the tip of the pyramid. To look at an investigation as a whole, there is so much more."
NTSB investigations have missed other general-aviation dangers by focusing on the cause of the crash and disregarding the cause of death or injury and whether people could have survived with better aircraft safety features.
The NTSB seeks little information about the cause of injuries or deaths in general-aviation crashes. Its nine-page aviation accident form asks pilots or aircraft operators to simply check a box indicating each occupants' degree of injury. NTSB reports almost never say anything about a cause of injury or death.
By ignoring "crashworthiness" — how well an aircraft protects occupants in a crash — NTSB investigations have overlooked lethal components such as helicopter fuel tanks that repeatedly ruptured after minor crashes and ignited fires that burned people to death.
A USA TODAY analysis of dozens of autopsy reports and crash records since 1982 found 58 helicopter crashes in which an occupant was unhurt by the impact only to be killed or injured by a post-crash fire. The 80 deaths and 33 injuries might have been prevented if the helicopters had sturdier fuel tanks, which were developed in the 1960s and installed in Army helicopters starting in 1970. The NTSB noted the dangers of such fires in 1980.
But it has since paid little attention to the issue. In 57 of the 58 crashes, NTSB investigators never noted the cause of death or a fuel-tank problem. They instead blamed pilots, maintenance workers or helicopter components for causing the crash. In January, the NTSB noted a problem with Robinson helicopters catching fire after crashes.
The NTSB's disregard of injury causes has alarmed safety experts and angered crash survivors such as Dave Wallace. He was burned over 45% of his body on May 28, 2005, when his Robinson R-44 helicopter erupted in flames after rolling onto its side during a landing in Southern California.
Released after 10 weeks in the burn unit at Arrowhead Regional Medical Center in Corona, Calif., Wallace got a call from NTSB investigator Patrick Jones. "He didn't ask me one thing about the fire," Wallace said, recalling instead accusatory questions about whether his helicopter was carrying too much weight and if Wallace had properly maintained his pilot's log book.
"I felt like I was under interrogation for murder," Wallace said.
Jones' final report notes that the R-44 had "impacted into a dry stream bed" where "the helicopter burst into flames" and that the three occupants "sustained burns while exiting the burning helicopter." But the report blames only Wallace — for "failure to maintain adequate main rotor rpm and directional control."
NTSB spokesman Peter Knudson said the report focused on the pilot, not the fire, because fires "are not at all unusual" and Wallace's R-44 hit the ground with a force that it is not designed to withstand.
In January, the NTSB noted seven R-44 crashes since 2008 in the U.S. and Australia in which 10 people were killed by fire and urged the FAA to require R-44 owners to install components that would help prevent post-crash fires. The FAA said in reply that the R-44 "does not appear significantly different" from similar helicopters in catching fire and will continue to study the helicopter.
Over the past 10 years, safety experts have pushed the NTSB to more precisely determine the causes of general-aviation crashes — and the causes of death — so that defective designs or faulty parts can be fixed.
In 2011, the U.S. Joint Helicopter Safety Analysis Team, an industry/government group seeking to reduce helicopter crashes, urged the NTSB to improve the "quality and depth" of its helicopter-crash investigations.
"Many accidents are not receiving in-depth onsite investigation by NTSB investigators," a team report said. "Investigations are being performed by telephone interview or by personnel whose primary function is not accident investigation." The personnel are FAA inspectors, who often conduct on-scene investigations for the NTSB.
NTSB investigators who miss a crash scene rely on reports from FAA personnel and local police, on weather data, flight logs, phone interviews — and on inspections of an aircraft and its components done by the manufacturers, even if a manufacturer is being sued for the crash.
"Many times what happens now is that when the accident occurs, the technical rep of the (manufacturing) company will call the NTSB and say we'll be party (to the investigation), we'll go out there and let you know what we see … the only people on scene would be perhaps an FAA guy and the field rep of the manufacturer," said Douglas Herlihy, a former NTSB investigator who now reconstructs crashes, often for plaintiffs in lawsuits against manufacturers.
