health
two weeks in the freezer
Mailed frozen meals are a lower-cost way to feed homebound people with dementia, but the person must store, pick and reheat two weeks of them; in a US pilot trial, daily delivered meals did not significantly lower nursing-home placement compared with frozen meals, and interviews with nine recipients and seven caregivers found problems storing the boxes and knowing when they would arrive
Source: Thomas et al., “Home-Delivered Meals and Nursing Home Placement Among People…” Open ↗source: primaryLast updated: 2026-11-17
Problem statement
In the United States, home-delivered meal programmes such as Meals on Wheels feed hundreds of thousands of homebound older people, and the trial authors cite evidence that 40% of recipients have a dementia diagnosis. There are two main ways to deliver. A driver can bring a ready-to-eat meal several times a week, chat, and make an informal wellness check. Or the programme can mail a box of frozen meals every two weeks, which the authors call "a lower-cost alternative" with the same nutrition. But frozen meals, they write, "require additional steps for food storage and preparation that require cognitive abilities (e.g., prospective memory, cognitive sequencing) that are often impacted among" people living with dementia. In a pilot trial that enrolled 243 homebound people with self- or proxy-reported dementia in Texas and Florida in 2021, the daily-delivery group was less often admitted to a nursing home within six months, but the difference was not statistically significant, and the pilot was not powered to find it. In interviews in the same trial, a caregiver described having no room in the freezer for a delivery, and a recipient said the arrival time was always a surprise. The authors also list reading instructions and operating a microwave as challenges "of particular relevance for those with cognitive impairments". The frozen-meal recipients who were quoted said they managed the reheating. But the authors saw signs that mainly higher-functioning people chose to be interviewed; caregivers often preferred to be interviewed in place of recipients with more impairment, so those recipients were not heard from directly. The unsolved problem is how to make a two-week supply of meals usable by a person whose memory is failing and who often lives alone, without paying for a daily visit.
Why this matters
The trial authors write that about 5,000 programmes served more than 880,000 older adults in 2019, and that over half of recipients said the meals provided more than half of their food for the day. In the trial, 49.0% of participants lived alone. A 2026 scoping review of cooking and mealtime support in early dementia reports that about 60% of people living with dementia at home cannot complete parts of mealtime routines, such as food shopping or preparing meals, on their own. Health care organisations are increasingly providing meals too, the trial authors note. On the authors' reasoning, the cheaper model asks more of the people least able to manage it. The trial's measures were nursing-home placement and whether people were still receiving meals, not whether the meals were eaten.
What’s been tried and why it hasn’t worked
(1) A head-to-head pilot trial. Of 325 eligible people on waiting lists at three Meals on Wheels programmes who were randomised, 243 enrolled (128 daily delivery, 115 mailed frozen). After six months, 25 (10.1%) had entered a nursing home. The adjusted log hazard ratio for daily versus frozen meals was −0.67 (95% CI −1.52 to 0.19), favouring daily delivery but not significant. The difference between the arms in the share admitted was 6.0 percentage points (95% CI −14.0 to 1.5). The authors call the trial a feasibility success and say a full trial is warranted. The trial was run in 2021–22 under COVID-19 rules that weakened the daily model: one programme had drivers hang the bag on the door and step away, and another cut daily deliveries to twice a week, so some "daily" meals arrived chilled and may have needed reheating. More people in the frozen arm lived alone, and the authors note "the challenges that participants expressed for transporting, storing, and reheating frozen meals that were likely exacerbated if living alone." Only 68.8% (daily) and 62.6% (frozen) were still receiving meals at six months, against more than 80% in a national survey of programme participants. (2) What recipients said. In telephone interviews with nine people living with dementia and seven caregivers, people were satisfied with both kinds of meal: they saved time and money and reduced caregiver stress. The authors' summary names these challenges for frozen meals: reading instructions and operating a microwave, "confusion about the timing of deliveries," and storing large quantities. One caregiver said two boxes, 20 meals, came every other week, which "meant I had a storage problem". The programme "finally got that straightened out", but she still keeps meals in a small