Construction Safety

In Construction, Stop Counting and Start Correcting

TJ Lyons | June 1, 2021

On This Page
Person using a calculator

Edwards Deming said, "Without data, you're just another person with an opinion." He was correct. Before jumping into a pool, you stick your toes in, and that's all you need. Too cold, and you stay out of the water. You don't need to grab a thermometer to confirm what you know. But in construction, there is a tendency, almost a preoccupation, to collect data, lots of data.

Typically, it's not what is counted but how many times we confirmed what we were looking for. So, let's be honest about the following.

  • We continue to collect the same data to be completely sure what we see and what we know can be proved by numbers.
  • We forward the data we know is hurting people and ask others for comments while people are still being injured.

Once, I noted from my field inspections (what I saw, at least) workers that, even when tied off, would still hit the ground if they fell. I developed a solution for that (the Do the Math program), then I asked our teams to spend the next 30 days looking only for this condition. From that, we developed one set of data. Some 73 percent of workers who said they were protected would have fallen and hit the ground while still attached to their "lifeline." We then acted on that data and spent a few months coaching everyone we found working 10 feet from the ground tied to 20 feet of a lifeline. We simply stopped the worker, grabbed a Sharpie and piece of wood to draw on, and went over the math with them. This "Do the Math" effort provided a quick focus, some "No kidding!" moments, and the hazard went away. We acted on what we saw and what we had measured.

The focus on falls is a great example of developing data and acting on it, but it took the relationship to fatalities to drive the effort. From a national Occupational Safety and Health Administration (OSHA) and industry emphasis on falls starting in 2012, fatal falls are now declining across the construction industry. Acting on what we see and know is why we're seeing this improvement.

Bar chart - "Number and rate of fatal falls to a lower level, construction, 2003-2018"

The construction industry needs to rethink how comfortable we are with data and start getting uncomfortable with conversations we need to have. On my wall hangs this quote from Exploring the Dangerous Trades by Alice Hamilton. Early in the 20th century, Ms. Hamilton was the mother of industrial safety. In her autobiography, she laments about her tendency to avoid speaking about the dangers she discovered to the owners of the factories she visited. She shares a friend's advice: "So often, when I have succeeded in breaking down the hostility of an employer and in establishing a friendly relation with him, I have been tempted to let it go at that, to depart without risking unpleasantness. Then I have remembered Julia Lathrop [a fellow social reformer] and have forced myself to say the unpleasant things which had to be said."

Now is our time to have those conversations. Act on what we know, stop collecting more data confirming what we already know, and make those changes we are reluctant to offer up. A few examples follow.

Energized Electrical Work

Our tolerance for working on active electrical systems always surprised me. I cannot open my dishwasher if it is powered up. A light indicates it's on, and a lock protects me from me. Rather than coordinate alternate power supplies or find the time when we can just shut off the electricity, the industry has devised a system of permits to allow "hot work." We make allowances for working on electrified things.

The construction industry needs to step back and consider the current approach where we expect workers to prepare to be hurt. Rather than prohibiting working where they could be exposed to burns, shock, or electrocution, we have standards and programs that require a worker to don arc shields and arc-resistant gear and (I cannot believe I am typing this) wearing clothing that will not melt. Consider the following language of this OSHA publication, 1910.269 App E.

1910 Subpart R Subpart Title: Special Industries Standard Number: 1910.269 App E

Title: Protection From Flames and Electric Arcs. Appendix E to § 1910.269-Protection From Flames and Electric Arcs

I. Introduction

Paragraph (l)(8) of § 1910.269 addresses protecting employees from flames and electric arcs. This paragraph requires employers to: (1) Assess the workplace for flame and electric-arc hazards (paragraph (l)(8)(i)); (2) estimate the available heat energy from electric arcs to which employees would be exposed (paragraph (l)(8)(ii)); (3) ensure that employees wear clothing that will not melt, or ignite and continue to burn, when exposed to flames.…

Energized systems are recognized as high-hazard work, and few areas are scrutinized more closely; yet, that approach is not working. How is this related to data? Do we need to keep confirming our error?

Bar chart - "Number and rate of electrocutions in construction, 2003-2015"

I was once asked to review a plan to replace a large conductor powering a portion of a hospital. The existing energized line was installed in a WWII-era tunnel and was in poor shape, and the replacement was to run alongside it. The existing conductor was so deteriorated; we were to install fire-resistant blankets over the existing lines in case it arced as we ran the new line. No concerns were noted about the danger to the team covering the lines they expected to arc. We got on a call, and I said, "No, we are not willing to take that chance!" I said the client needed to find another way that would not put the workers at risk. Within a few days, a generator was found, a connection was found to bypass the failing line, and work in the tunnel was completed, thus eliminating the risk.

When You Have the Data

My second example shows the value of first collecting some data, acknowledging what we already knew (it hurts to fall off a ladder), and act on it. Below was forwarded by my boss Tony O'Dea summing up the concern.


As if we need more stats to tell us that ladder can be dangerous if not used and set up properly, here are more:

From the Centers of Disease Control and Prevention, a study by PhDs Socias, Menendez, Collins and Simeonov, examined ladder falls and discovered that, while falls remained the leading cause of death, 43% Of fatal falls in the last decade have involved a ladder. Among workers, 20% of all fatal falls involved ladders. Among construction workers, 81% of fall injuries treated in US emergency rooms involved a ladder.