"If you (the NTSB) are not there, you've got the representative from the company at the scene. His job is to skew the facts, to ignore the product difficulties and to remove the question of liability," Herlihy said.
DeLisi of the NTSB said the board's five presidential appointees have the final say in determining the cause of a crash. The FAA, with 3,000 inspectors, can often get to a crash scene more quickly than one of the NTSB's 48 regional investigators, DeLisi said. Manufacturers "work very effectively in providing technical assistance," and their analytical work is overseen by an NTSB investigator.
"We have a very active role in managing the parties," DeLisi said.
The reliance on manufacturers to inspect their parts for malfunctions and defects has long raised concerns about a conflict of interest.
A 2000 report by the RAND Corp. research group said manufacturers are "essential" to aviation investigations and give the NTSB the "technical capability" to determine crash causes. The NTSB has 90 aviation investigators who handle roughly 1,500 crashes a year.
But manufacturers may "threaten the integrity" of investigations, the report added. Some manufacturers' representatives to the NTSB feel beholden to "corporate managers who are equally, if not more, concerned about the potential liability and corporate image problems associated with a major plane crash."
RAND recommended that the NTSB reduce its reliance on manufacturers and make greater use of people with no financial stake, such as government experts, consultants and scholars. That has not happened.
"There are cases where the investigation is less than stellar," said John Goglia, an NTSB board member from 1995 to 2004. "There are some particular problems with general aviation: The NTSB doesn't do all the investigations. Some of them the FAA does for the NTSB."
One consequence of limited investigations has been a series of verdicts in civil lawsuits that contradict NTSB findings. They include:
• $48 million against British parts maker Doncasters, including $28 million in punitive damages, for a 2006 crash that killed six in a deHavilland airplane. A jury blamed defective turbine blades that caused one of two engines to fail. The NTSB could not determine why the turbine blades broke and blamed the pilot for "failure to maintain airspeed" after the engine failure. Doncasters paid the $48 million.
• $89 million against Lycoming Engines and Precision Airmotive, including $65 million in punitive damages, for a 1999 crash that killed four flying in a Piper Cherokee. A jury blamed a defective carburetor that caused engine failure. The NTSB, whose investigative file covers just 40 pages, found no evidence of any malfunction and blamed "the pilot's loss of control." The case settled for $21 million.
• $480 million against Cessna Aircraft Co., including $400 million in punitive damages, for a 1989 crash that injured three. A jury concluded the crash was caused by a defective pilot's seat that slid backward unexpectedly, causing pilot James Cassoutt to lose control of the plane. Cassoutt had told the NTSB the seat slid back, but the NTSB concluded he "inadvertently stalled the aircraft." The case settled for $41 million.
• $26.1 million against Textron Lycoming, including $6 million in punitive damages, for a 2008 crash that killed three and which a judge found was caused by a defective carburetor that caused engine failure. The NTSB assigned Precision Airmotive to analyze the airplane's carburetor, which Precision manufactured. The company found that the carburetor was "completely filled with fuel" — a condition linked to engine failures in the past — but its three-page report to the NTSB says nothing about the history of the problem.
"Why would you not disclose that when the NTSB is relying on you and you know the (NTSB) field-office guy doesn't have a specialized knowledge of carburetors?" said Portland attorney Matthew Clarke, who filed the lawsuit.
The NTSB blamed the pilot's "improper decision" to fly into low-visibility conditions.
Pilots also were blamed — and later exonerated — in numerous cases involving air-traffic controllers. Since 1997, the U.S. government has paid $60 million to settle 41 claims over crashes that the NTSB blamed on pilots but which claimants said were caused by air-traffic controllers, federal records show.
