Safety ceilings must be set for the most sensitive population

What it is

A safety ceiling calibrated to a healthy adult's lab pain threshold is not safe for a meaningful share of real users — any tactile or thermal output ceiling has to be set for the most perceptually vulnerable segment of the target user base, not for the participants a typical usability test recruits. These two groups often barely overlap at all: usability testing tends to recruit healthy, attentive, cooperative adults — exactly the lowest-risk end of the population.

Why it happens

Several independent factors can raise a given user's actual injury risk without the device's output changing at all. Peripheral neuropathy (common in diabetes, and in some older adults) blunts pain and temperature warning signals, letting damage accumulate quietly before it is felt or reported. Thinner, more fragile skin (in infants, older adults, or under certain conditions or medications) means the same force or temperature reaches tissue-damaging levels at a lower intensity than in the healthy young adult skin typically used in lab studies. Limited attention or reaction capacity (distraction, high cognitive load, certain disabilities) means more time passes between a stimulus starting and the user actively withdrawing or intervening than in an attentive, cooperative threshold-study participant. None of these factors are visible to the device itself, so a single uniform ceiling is only as safe as the weakest case it was built to cover.

Studying it

The evidence base here comes more from clinical and occupational-safety literature than from routine human-factors usability testing: case reports of burns in diabetic neuropathy patients, pressure-injury risk research in populations with reduced mobility or sensation, and clinical data on skin tolerance in children and older adults. These populations are systematically absent from typical usability recruitment samples, which means the "user acceptance" a routine test produces is inherently biased toward underestimating risk.

Where it stops holding

"The most sensitive plausible user" is a judgment bounded by the product's actual, reasonably foreseeable user base — it does not mean every product must be engineered for the most extreme case imaginable across all of humanity. A medical device explicitly intended for diabetic patients must design around neuropathy-related insensitivity as a near-certainty, while a niche product with a narrow, self-selected enthusiast user base has a correspondingly narrower worst case to cover. The boundary here is the product's actual, foreseeable user population — not the handful of people who happened to walk into the usability lab, since that sample is itself pre-filtered toward health and cannot stand in for the former.

Applying it

• Before setting any intensity, temperature, or duration ceiling, list the perceptually vulnerable groups plausibly present in the target user base (older adults, people with neuropathy such as diabetics, children, anyone with reduced touch/pain sensation from surgery or medication), and set the ceiling at the end of published safety data that covers those groups — not at the median of healthy young-adult lab data.
• Do not rely on "the user says it hurts, so stop" as the sole safety mechanism — feedback from perceptually impaired users may be delayed or absent. Build hard engineering ceilings, independent of active user report, set separately across intensity, temperature, and duration, and enforce them in software.
• How to check: at the design-review stage, check the current ceilings against clinical or occupational-safety data for the most sensitive known subgroup in the target population — not just feedback from healthy participants in internal usability tests — as a dedicated "worst-case user" review, confirming the ceiling still holds a margin even under that assumption.

Related

• Same group: A4.07.1 Force feedback and thermal feedback both have a safety ceiling · A4.07.2 Sustained moderate stimulation can also cause harm
• Nearby: A10.06 Error-proofing design · A3.17 Types and effects of hearing loss
• Search terms: vulnerable population · sensory neuropathy · worst-case user · inclusive safety margin