In a new study of industrial sites, 63% said fewer than one frontline worker in 20 reports a hazard each month, or that they do not track it at all. Recordable injury rates are at a series low, but the leading indicator that should warn of the next one is running close to empty.
By the measure most boards watch, industrial safety has never looked better. Private employers recorded 2.3 injuries and illnesses per 100 full-time workers in 2024, down from 2.4, and their 2.5 million reported cases were the fewest since the series began in 2003. But recordable cases are a lagging indicator: they count harm that has already happened. The measure meant to warn of the next serious incident is what workers see and report before anyone is hurt, and new research suggests that at most sites it barely registers, and at the largest ones nobody can say what it is.
Teamforce AI's Frontline Risk Study 2026, released on October 29 and developed with the University of Iowa's Vaughan Institute of Risk Management & Insurance, asked 60 professionals across manufacturing, chemicals, pharma, food, energy, construction and agriculture how many frontline workers submit a safety, hazard or near-miss report in a typical month. 38% put it below 5%. A further 25% said their site does not measure it. Only 22% put participation above 30%.
The low numbers are also stuck. Of the 23 sites below 5%, 19 said participation had not changed in twelve months. The full study is candid about its limits: the survey was voluntary, so it probably leans towards safety-engaged respondents, and figures based on all 60 carry a margin of about 13 points either way. Even at the low end of that range, about half of sites would have thin or unmeasured reporting. As Jim Lewis, the Vaughan Institute's executive director, put it in the release, "You can't predict and prevent what your workers aren't telling you."
The most uncomfortable finding concerns scale. Half of respondents at sites with more than 1,000 employees, and 57% of those in corporate or multi-site roles, could not state their reporting rate. Combined, 53% of large or corporate respondents could not give a figure, against 12% at smaller sites. The organisations with the most people exposed to risk, and usually the most mature EHS functions, are the least able to say whether those people are speaking up.
Technology helps, but not on its own. Among sites that measure reporting, half of those with a digital channel such as an app or EHS software (15 of 30) exceeded 15% participation, while none of the 15 without one did. Yet across all 39 sites with these tools, 20 were still under 5% or not measuring. Closing the loop is a separate weakness: 36% of sites with corrective-action software and 30% without named verifying that fixes actually worked as a top need, a gap the authors found was not significant. Teamforce chief executive Vivek Kumar summed up the problem: "a closed ticket shows someone acted, not that the fix worked."
Other recent research explains why a reporting channel alone does not fill up. A TRADESAFE survey of 503 blue-collar workers, reported in July, found that one in four had witnessed a near miss in the past year and not reported it, including 33% in manufacturing. 41% of manufacturing respondents cited fear of retaliation. The most common reason for staying silent overall was a belief that nothing would change (36%), and only 56% agreed that reporting a safety issue would lead to real action. The trust gap runs along the org chart: 66% of managers said they felt very safe speaking up, against 36% of frontline workers.
A University of Alberta study of nearly 4,700 mining workers and contractors, covered by Safety+Health in February, found the same pattern underground. A novice, developing workforce was the leading reason hazards went unreported (18%), followed by fear of career repercussions (15%) and social repercussions (12%). Temporary and contract workers worried about losing their jobs, future positions or overtime. Lead author Lianne Lefsrud said workers "described being more afraid of the social consequences of reporting hazards than of the hazards themselves."
That matters for how plants read their own data. A falling recordable rate and a near-empty hazard log can look like the same good news. They are not. A site where few people report may simply be a site where few people feel able to, and its contractors, newest starters and frontline crews are the people the studies above suggest are most likely to stay silent.
These steps turn hazard reporting from an assumed habit into a number the plant can manage against.
Industry has spent decades driving recordable cases down, and the 2024 figures show it worked. The next gains depend on hearing about hazards before they become cases, and on most sites that conversation has barely started. Plants that can name their reporting rate, and show workers what each report changed, will know where the next injury is coming from before it arrives.

Guide
Alberta researchers found temporary and contract workers among the most afraid that a hazard report would cost them work. This guide explains how a contractor's safety maturity predicts its performance, and how to tier the firms whose crews may be staying quiet.
Download
Case Study
Only sites with a digital channel cleared 15% reporting participation, yet half of the sites with tools were still under 5% or not measuring. This case study follows how one operator evaluated configurable field software and what changed once crews were using it.
Download
Report
Just 36% of frontline workers feel very safe speaking up, against 66% of managers. This report examines who today's frontline workers are and what they value, the starting point for a reporting culture they will actually join.
Download