Former Marine Gavin Heyworth was blamed by the NTSB for a 2003 midair helicopter collision that killed two others because of his "failure to comply" with air-traffic-controller instructions. But after a week-long bench trial, a federal judge ruled that FAA controllers failed to keep the two helicopters apart and had acted "negligently and carelessly." The Justice Department settled claims against the FAA over the crash for $10.1 million.
The NTSB finding "was one of the worst things that ever happened to him," said Heyworth's attorney James Pocrass. "He felt responsible for the death of those two people. The vindication he had from the judge was bigger than winning the case."
The NTSB has become increasingly concerned about general aviation as crash and death rates have remained stagnant for 15 years while commercial crashes have practically ceased. In 2012, the NTSB elevated general-aviation safety to its annual list of "most wanted" improvements in transportation safety.
"That was a breakthrough," the NTSB's DeLisi said.
The NTSB, with a staff of 400, has focused since its creation in 1967 on hazards to the greatest number of people in aviation, rail transit, pipelines, waterways and roadways. The NTSB helps investigate 19 foreign airline crashes a year on average, including the recent disappearance of Malaysia Airlines Flight 370. Foreign investigations are "a particular challenge" because of the board's domestic obligations, the NTSB wrote in a recent annual report.
In 2012, the NTSB said that in many general-aviation crashes "pilots did not have the adequate knowledge, skills or recurrent training to fly safely." At the opening of the 2012 forum, Hersman, then the NTSB chairwoman, said, "GA pilots are not learning from the deadly mistakes of their brethren."
The emphasis on pilots and pilot problems has steered the NTSB away from mechanical problems or ways to increase the chances of surviving a crash, some experts say.
"There is essentially no requirement that specific injury data be collected. They don't even collect impact data, like angle (of the crash) or velocity," said retired Army colonel Dennis Shanahan, former commander of the U.S. Army Aeromedical Research Laboratory who has consulted for the NTSB.
"This has severely impeded researchers' ability to find out what's going on from an injury basis," Shanahan added. "If you're interested in a particular aircraft and you want to know what the injuries are, it can't be done."
Harry Robertson, an aviation-safety pioneer who developed stronger helicopter fuel tanks in the 1960s and is in the National Aviation Hall of Fame, said general-aviation deaths cannot be substantially reduced until the NTSB understands their cause.
"The most important thing they can do first is improve the quality of their investigations as to what it was that caused the injury and the death," Robertson said.
And once causes are known, "engineers can step in and figure out the ways to assure an improvement in crashworthy capabilities. … Trying to improve crashworthiness is very hard if you don't know what broke."
The NTSB and its predecessor agencies have recognized for decades that general-aviation crashes are often mild enough for occupants to survive.
In 1980, the NTSB said, "General-aviation aircraft are unnecessarily lethal in crash situations which should be survivable."
To reduce automotive deaths, the National Highway Traffic Safety Administration gathers data on what kills or injures people in crashes. The findings have led NHTSA to mandate a wide range of safety equipment that has reduced the number of deaths in motor vehicles to 21,667 in 2012 from 35,025 in 1988, even as the amount of driving increased 50%.
A 2002 report written for the Department of Transportation by Sue Baker, the Johns Hopkins expert, urged the NTSB to follow NHTSA's approach and collect detailed information on a random sample of general-aviation crashes. That hasn't been done.
Shanahan said the NTSB was interested when he was consulting there in 1989, but money was not available.
The NTSB has instead conducted in-depth studies on specific general-aviation aircraft and flying conditions that are causing extensive crashes, such as home-built airplanes and medical helicopters. The studies have found causes and patterns that were not found during individual crash investigations and have led to safety recommendations and improvements.
But the NTSB safety recommendations come years or decades after a problem arises.
In May, an NTSB report noted "undue hazards and risks" of agricultural flights such as pilot fatigue and poor aircraft maintenance and urged safety improvements.
The report did not note that since 1982, 4,200 agricultural airplanes and helicopters had crashed, killing 394 people and injuring 1,224.
In two-thirds of the crashes, the NTSB blamed the pilot.