Skip to content
problem genome

One of 592 unsolved problems in the Problem Genome, a collection written up for student design and engineering teams.

health · digital

the button left unpressed

Older people who fell alone and could not get up mostly did not use the alarm they had (78% of such falls at home in a UK over-90s cohort; 11 of 13 in a small German sample), and in one small US home pilot a commercial wearable fall detector caught 1 of the 4 falls that happened while it was worn, and 83 of its 84 alarms were false

Source: Fleming et al., “Inability to get up after falling, subsequent time on floor…” Open ↗source: primaryLast updated: 2026-11-17

Problem statement

A personal alarm (a pendant or wrist button linked to a call centre, called a social alarm, Hausnotruf or trygghetsalarm in different countries) is widely issued to older people, many of whom live alone. One of its main jobs is the moment after a fall when the person cannot get up. Studies in the UK and Germany found that at that moment the button is often not pressed. In a cohort of 110 people over 90 in and around Cambridge, UK, followed for a year, the person had some form of call alarm in 99% of the falls where they were alone and could not get up, yet in 80% of those falls they did not use it. That figure includes care-home call bells (94% not used); in falls at home in the community it was 78% (28 of 36), and in sheltered housing 59% (23 of 39). In 37 of the 38 "long lies" of an hour or more despite an installed alarm, it was not used. In Germany, in an interview study of 52 alarm subscribers, 11 of the 13 who had fallen alone at home in the previous three months and lay on the floor for more than five minutes did not use the alarm (the paper gives this as 83% in its abstract and 84% in its discussion). In Norway, in 19 interviews with people with dementia in a trial, 54% did not know they had an alarm. Automatic fall detection was meant to remove the need to press. In a four-month home pilot in the US, one commercial wearable detector found 1 of the 4 falls that happened while it was worn among those who completed the study, raised 84 alarms of which 83 were false, and missed falls that happened while it sat on its charger at night. The unsolved problem is an alarm that is actually on the person, and actually raised, at the moment they are on the floor.

Why this matters

Lying on the floor for a long time after a fall is common in the very old and dangerous. In the Cambridge cohort, 82% of falls happened when the person was alone, 80% of those who fell could not get up after at least one fall, 30% had lain on the floor for an hour or more, and a long lie was strongly associated with serious injury, hospital admission and moving into long-term care. Almost no one who did press their alarm lay for more than an hour, which suggests the alarm helps when it is used (our reading). Alarms are paid for by municipalities and subscribers on the assumption that they are used. The Norwegian authors conclude that installing one "does not automatically mean that it is in use" and that people with dementia often feel a false sense of security from it.

