Elimination is the most effective chemical hazard control. It means completely removing the hazardous chemical from the workplace so no one is exposed to it at all. This ranks at the top of the hierarchy of controls, a five-level framework developed by the National Institute for Occupational Safety and Health (NIOSH) that guides how workplaces should approach any hazard, chemicals included.
The five levels, ranked from most to least effective, are: elimination, substitution, engineering controls, administrative controls, and personal protective equipment (PPE). The higher you go on this hierarchy, the less you rely on human behavior to keep people safe, and that’s what makes those top-tier controls so much more reliable.
Why Elimination Ranks First
Elimination works because it removes the source of danger entirely. If a chemical isn’t present, it can’t cause harm. There’s no exposure limit to monitor, no protective gear to maintain, no training required to handle it safely. A manufacturing facility that finds a way to drop a toxic solvent from its process altogether has solved the problem permanently for every worker on every shift.
The reason this matters so much comes down to reliability. Engineering controls can malfunction. Administrative procedures can be skipped. PPE can be worn incorrectly. But a chemical that no longer exists in the workplace cannot hurt anyone under any circumstances. That’s why both NIOSH and OSHA place elimination at the top of the pyramid: it requires zero ongoing human interaction to remain effective.
The challenge, of course, is that elimination isn’t always possible. Some chemicals are essential to a process and can’t simply be removed. When that’s the case, the hierarchy guides you to the next best option.
Substitution: The Next Best Thing
Substitution means replacing a hazardous chemical with one that’s less dangerous. A workplace might switch from a toxic cleaning solvent to a non-toxic alternative, or swap latex gloves (which cause allergic reactions in some workers) for a non-latex material. The EPA identifies substitution as one of four “inherently safer” design approaches, noting that it can potentially eliminate the underlying hazard altogether.
The line between elimination and substitution is sometimes blurry. If you replace a carcinogenic degreaser with a water-based cleaner, you’ve eliminated the carcinogen and substituted a safer product in one move. What matters is that both approaches address the hazard at its source rather than trying to manage exposure after the fact.
Substitution does require careful evaluation. A replacement chemical needs to actually be safer overall, not just trade one risk for another. The EPA recommends considering not just toxicity but also flammability, reactivity, and the conditions under which the substitute will be used.
Engineering Controls: Building Safety Into the Environment
When you can’t eliminate or substitute a chemical, engineering controls are the next most effective option. These are physical changes to the workplace that reduce exposure without requiring workers to do anything differently. Common examples include local exhaust ventilation (like fume hoods that capture chemical vapors at the source), enclosed systems that prevent chemicals from reaching the air, and automated processes that keep workers physically separated from hazardous substances.
Ventilation systems are among the most widely used engineering controls for chemical hazards. A properly designed local exhaust system pulls contaminated air away from the worker’s breathing zone before it can spread into the room. General ventilation, which dilutes airborne chemicals by circulating fresh air through the space, serves as a secondary layer but is less effective than capturing vapors directly at the source.
Engineering controls share an important trait with elimination and substitution: they work without relying on worker behavior. A fume hood protects everyone in the lab whether or not they remembered their safety training that morning. That’s a fundamentally different kind of protection than reminding someone to wear a respirator.
Administrative Controls: Changing How Work Gets Done
Administrative controls don’t remove the hazard or block exposure physically. Instead, they reduce how much contact workers have with dangerous chemicals by changing procedures, schedules, or information. Examples include rotating workers through tasks so no single person accumulates excessive exposure, scheduling high-exposure work during shifts with fewer people present, posting warning signs near chemical storage areas, and training employees on safe handling techniques.
OSHA groups several specific measures under this category:
- Procedures: equipment inspections, preventive maintenance checklists, lockout/tagout protocols
- Training: hazard communication, confined space entry, safe work procedures for chemical handling
- Warnings: signs, labels, alarms, and written instructions
These controls are useful, but they’re inherently less reliable because they depend on people following rules consistently. A warning sign only works if someone reads it. A rotation schedule only works if a supervisor enforces it. That’s why OSHA recommends using administrative controls together with higher-level controls rather than as a standalone solution.
Why PPE Is the Last Resort
Personal protective equipment, including gloves, respirators, goggles, and chemical-resistant suits, sits at the bottom of the hierarchy. It’s the most common form of chemical protection in many workplaces, but it’s also the least effective because it places the entire burden of safety on the individual worker.
Research on PPE failure modes reveals just how many things can go wrong. A human factors study analyzing PPE removal procedures identified 103 distinct ways the process could fail. The highest-risk failures involved people moving between clean and contaminated areas, removing gloves, and removing protective aprons. Contributing factors ranged from individual body size (workers with shorter arms had more difficulty keeping contaminated gear away from their body during removal) to environmental constraints (small changing areas made it more likely for people to bump into each other and spread contamination).
Even the design of the equipment itself introduces risk. PPE that’s too large or too small for a particular user is more likely to have undetected breaches. Longer hoods increase contamination risk. Aprons that must be pulled over the head rather than torn away create additional exposure opportunities during removal. These aren’t hypothetical problems. They’re documented failure patterns that occur even in controlled settings with trained personnel.
PPE also requires ongoing costs: purchasing, fitting, replacing, training, and monitoring compliance. A pair of chemical-resistant gloves degrades over time and with repeated chemical contact. Respirator cartridges expire. Fit testing needs to be repeated. None of these issues exist with elimination, and they’re minimal with good engineering controls.
Combining Controls for Real-World Protection
In practice, most workplaces need a layered approach. OSHA explicitly states that a combination of control methods often provides the best level of protection. A chemical processing facility might substitute a less volatile solvent (substitution), install local exhaust ventilation over the work area (engineering), train workers on safe handling procedures (administrative), and provide chemical-resistant gloves for tasks involving direct contact (PPE).
The hierarchy isn’t meant to suggest you pick one level and stop. It’s a prioritization tool. You start at the top, implement the most effective controls you can, and then layer in lower-level controls to address whatever residual risk remains. The goal is to push as much of your protection as possible toward the top of the hierarchy, where it doesn’t depend on a worker remembering to zip up a suit correctly at 3 a.m. on a double shift.
The EPA frames this as a spectrum of reliability. Inherent approaches like elimination and substitution are “generally highly reliable” because they address the hazard itself. Administrative safeguards tend to be less reliable because they depend on consistent human performance, something that varies with fatigue, training quality, staffing levels, and dozens of other factors that no workplace fully controls.