Confined space fatalities are still happening at a rate that should embarrass our industry. OSHA reports that approximately 90 workers die in confined spaces every year in the United States — and the majority of those deaths involve spaces that had written programs. The program existed. The entry happened anyway. People died.
That tells you the problem isn't a lack of paperwork. The problem is the gap between what the written program says and what actually happens at the entry point.
I've walked into hundreds of facilities across oil & gas, mining, electrical, and industrial construction. OSHA has too. The violations we both find concentrate in three areas with remarkable consistency — and none of them are complicated to fix once you know where to look.
If your confined space program has never been tested against a real entry — permit pulled, atmospheric testing documented, attendant stationed, retrieval equipment staged — it isn't a program. It's a binder.
Gap #1: Atmospheric Testing That Doesn't Match the Hazard Profile
The most common citation under 1910.146 isn't about skipping atmospheric testing — it's about testing for the wrong things. A facility tests oxygen levels and calls it done, when the space has a documented history of hydrogen sulfide accumulation, or sits adjacent to a petroleum storage area with LEL concerns. The tests don't match the identified hazards.
Your atmospheric testing protocol needs to flow directly from your hazard identification for each specific space. If the space contains or connects to anything that produces H₂S, you test for H₂S. If there's any combustible material history, you test LEL. Oxygen deficiency/enrichment testing is the floor, not the ceiling. Build a space-specific testing matrix and put it on the permit.
The second atmospheric testing failure is sequencing. Workers enter before testing is complete, or they test at the top of the space and call the bottom safe. Atmospheric conditions stratify. Heavier-than-air gases settle. Carbon monoxide and oxygen-deficient conditions can exist at the work level while the reading at the entry point looks clean.
Your program needs to specify testing locations — top, middle, and bottom — before entry is authorized. That requirement needs to be on the permit itself, not buried in a procedure document nobody reads at the opening.
Gap #2: Attendant Duties That Exist Only on Paper
OSHA 1910.146(i) spells out attendant duties clearly. In practice, those duties routinely collapse to one thing: standing near the hole. The attendant isn't maintaining accurate counts of authorized entrants. They don't know how to recognize behavioral effects of exposure. They haven't been trained on non-entry rescue procedures. And in most cases, if something went wrong, they would enter the space to help — which is how body counts become multiple fatalities.
Attendant training needs to be documented and specific — not a checkbox on a general awareness course. Attendants must know the specific hazards of the specific space they're assigned to. They must practice accounting for entrants, know the emergency numbers, understand the rescue plan, and be trained and drilled on non-entry rescue. If your attendant would go in after a downed worker, your rescue plan isn't a rescue plan.
In 25 years I have never seen a successful non-entry rescue from a space where the attendant wasn't trained specifically on that procedure before the entry. Train your attendants on retrieval equipment — not just where it's stored, but how to use it under stress without entering the space.
Gap #3: Rescue Plans That Can't Execute
The most dangerous confined space rescue plan I see regularly is one that depends entirely on local emergency services as the primary response. OSHA requires that rescue services be available — available means capable of responding in time. In remote oil & gas locations, rural mine sites, or large industrial campuses, local EMS response time may be 20-40 minutes. A victim with compromised oxygen or H₂S exposure doesn't have 20-40 minutes.
Your rescue plan must be realistic for your actual location and your actual hazards. If EMS can't respond in time, you need on-site rescue capability — trained, equipped, and practiced. OSHA allows retrieval systems as the primary rescue method where non-entry retrieval is feasible. Build your program around that. If entry rescue is required, you need people trained to NFPA 1006 or equivalent, with equipment staged and accessible, not locked in a trailer across the facility.
What To Do Before OSHA Shows Up
- Inventory every confined space on your site — including spaces that may not be classified as permit-required. Classification errors carry their own citation risk.
- Build space-specific hazard profiles and confirm your atmospheric testing protocol addresses each identified hazard, not just the standard ones.
- Audit your attendants by observation, not paperwork — watch them work an entry and assess whether they're actually performing the duties the program requires.
- Time your rescue response from an actual entry point to when equipment would be staged and rescue-ready. If that number is uncomfortable, your rescue plan needs revision.
- Run a tabletop or drill annually at minimum — not just initial training. Confined space emergencies are rare enough that skills degrade without practice.
These aren't advanced concepts. They're the basics — executed consistently, documented accurately, and verified by people who have actually seen what confined space incidents look like in real life.
Need a Confined Space Program Review?
Texas Sentinel Safety Group conducts compliance gap analyses that identify exactly where your program falls short — before OSHA does. Field-experienced, documentation-ready, actionable.
Request a Compliance Audit