What’s been tried and why it hasn’t worked

(1) The pendant alarm. The Cambridge study gathered the reasons from participants and carers. Barriers appeared at three stages: not seeing the point of having one, not getting into the habit of wearing the pendant, and, after a fall, not raising the alarm, either as a decision or as a failed attempt. Decisions included "I didn't want to use the call alarm, although I was wearing it, for fear of being taken into hospital" and "Didn't need to," from a man who fell twice while trying to get up. Failed attempts included a person who fell in hospital and could not reach the alarm, a man whose pendant did not work after a power cut, a woman who did not think her pendant would work in the street outside her flat, and a woman found confused on the floor who thought she had pressed it but had not. The authors call for work "on access and activation issues for design of call alarms". In the German sample of 52 subscribers, satisfaction was very high, yet 14 of the 52 (27%; the abstract says 24%) answered that they "never wear the alarm button" when alone at home, and only 7 of the 52 wore it 24 hours a day (the abstract says 14%). (The German abstract says "den Notrufsender nie zu nutzen", in our translation "never to use the alarm transmitter".) The most common reason given for not using the alarm after a fall was the wish to be independent and manage it themselves (8 people). None of five satisfaction measures predicted use; only how relevant the person thought the alarm was did. In Norway, of 82 people with dementia reaching the trial's two-year assessment, 62.2% had a social alarm; in 19 interviews at 18 months, 54% did not know they had one and 46% were not wearing it; yet at 24 months 60.8% of those answering said the alarm contributed to their safety; in one home the wife wore her husband's alarm in case he fell. (2) Automatic fall detection. In the US test, 18 older adults were given a commercial wearable fall detector; 8 completed four months. Sensitivity among completers (7 analysed; 3 participants were excluded for carrying the device off the body) was 0.25 against the manufacturer's laboratory figure of about 94%, and positive predictive value was 0.02. Specificity (0.92) matched the laboratory figure closely, and false alarms did not measurably change adherence in the five days after them. Of 84 alarms, 83 were false, most of them during normal use or when the device was dropped or put down (about a fifth had no known cause). Eight falls happened while the device was not being worn, half of them with it on its charger, in the early morning or at night. Adherence fell by about the halfway point. The authors conclude that "the device is inaccurate in real world settings given the very low sensitivity observed"; it was one device and a small sample. Real-world detection is not hopeless: a prototype accelerometer sensor from a group at the University of Oulu, Finland, worn by 16 older people for 15,500 hours caught 12 of 15 real falls (sensitivity 80%), with about one false alarm per 20 to 40 hours of use (Kangas et al. 2015, abstract read). A 2026 scoping review of 243 fall-detection studies found more than half relied on simulated laboratory falls and only 21 had real-world or mixed validation with older adults; it names false alarms, battery management, comfort and long-term wear as the barriers. (3) Sensors that need nothing worn. Passive room sensors that raise an alert when someone is lying on the floor are sold (the 2023 Norwegian authors name one), but those authors note there is little research to confirm their appropriate use in home-living people with dementia; and, by their nature, they cover only the rooms where they are fitted (our note). (4) Pendants that show the alarm has gone. At least one maker's pendant already confirms a press with a light: on the Careium Enzo, worn at the wrist or neck, "The LED light changes colour to reassure you that help is on the way" (maker page, linked under Sources). The same page says the pendant is waterproof and safe to wear in the shower or bath, and offers an optional Easy Press Adapter for users with arthritis or weak grip. It describes no signal for an alarm that was not sent or a pendant out of range of the base unit, and nothing about pressing it from the floor.

What would unlock progress

The failures sit at different links of one chain: owning it, wearing it (including at night, when it is taken off for bed or left to charge), being able to press it from the floor, knowing it has worked, and being willing to call. A design that treats the whole chain as the product, rather than the button or the detection algorithm alone, is the opening. The users' own reasons point to concrete targets: a form people keep on through the night without charging gaps (and, our addition, in the shower); activation that works when the button is hard to reach (and, our addition, with weak or arthritic hands and lying in any position); clear feedback that the call has gone (one maker's pendant already shows this with a light; see What's Been Tried), also when the pendant is out of range of the base unit or the base has lost power (two makers' own pages describe speech pendants with a built-in microphone and speaker, through which the wearer talks to the call centre, one giving a range of 50 m indoors from the base unit; neither says what the wearer is told when the call does not go through); and a way to ask for "help to get up" that does not feel like summoning an ambulance, since fear of hospital stopped one woman pressing. The Cambridge authors also note that people need training in ways to get up from the floor.

How we’d know it’s being solved

(a) The share of falls alone, with inability to get up, in which help is summoned: here about 22% at home in the community (UK) and 2 of 13 (Germany); the target is most. (b) Wear time: share of users wearing the alarm 24 hours a day (14% in the German sample), measured by the device, not self-report. (c) For automatic detection, sensitivity and false alarms per user-day measured in homes over months, not in simulated falls (0.25 and 83 of 84 false in the US test). (d) No falls occurring while the device is off the body or charging. (e) Fewer long lies of an hour or more.

