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

counting sips nobody feels

Thirst weakens with age, so care-home staff must track what residents drink — but in one UK study the few staff fluid charts returned agreed poorly with observed drinking, and in a Spanish robot-supported hydration trial the authors could not separate real drinking from differences in record-keeping

Source: Vallès-Carvajal et al., “Systematic Hydration Rounds Versus a Socially…” Open ↗source: primaryLast updated: 2026-11-12

Problem statement

Thirst is an interoceptive signal — the body telling the brain it needs water — and it weakens with age. A Spanish nursing-home study explains low drinking partly by "reduced thirst sensation and memory problems associated with age," which "can contribute to a refusal to drink, either because the need is not perceived, or because they do not remember that they have not been drinking." When residents cannot feel or report the need, staff have to offer drinks and track intake, and that is where the system breaks. In a study in UK residential homes, few staff-completed fluid charts were returned, and those correlated poorly with researcher observation (r = 0.122, p = 0.818; the charts averaged 702 ml/day below observation), while residents' own diaries were accurate but only for those "physically and cognitively able to complete it." In a Lleida (Spain) nursing home, the recording system showed only whether a resident had drunk, not how much, and staff called recording an excessive workload. When the same research group compared staff hydration rounds with rounds supported by a socially assistive robot (reminders, praise and a digital intake record), the robot floors recorded less intake (with no significant difference in intake trajectories over the month), but the authors conclude it is "not possible to determine" whether that reflects real drinking, the different documentation procedures, or both. Research prototypes of weighing cups and coasters exist, but a 2021 review found very few tested with older people and none widely available. The unsolved problem is measuring what residents with weakened thirst actually drink, under real care-home conditions and without adding staff work, so that hydration programmes can be run and judged at all.

Why this matters

The Lleida trial cites a meta-analysis estimate that 34% of nursing-home residents are dehydrated (by serum osmolality) and reports that fluid intakes below 1,500 ml/day are found in 24% to 85.5% of residents across studies, against ESPEN's recommended minimum of 1.6 l for women and 2.0 l for men over 65. A related paper lists the harms of dehydration as including falls, acute kidney injury, urinary tract infections, delirium, hospitalisation and death. Interventions to fix this — rounds, reminders, flavoured drinks, robots — are judged by the fluid intake they produce, and in nursing homes that intake is commonly recorded by care staff (both arms of the Lleida trial were scored on staff records), so if the records cannot be trusted, neither can the evaluations.

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

(1) Staff fluid charts. The 2015 UK FISE study (22 residents, Norfolk) found staff charts unreliable (few returned; r = 0.122, p = 0.818, so the comparison rests on a small number of charts), and the Lleida trial authors note that staff "tend to record beverages offered rather than those consumed." (2) Resident self-report. The Drinks Diary matched observation well (r = 0.93 in the paper's abstract, 0.851 in its results section; 19 of 22 correctly classified as drinking enough or not), but 12 of the 22 participants already scored below 27 on the MMSE and the tool is limited to residents able to fill it in — the residents whose thirst and memory are weakest are the ones it cannot serve. (3) Apps for staff. An English hydration app (Hydr8), co-developed with care homes, was valued (staff reported better understanding of hydration and better communication of intake data), but "technical glitches hindered the seamless embedding of Hydr8 into everyday practice" (abstract). (4) A robot. In the Lleida quasi-experiment (57 residents, four floors, one month), two floors received a robot that delivered reminders and positive reinforcement and prompted staff to complete a digital record after each round; the robot group was associated with 354 ml/day lower recorded intake after adjustment, but it also started with lower intake and more dependent residents, and its real-time digital record differed from the paper record kept at the end of each shift on the other floors. No significant group × time difference in intake trajectories was found, and daily checks confirmed all scheduled records were completed in both groups — so the problem was not missing records but records that cannot be compared. The authors say technological tools "require adequate integration into routine clinical practice." (5) Sensor cups, coasters and holders. A 2021 review of automatic fluid-intake monitoring for older people (Cohen, Fernie and Roshan Fekr) found that sensors in containers or surfaces were the only approaches that estimated volume accurately (most container techniques above 90% accuracy), but that "very few were tested on this age group," only four systems were tested in free-living conditions, the user "must drink from a specific container or in a specific location," weight-based containers must be set down and calibrated per container, and "there is no widely, commercially available product that aids monitoring and tracking senior liquid intake automatically." Prototypes it describes include a load-cell coaster (WaterCoaster, 20 users, data lost to Bluetooth drop-outs and coaster movement) and a load-cell cup holder tested over five months in a nursing home (Zimmermann et al. 2017, IEEE SAS; read only as summarised in the review). A Dutch validation and feasibility study of a "Smart Mug" in nursing-home residents (anticipated start 2016, planned n = 20, hypothesis of validity r ≥ 0.5) excludes residents with cognitive impairment; no results are posted. Commercial sensor cups exist, and one has been studied in use: a case study of the Auxivia smart drinking glass, which "can measure the amount of water drunk", observed three French nursing homes (9 hours at each site, 29 interviews) and describes how each home's work arrangements shaped its use; its abstract reports no accuracy check against observed drinking (abstract read). No independent accuracy data in care-home residents with cognitive impairment were found in this search.