Slowly, the construction industry has recognized the dangers that ladders bring to a worksite and moved to alternatives like mobile elevating work platforms; we are seeing the trends. There has been a nice decline in fatalities from ladders. A recent headline from the Centers for Disease Control and Prevention (CDC) confirms what we see (not measure) in the field—fewer ladders.

The National Campaign to Prevent Falls in Construction is showing signs of success. Construction still has the most fatal falls out of all industries and represents 51 percent of all falls nationally. However, in the past 2 years, we have seen a decrease in many of the types of falls. While fatal construction falls increased in 2014 and 2015, reaching a peak in 2016, there has been a steady decline since 2017 in fatal falls from: structures and surfaces, scaffolds, staging, ladders and roofs.

Scott Breloff, PhD; Elizabeth Garza, MPH, CPH; Scott Earnest, PhD, PE, CSP; Alan Echt, DrPH, CIH; Christina Socias-Morales, DrPH; Jeanette Novakovich, PhD; "Stand-Down for Falls in Its 7th Year: Fatal Falls Are Falling," National Institute for Occupational Safety and Health, CDC, August 10, 2020.

When you have data that shows "43 percent of fatal falls in the last decade have involved a ladder," you don't inspect the ladders more closely or choose another type of ladder; you eliminate ladders. When I was involved in coordinating the Ladders Last effort in the United States, I had called the idea "Ladders No More," and one of the executives in the firm said, "You cannot say that. What about the companies that make ladders?" I replied, "IBM used to produce typewriters."

The Case of Collecting and Using the Wrong Data

I often pose this question to other safety professionals: "Would you consider a small firm with two OSHA recordables in a year an unsafe firm?" The answer is typically no. I counter, "But what if one of the recordables was a fatality?"

Industries use OSHA recordable rates developed in the 1970s to gauge a firm's safety performance. To put this in perspective, at that time, we also used carbon paper and laced up our leather ski boots to our wooden skis. Recordable rates are used to award firms, award work, and bar contractors from bidding work, and most firms display their current rates as indicators of their safety success. This is as traditional as using Herbert William Heinrich's injury pyramid to determine when the next injury will occur—another data source that is in question.

A few decades ago, Professional Safety Magazine published an article about the practice of posting: "It has been __ days since our last accident" at a factory. The author determined it was a useless indicator of safety. That comment was based on the presence of chance and probability.

In a 2020 study from the Construction Safety Research Alliance, the authors used 17 years of data and 3.2 trillion worker hours stated to determine "there is no discernable association between Total Recordable Incident Rates (TRIR) and fatalities" and "in nearly every practical circumstance, it is statistically invalid to use TRIR to compare companies, business units, or teams." It's a great read. They recognize the limits of the TRIR and clearly found some firms simply have insufficient manhours to "return a meaningful TRIR." Using data to question tradition—love that. Is there a more meaningful gauge of how a company is doing? Lost time is a good indicator, but I suspect it's a mix of the following.

  • The number of inspections by operational folks where significant findings are reported, not missing safety glasses but missing railings
  • The number of inspections where unsafes are reported. Again, not if someone is wearing gloves, but if they have their hand in a blender.
  • The number of safety manager inspections and where unsafes are reported
  • The number of lost work days
  • A review of incidents—and the energy behind it. Height, mechanical, tool, etc. speaks to severity and my energy friend, Dr. William Haddon Jr., and his energy release theory.

Finally, here's a comment on data collected during inspections and a suggestion to wrap this up. There is an unspoken and hidden bias during inspections by on-site staff. We are reluctant to admit our safety efforts did not meet the mark, but firms still rely on that daily data source. Many safety managers will do what they can to avoid identifying and documenting any life threat on their project. That's human nature—the need to protect ourselves while correcting unsafes—but the data we need will be skewed. The severity will be downplayed, and that killing condition remains unnoticed.

My suggestion is also this challenge to my readers. Move toward independent and comprehensive inspections by someone unrelated to that site. Tracking only the amount of potential life threats found and any significant contributor like lifting straps that could fail during use or a crawler crane just a bit off-level in the mud. That is why when the insurer stops by for an inspection, they must be allowed to free-range, and these professionals must be brutal in their inspections. Every little thing can be a big thing, and the search must start with that philosophy. If you remove the numerous contributors to an accident, it will not occur.

The next time you head out to the field to gather some data, look at the little things that in perfect alignment will kill. It won't be if someone is wearing their gloves; it will be the upside-down cable clamp that may fail if someone falls against that safety cable or a stop sign missing where the site trucks exit across the public crosswalk.

We have all the data we need, with the exception of perhaps the nano-sciences. We need to focus on the data we know that reflects the OSHA Fatal Four Hazards: falls, electrical exposure, struck-by, and caught-in/between situations. Each involves a killing condition, and we have the data to confirm the threat, and, yes, energy will always be a contributor to the incident. Not just electricity but passing vehicles, the distance you fall, and things that crush you.

If we spend more time making the corrections rather than count the observations, we will succeed.

Opinions expressed in Expert Commentary articles are those of the author and are not necessarily held by the author's employer or IRMI. Expert Commentary articles and other IRMI Online content do not purport to provide legal, accounting, or other professional advice or opinion. If such advice is needed, consult with your attorney, accountant, or other qualified adviser.