Where to start

Facility-free door (make a thing): automatic detectors are already sold (the US pilot tested a commercial one), and at least one maker's waterproof pendant, worn at the wrist or neck, has a light that changes colour when help is on the way and an optional Easy Press Adapter that "reduces the force needed to trigger an alarm which is ideal for users with arthritis or weak grip strength" (Careium Enzo maker page, linked under Sources); two makers also sell speech pendants with their own microphone and speaker, so within range of the base unit the wearer hears the call centre through the pendant (maker pages linked under Sources). Check what is sold first and do not remake them. The team's object starts from the "sent" light that is already sold and addresses what the page shows is still open: telling the user when the alarm was NOT sent, because the pendant is out of range of the base unit or the base has lost power, and pressing from the floor, in whatever position the person lies and wherever the pendant has swung. Design and prototype an alarm button housing (no radio needed: a switch, a buzzer, a vibration motor and lights) that can be pressed with a palm, a forearm or the chin as well as a weak grip, and that gives clearly different felt and heard signals for "pressed and sent" and "pressed but not sent" (the out-of-range and power-cut cases, simulated with a switch). Bench-test it seated or at a bench: a force gauge for the press force, and a jointed mannequin or padded rig that mimics a wearer lying on their side, back or front for reach; record the force and reach each press needs. No partner, facility or data is needed beyond a makerspace. No team member gets down on the floor to act out a fall unless they choose to and can get up unaided; the rig does that job. User door: older adults living alone can be reached through a senior centre or a sheltered-housing common room; some will already have an alarm and can say when they take it off and why. In supervised sessions seated at a table (no one is asked to lie on the floor or simulate a fall), they press the team's housings and a standard pendant on a force gauge, with gloves that mimic reduced grip if wanted, tell the "sent" and "not sent" signals apart, and rate whether they would keep it on overnight and in the shower; collect their reasons in their words, as the Cambridge study did. A home study of wear time and real-fall use is designed, not run: write the protocol, with device-logged wear time and summoned-help rates as outcomes. Sessions with older adults need the campus's ethics review before they start; any work with people with dementia is design-the-test only. Relevant skills: product and industrial design, human factors, simple electronics.

Genes

Related problems

Stuck the same way, in other fields.

field: food & the living worldninety-eight percent of u.s. chemicals, never assessedOnly 1.8% of U.S. commercial chemicals have a published human health assessment from EPA or ATSDRtested in the wrong worldsource: primaryfield: food & the living worldthe same mistake, the same bananaCommercial banana monoculture faces existential threat from repeat of previous variety's destructiontested in the wrong worldsource: primaryfield: food & the living worldfew days from first sign to empty pondWhite spot syndrome virus cut Brazil's shrimp production by a fifth in a single year — and fast, sensitive tests to prevent the next outbreak are still held back by costtested in the wrong worldsource: primaryfield: food & the living worldthe eye that doesn't exist yetMapping Earth's ecosystems and minerals from space requires an imaging spectrometer nobody has builttested in the wrong worldsource: primary

Sources

  • Fleming J, Brayne C, Cambridge City over-75s Cohort (CC75C) study collaboration. "Inability to get up after falling, subsequent time on floor, and summoning help: prospective cohort study in people over 90." BMJ (Clinical research ed.) 2008;337:a2227, doi:10.1136/bmj.a2227, accessed 2026-11-17. Link ↗
  • Heinbüchner B, Hautzinger M, Becker C, Pfeiffer K. "Satisfaction and use of personal emergency response systems." Zeitschrift fur Gerontologie und Geriatrie 2010;43(4):219-223, doi:10.1007/s00391-010-0127-4 (article in English with abstracts in English and German; full text read through library access), accessed 2026-11-17. Link ↗
  • Puaschitz NGS, Jacobsen FF, Berge LI, Husebo BS. "Access to, use of, and experiences with social alarms in home-living people with dementia: results from the LIVE@Home.Path trial." Frontiers in aging neuroscience 2023;15:1167616, doi:10.3389/fnagi.2023.1167616, accessed 2026-11-17. Link ↗
  • Chaudhuri S, Oudejans D, Thompson HJ, Demiris G. "Real-World Accuracy and Use of a Wearable Fall Detection Device by Older Adults." Journal of the American Geriatrics Society 2015;63(11):2415-2416, doi:10.1111/jgs.13804, accessed 2026-11-17. Link ↗
  • Telemedcare Ltd, "Home Alarm and Speech Pendant!" (maker's product sheet, undated), accessed 2026-11-17. Link ↗
  • Climax Technology Co, Ltd, "WTRVS-8/15 Series" (maker's product page), accessed 2026-11-17. Link ↗
  • Kangas M, Korpelainen R, Vikman I, Nyberg L, Jämsä T. "Sensitivity and False Alarm Rate of a Fall Sensor in Long-Term Fall Detection in the Elderly." Gerontology 2015;61(1):61-68, doi:10.1159/000362720 (abstract read via Europe PMC), accessed 2026-11-17. Link ↗
  • Chen L, Yao W. "Fall detection and pre-impact prediction technologies in older adults: a scoping review of translational maturity and public health integration." Frontiers in public health 2026;14:1737644, doi:10.3389/fpubh.2026.1737644, accessed 2026-11-17. Link ↗
  • Careium, "Enzo Personal Alarm Trigger" (maker's product page), accessed 2026-11-17. Link ↗