What would unlock progress

A measurement that does not depend on the resident's thirst, memory or literacy, or on staff writing things down. Weighing cups and coasters already work in the lab; what they left unsolved is care-home use: residents who drink from whatever cup is handed to them, assisted drinking, attribution of each reading to the right person without typing, and validation against observation in residents with dementia (the Dutch Smart Mug study excludes them). Intake then has to be summarised for staff at a glance. It has to cope with thickened drinks (the Lleida trial found dysphagia independently associated with 231 ml/day lower intake, and textures were recorded as liquid, nectar or pudding), shared and refilled cups, spills, and drinks taken outside meal rounds. With a trustworthy intake signal, the existing strategies (rounds, reminders, social drinking, robots) could finally be compared fairly.

How we’d know it’s being solved

(a) Agreement with direct observation over a day (the FISE benchmark: the Drinks Diary reached r = 0.93 by the paper's abstract and 0.851 by its results section; staff charts r = 0.122), including for residents with dementia and for thickened fluids; (b) staff minutes per resident per day spent recording, compared with current practice; (c) share of residents correctly classified as drinking enough or not against the ESPEN thresholds. Beware the proxy: drinks offered or served are not drinks drunk.

Where to start

Facility-free door: start from the published designs in the Cohen et al. review (load-cell coaster, load-cell cup holder), rebuild one, and stress-test the failure modes those designs left open: refills, cup swaps between people, set-downs on other surfaces, spills, and thin versus thickened liquids, against a kitchen scale with volunteers. Design door: design the staff-facing side — what a carer on a busy shift needs to see (who has not drunk since breakfast) and how a reading gets attached to the right resident without typing — and test it as a paper or screen prototype with care workers. Partner door (design the trial, don't run it): with a care home, design a one-day validation protocol comparing the device, staff charts and direct observation, modelled on the FISE design, with a pre-registered analysis plan. Access note: testing with residents needs a care-home partner and ethics review; the facility-free door uses volunteers. Relevant skills: embedded sensing and load cells, product design, interaction design for care work, nursing or dietetics input.

Genes

Related problems

Stuck the same way, in other fields.

field: the built worldthe wrong ruler for concrete's lifetimeDurability rules judge low-clinker concrete by how fast it carbonates — but carbonation alone does not corrode rebar, 16 competing test standards cannot be converted to real-world rates, and a standard's conservative carbonation rates would write off reusable precast elementsmeasured the wrong thingsource: primaryfield: society & learningnobody measures the earRoughly 11 million U.S. workers are predicted to get less attenuation than they need from their hearing protection, and no standard exists for measuring the noise dose that actually reaches the earmeasured the wrong thingsource: primaryfield: society & learningnew curricula, teaching to the testCountries rewrite curricula around competencies, but where exams reward content, teaching follows the testmeasured the wrong thingsource: primaryfield: society & learningnobody counts the whole dayRail, maritime, air and passenger transit all have workers whose on-call time is not well tracked, and a complete picture of time at work is needed to determine impacts on sleep, fatigue and injurymeasured the wrong thingsource: primary