freezer box with refreezable packs because "they just don't fit." A recipient said the FedEx arrival time "was always such a surprise" and that knowing ahead "would be more helpful". On reheating, the quoted frozen-meal recipients coped: one follows the directions and adds a minute, another uses the oven. The authors saw evidence that higher-functioning people were the ones choosing to take part: caregivers often answered the phone and preferred to be interviewed in place of the recipient, citing the recipient's impairment. Even so, the nine recipients' cognitive screening scores ranged from 10 to 32, "indicating a broad range of function", and four were likely cognitively impaired. (3) Delivery fidelity was high, with weak points in the daily arm. A separate fidelity analysis found all three sites in the top quartile of adherence to the delivery protocol (16, 14 and 14 of 18 points), with high adherence in both arms, so the frozen meals were delivered as planned. The low scores were in the daily arm: all three sites scored low on keeping the face-to-face delivery model because of COVID-19 precautions, and two scored low on deliveries per week; the authors say this may have weakened the daily arm's effect. What the person did with the meals after the door is not measured. (4) Appliances that read the package. Barcode "scan-to-cook" microwaves exist: one maker's support page says a phone app can scan the UPC code on "any one of 3,000 frozen foods" and send the directions to the phone, which can then start the microwave. They need a smartphone, an app and a networked microwave, and whether a programme's meals are in the database is not known. No source read tested such a device with people with dementia. (5) A packaging redesign for home-delivered meals. The Live Well Collaborative worked with the Council on Aging of Southwestern Ohio to develop "new packaging for home delivered meals in a comprehensive system". It ran 29 interviews, rode along on a delivery, and had older adults test the redesigned packaging. Feedback covered "legibility, visual contrast, instructional design, and manual manipulation of the package." The project page does not mention frozen meals, memory loss or dementia, gives no year, and reports no measured results. (6) A large trial of the same comparison, not aimed at dementia. The same Brown University team is running Deliver-EE (NCT05357261), which randomised 2,300 older adults on waiting lists at fourteen Meals on Wheels programmes to daily delivery five days a week or "a shipment of 10 frozen meals, every two weeks". Its primary outcome is days spent in hospitals and nursing homes over six months; the registry lists primary completion as August 2027 (estimated), and the trial's page says enrolment closed in August 2025. It enrols people aged 66 and over and excludes anyone unable to take part in a phone survey, for example through "lack of comprehension of study purpose". The registry entry does not mention dementia, and the frozen arm is the standard box. A 2025 interview study from inside the trial (54 participants) reports that meals gave "much-needed nutritional stability" and relieved stress, but it does not report findings on storage, freezer space or delivery timing. (7) A programme that asked about freezers. An Indiana frozen-meal programme, in a 2018-grantee report published by the federal Administration for Community Living, asked clients, in an assessment used at the start and end of the programme, whether they had a standard-size refrigerator, with answers including "Dorm size refrigerator", "No freezer" and "Use cooler". Its satisfaction survey asked "Did your deliveries come when expected?" and "Did you find yourself eating all of the meals you receive?". These are self-reports from older adults at risk of food insecurity, not people with dementia. (8) A meal "pill box". In 2024 UK news coverage described a large "'pill box' style meal organiser" made by a meal-delivery company, which the coverage says was "invented to highlight the issue" of elderly relatives forgetting to eat. The coverage read reports no testing and does not say it was built for a frozen supply.
What would unlock progress
A frozen-meal "system" designed around the person who has to use it: a box size that fits a small freezer, a clear sign of which meal is today's, one-step reheating that does not depend on reading a label, and a way for a caregiver or the programme to know that meals are being eaten rather than piling up. Some of the pieces exist in other fields (pill organisers by day, picture-based instructions, scan-to-cook appliances). The team's larger trial, Deliver-EE, compares daily delivery with the standard frozen box in older adults in general and excludes people who cannot manage a phone survey, so it will not show whether a redesigned box helps people with dementia. A follow-on trial in people with dementia would be the place to test such a change, alongside the comparison of daily and frozen delivery.