Also cited in the brief

  • WHO, "Priority assistive products list" (2016) Link ↗
Revision notes

This brief gets checked against its sources and contains the following revisions. Future revisions get added here.

Worth knowing

  • the German paper gives its key figures in two forms (83% and 84%; 24% and 27%); the page prints both
  • the speech-pendant and confirm-light descriptions come from makers' own pages, not from tests

Full check

Sources read

  • Fleming & Brayne, falls in people over 90, Cambridge cohort (BMJ, 2008): full text
  • Heinbüchner et al., satisfaction and use of personal alarms (Zeitschrift für Gerontologie und Geriatrie, 2010): full text
  • Puaschitz et al., social alarms in people with dementia, LIVE@Home.Path trial (Frontiers in Aging Neuroscience, 2023): full text
  • Chaudhuri et al., real-world accuracy of a wearable fall detector (Journal of the American Geriatrics Society, 2015): full text
  • Chen & Yao, scoping review of fall detection (Frontiers in Public Health, 2026): full text
  • Kebede et al., rapid review of technology-enabled care (PLOS Digital Health, 2026): full text (searched in part)
  • Kangas et al., long-term test of a fall sensor (Gerontology, 2015): abstract only
  • Careium, Enzo personal alarm trigger (maker's page): full text

What changed

  • Not found in the sources, replaced with what they say: the Finnish sensor was placed in Finland; the record gives only the research group's university, and the page now says so; the bathroom and weak hands were listed among users' own reasons for not having the alarm on; the sources report taking it off for bed, leaving it to charge at night and not reaching it, and the page now says that, keeping the shower and weak hands as its own design targets.
  • Stated more strongly than the source: the alarm is now described as widely issued to older people, many of whom live alone, with the moment after a fall as one of its main jobs, as the sources put it; only the UK and German studies measured whether the alarm was used after a fall; the page no longer says three countries did; that the alarm works when used is now marked as the page's own reading of the Cambridge finding that almost no one who pressed lay on the floor for over an hour.
  • Numbers: the Norwegian 54% is now given as the paper gives it, from 19 interviews, without saying it is a share of 19 people; the causes of the 83 false alarms are now given in the paper's proportions: most during normal use or when the device was dropped or put down, about a fifth with no known cause.
  • Title: the long title now says the detector caught 1 of the 4 falls that happened while it was worn (display title unchanged).
  • Entry points: the make starting point now builds on a maker's pendant that already lights up when help is on the way and an easy-press adapter already sold; what it leaves open is telling the user when an alarm was not sent, and pressing from the floor; the user starting point says some older adults at a senior centre will already have an alarm, not many.
  • Other: a fourth earlier attempt was added: a maker's pendant that confirms a press with a light, which describes no signal for an alarm that was not sent.

Left open

  • the Finnish sensor's real-world results (its paper was read only as an abstract)

Separate check of specific claims: Two makers' own pages on speech pendants replaced two sentences saying this had not been checked, a product-list note now rests on the WHO list itself, and two citations were added to the source list.

Earlier full check: The German figures were restated from the paper's results (11 of 13 who fell alone; 14 of 52 who never wore the button), the long title now begins 'Older People', who pays for alarms was corrected, and two product statements were marked as not yet checked.

Separate check of specific claims: The make starting point was narrowed to pressing from the floor and a 'pressed but not sent' signal, with safety lines for bench tests and an ethics line for sessions with older adults.

Earlier full check: The titles were corrected (the 80% figure included care-home call bells; display title 'The Button Nobody Pressed' → 'The Button Left Unpressed'), figures were given with their counts, and a positive real-world detector result and the Norwegian safety finding were added.