Sources

  • Vallès-Carvajal C, Selva-Pareja L, Masot O, Falcón-Villaitodo M, Camí C, Botigué T. "Systematic Hydration Rounds Versus a Socially Assistive Robot to Promote Fluid Intake in Institutionalized Older Adults: A Quasi-Experimental Study." Nutrients 2026, doi:10.3390/nu18162715, accessed 2026-10-22. Link ↗
  • Paraíso-Pueyo E, Vallès-Carvajal C, Camí C, Botigué T, Selva-Pareja L, Alzuria-Alós RM. "Hydration and Dehydration Prevention in Nursing Homes: Perspectives, Barriers, and Practices of Care Teams and Managers." Nutrients 2026, doi:10.3390/nu18040630, accessed 2026-10-22. Link ↗
  • Jimoh FO, Bunn D, Hooper L. "Assessment of a Self-Reported Drinks Diary for the Estimation of Drinks Intake by Care Home Residents: Fluid Intake Study in the Elderly (FISE)." The journal of nutrition, health & aging 2015, doi:10.1007/s12603-015-0458-3 (abstract, accessed 2026-10-22; full text, accessed 2026-11-12) Link ↗ Link 2 ↗
  • Steven A, Wilson G, Young-Murphy L. "The Implementation of an Innovative Hydration Monitoring App in Care Home Settings: A Qualitative Study." JMIR mHealth and uHealth 2019, doi:10.2196/mhealth.9892, accessed 2026-10-22. Link ↗
  • "Smart registration of fluid intake with the 'Smart Mug'; a validation and feasibility study," Netherlands trial register NL-OMON25903, accessed 2026-10-22. Link ↗
  • Cohen R, Fernie G, Roshan Fekr A. "Fluid Intake Monitoring Systems for the Elderly: A Review of the Literature." Nutrients 2021, doi:10.3390/nu13062092, accessed 2026-10-22. Link ↗
  • Gazza C, Marcilly R, Kovacs B, Schiro J, Pelayo S. "Integration of a new technology into a work system: a case study of a smart drinking glass in French nursing homes." Disability and Rehabilitation: Assistive Technology 19(4):1249-1261, 2024, doi:10.1080/17483107.2022.2162612 (abstract read via Europe PMC record, PMID 36622869), accessed 2026-11-12. Link ↗
Revision notes

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

Worth knowing

  • The drinks-diary study gives its main agreement figure two ways: a correlation of 0.93 in its summary and 0.851 in its results section. The page gives both.

Full check

Sources read

  • Vallès-Carvajal et al., hydration robot trial in a care home (Nutrients, 2026): full text
  • Paraíso-Pueyo et al., care staff focus groups on hydration (Nutrients, 2026): full text
  • Steven et al., Hydr8 hydration app study (JMIR mHealth and uHealth, 2019): full text
  • Cohen, Fernie and Roshan Fekr, review of sensor drinkware (Nutrients, 2021): full text
  • Werden Abrams et al., staff views on hydration care (International Journal of Older People Nursing, 2026): full text
  • Scoping-review protocol on care-home hydration (BMJ Open, 2026): full text
  • Jimoh, Bunn and Hooper, drinks diary study in care homes (2015): full text
  • Dutch trial registry entry for a smart mug study: full text
  • Gazza et al., smart drinking glass in French nursing homes (2024): abstract only

What changed

  • Numbers: the 702 ml/day gap between staff charts and direct observation: the charts were on average below what residents were seen to drink, and the page now says so in both places; the drinks diary's agreement with observation: the page now gives both figures the paper reports, 0.93 and 0.851, and says where each comes from.
  • Links: the drinks-diary study now has its full-text address beside its abstract address.

Left open

  • two prototype studies, a drink-weighing coaster and a nursing-home cup holder (known only through the review's account of them)
  • what the smart drinking glass case study found (abstract only)
  • the opening description of thirst as a body signal (a definition, not drawn from a source)

Separate check of specific claims: A smart drinking glass studied in French nursing homes was added as an earlier attempt, and the notes on which sources were read in full were brought up to date.

Earlier full check: A claim about how all hydration programmes are judged was brought in line with the trial, and an unsupported aside about thickener being sold over the counter was removed.

Separate check of specific claims: The long title was toned down, the robot trial's result and record-keeping were described more exactly, the hydration app's positive findings were added, and a review of sensor drinkware was added as an earlier attempt.