Tactile alarm research: what actually wakes deaf and hard-of-hearing sleepers
Short answer: the strongest published evidence on tactile wake-up signals for deaf and hard-of-hearing sleepers comes from one Australian research program using polysomnography-monitored sleep labs, and it found bed and pillow shakers woke most — not all — of the adults tested, with real variation by age, alcohol, and hearing-loss severity; that ceiling applies to every vibration device built on the same physical principle, ours included, not just the specific units the researchers used.
Full disclosure: we make a wrist vibration alarm and a bed shaker. Neither was tested in the research below — no consumer brand's product was — and this page says so plainly rather than borrowing the study's authority for a device it never touched.
The study behind almost every "vibrating alarms wake deaf sleepers" claim
Two closely related publications from Dorothy Bruck and Ian Thomas at Victoria University form the core of the evidence base this whole category leans on. The conference paper, Thomas, I., & Bruck, D., "Strobe Lights, Pillow Shakers and Bed Shakers as Smoke Alarm Signals," Fire Safety Science — Proceedings of the Ninth International Symposium (2008), pp. 415–424, and the journal paper, Bruck, D., & Thomas, I., "Smoke alarms for sleeping adults who are hard-of-hearing: comparison of auditory, visual, and tactile signals," Ear & Hearing, 30(1) (2009), pp. 73–80 (DOI 10.1097/AUD.0b013e3181906f89), report overlapping data from one research program: adults monitored with sleep-lab polysomnography, exposed to escalating tactile, auditory, and visual smoke-alarm signals during slow-wave (deep) sleep, tested in a university sleep lab or their own bedrooms.
Who was actually tested — and who wasn't
The hard-of-hearing group was 38 adults — 16 men, 22 women, mean age 54.4, ranging from 18 to 77 — with bilateral sensorineural hearing loss measured at 25 to 70 decibels. That's a defined range of hearing loss, not profound or total deafness, and it matters for how far these findings travel: someone with a milder measured loss in that band is a different case than someone profoundly or prelingually Deaf with effectively no residual hearing, and the paper doesn't claim to speak for that latter group specifically. A separate alcohol-impaired group — 32 young adults, normal hearing, mean age 21.2, tested at a 0.05 blood alcohol concentration — checked whether intoxication changed arousal thresholds in people who could otherwise hear fine.
What woke people, and what didn't
A 520 Hz square-wave audible tone was the most effective signal overall, waking 92 to 100 percent of participants in each group at or below a 75-decibel reference level. Tactile bed and pillow shakers landed in the middle for the hard-of-hearing group specifically: they failed to wake 17 to 20 percent of those participants at the tested intensity levels — roughly four in five did wake, but a meaningful minority did not, even in a monitored lab where people knew a test signal was coming. Strobe lights performed worst: roughly 75 percent of both groups failed to wake at the lowest intensity tested, a level that itself exceeded the relevant NFPA visual-alarm standard. None of these numbers describe a device that reliably wakes everyone — they describe a hierarchy of relative effectiveness, with a real failure rate attached to every signal type, tactile included.
Alcohol changes the picture more than most marketing admits
The alcohol-impaired group's numbers make the same shakers look considerably less reliable: 36 to 42 percent failed to wake to the tactile signal at the same benchmark intensities, roughly double the hard-of-hearing group's failure rate. That's worth sitting with — it means the limiting factor for a tactile alarm isn't only hearing status. A sedated, exhausted, or intoxicated sleeper with completely normal hearing can be harder to wake with vibration than a sober adult with moderate hearing loss. Assuming a bed shaker is a guaranteed wake-up regardless of the sleeper's state extrapolates well past what this data shows.
What the study's own authors flagged as limitations
Both papers are candid about where their numbers might not hold up outside the lab. The researchers note that their participants were primed — they knew they'd volunteered for a sleep study and consented to being woken by test signals — and describe their sample as "highly screened," both factors that plausibly make lab arousal thresholds lower than what an unwarned person would show on an ordinary night at home. They also flag that it's genuinely difficult to isolate why arousal rates differ across different bed-shaker studies in the wider literature — device design, placement on the mattress, natural night-to-night variation in sleep depth, individual differences between people, and intoxication status are all tangled together, and this study doesn't claim to have cleanly separated them.
What a 2022 review added — and what it says is still missing
A more recent scoping review, Smedberg, E., Ronchi, E., & Hutchison, V., "Alarm Technologies to Wake Sleeping People Who Are Deaf or Hard of Hearing," Fire Technology, 58(4) (2022), pp. 2485–2507 (DOI 10.1007/s10694-022-01265-8), surveyed 13 alerting technologies across audible, olfactory, tactile, and visual categories, and concluded that low-frequency audible alarms and tactile devices remain the "most reliable existing technologies" for this population. It's just as direct about what's unresolved: it calls for more research on older adults and bed/pillow shakers specifically, since hearing loss becomes more common with age while the tactile-arousal research base skews toward the Bruck and Thomas dataset above; it notes there's no standardized way to measure vibrational-device effectiveness across studies at all; and it flags an adoption gap — people who are deaf or hard of hearing don't necessarily use the technologies the literature calls effective — based on a survey of only 36 North American respondents, a small sample the review itself says limits that conclusion's reach.