Deck Researchers Scrutinize Railing Attachment

As deaths and injuries continue, an engineering team zeroes in on flawed deck details

A Virginia Tech task force examining the critical structural details for residential decks and balconies released detailed data and conclusions about deck-to-house attachments this spring (see "Load-Tested Deck Ledger Connections," 3/04). Now the team is shifting its focus to the other end of the deck. In a process that Virginia Tech professor emeritus Frank Woeste expects to take at least a year, the group will apply its mix of field investigation, engineering analysis, and laboratory testing to railing and rail post connections in hopes of identifying which details will prevent death and injury and which won't.

Each year brings a fresh crop of deck structural failures, occasionally with multiple casualties. Only the most dramatic failures, like last summer's Chicago disaster that killed 13 partygoers, get widespread press attention; no one knows how many deck calamities go completely unreported. But failures that kill or injure people are common enough for Richmond, Va., attorney John Conrad to make part of his living defending the resulting lawsuits, and Conrad has no trouble calling examples to mind. He joined Frank Woeste and Virginia Tech wood science professor Joe Loferski in presenting a three-day seminar titled "Liability Issues, Design Data, and Inspection Techniques for Wood Decks, Balconies and Porches" at Virginia Tech in April.



"There are a lot more cases than people realize," Conrad told JLC in February. "Usually it's the owner who gets sued, because they are usually covered by insurance. That's the economics of a lawsuit. But when the insurance won't cover the damages for any reason, plaintiffs can also turn to the designer or the builder." There's usually little room for argument, adds Conrad: "If a deck falls down and someone is hurt or killed, people expect the owner to pay. It's pretty much time to get the checkbook out."

Loferski sometimes serves as an expert witness in deck-related lawsuits. "I'll be hired by one party or the other, and my role is to help these people understand what actually happened," he explains. He says there's a reason the Virginia Tech team has zeroed in on ledger connections and post and railing attachments: "These are the two components on the deck that, when they fail, people get hurt or killed. I have not heard of anyone falling through a deck board, for instance. We haven't heard of a single column that has collapsed due to buckling or overload. The only ones we hear about are those two: the deck ledger and the handrails."

In contrast to total collapses caused by a failure of the ledger connections, says Loferski, railing failures usually involve only one or two people. "When ledgers fail, it's because a lot of fasteners let go at once. With railings and rail posts, a catastrophic failure can occur when only one or two fasteners fail. A couple of people lean over, and they fall."

News reports about major collapses typically include a bystander's opinion that the deck was overloaded with people, but Loferski gives that notion little credence. "If you look at the design code loads — 40 pounds live load plus 10 pounds dead load — it's hard to make that up with people. Even if you had the whole deck full of 300-pound people, I don't think you could do it."

For a railing, says Loferski, it might be possible in theory to exceed code-specified design loads. "The design code load is 200 pounds in any direction, applied at the top of the railing. In ASCE 7, the document that codes refer to for loads on buildings, there's an additional provision of 50 pounds per lineal foot of railing, and if the posts are 10 feet apart, that would govern. That sentence from ASCE 7 didn't get into the building code; but if you had posts spaced 10 feet apart, and people leaning against the whole length of the rail, it may be that you could overstress a post."

But in the real world, says Loferski, the railing failures he knows about had nothing to do with overloading: The structures were clearly inadequate. "One or two of them might have worked when they were new, but even those were underbuilt. Typically, the railings that failed were less than ten years old, and they tend to be built with untreated wood that has deteriorated in service, and with ungalvanized nails that have rusted. Often people used finish nails or screws that were never adequate for the design loads even before they started to rust."

Even as Loferski spoke, news reports indicated that the 2004 deck failure season was already underway. In late February, seven men brawling at a New Jersey party broke through a railing and fell 25 feet (five went to the hospital and two left the scene, police said). And in DeKalb County, Ga., four adults and two toddlers ended up at the emergency room with minor injuries and charcoal burns after a deck collapsed during a cookout. The six were fortunate, DeKalb County Fire Department captain Eric Jackson told TV reporters on the scene: "They were able to walk away."