How we’d know it’s being solved
(a) In a properly approved study, the share of mailed meals eaten within their intended day or week by recipients with memory loss, against the current box. (b) Task success and errors when older adults (and later, under ethics review, people with mild dementia) pick and prepare "today's" meal from a two-week supply, with and without the redesign. (c) Freezer fit: the share of recipients who can store a full delivery without extra equipment. (d) In the follow-on trial, whether a redesigned frozen model narrows the gap in nursing-home placement and in six-month retention (62.6% frozen vs 68.8% daily in the pilot). Beware the proxy: meals delivered are not meals eaten.
Where to start
Facility-free door: measure the storage problem the caregiver described. Buy several frozen ready meals sold in shops, measure their trays and outer boxes, and work out the space a two-week delivery takes (in the trial, one caregiver received two boxes, 20 meals in all, every other week). Compare it with the freezer compartments of small fridges, measured by hand or taken from makers' listed sizes. Then draft a short arrival notice that says which day the next box will come, since a recipient said the arrival "was always such a surprise", and test both for clarity with classmates and family members. Object door (reachable users): design a two-week meal box that fits a small freezer compartment, and a freezer organiser that shows "today's" meal at a glance. Test it with older adults at a senior centre or a community kitchen, using empty boxes, dummy trays and a cardboard mock-up of a small freezer, so that no food is served and nothing is heated. Start from the Live Well packaging work and add what it did not address: a multi-week frozen supply, freezer space and knowing which meal is today's. Do not redo its label and handling work. A day-by-day meal "pill box" has been made to highlight the issue of missed meals, with no testing reported, so the organiser's job is to make that idea work inside a small freezer and be tested. Unlike scan-to-cook microwaves, the box needs no phone, app or new appliance. Partner door (design the test, don't run it): with a home-delivered meal programme, design a protocol for testing the box with people with mild dementia who live alone, including how to learn whether meals are eaten without burdening them. Start from the Indiana programme's assessment question on refrigerator size and freezer and its self-report questions on delivery timing and meals eaten, and work out what a person with memory loss could not reliably report. Access line: testing with people with dementia needs ethics review and the programme's partnership; the deliverable is the protocol. Skills: packaging and product design, information design, human factors, gerontology.
Genes
Related problems
Stuck the same way, in other fields.
field: digital & spaceeurope's satellites run on parts others can embargoWell over half the electronic components in European spacecraft come from outside Europe, where export rules like ITAR can put the best technology out of reachbuilt without the people who'd live with itsource: primaryfield: society & learningexperiments without facesDigital learning platforms can now run cheap randomized experiments on hundreds of thousands of students — but cannot answer "for whom does it work?" Because they may not hold student demographic databuilt without the people who'd live with itsource: secondaryfield: food & the living worldfresh at dawn, half price by duskIn small-scale fisheries the fish is rarely thrown away — it is sold for half price as it warms, and no one can objectively grade the quality being lostbuilt without the people who'd live with itsource: primaryfield: water & energydilute, dispersed, and too potent to ignoreHundreds of thousands of engines, flares and coal mine shafts emit dilute methane, where conventional solutions often fall shortbuilt without the people who'd live with itsource: primarySources
- Thomas KS, Bunker J, Gadbois E, Hilgeman M, McCreedy E, Mills W, et al. "Home-Delivered Meals and Nursing Home Placement Among People With Self-Reported Dementia: A Pilot Pragmatic Clinical Trial." JAMA Network Open 6(12):e2347195, 2023, doi:10.1001/jamanetworkopen.2023.47195, accessed 2026-11-17. Link ↗
- Gadbois EA, Bunker JN, Hilgeman M, Shield R, McAuliff KE, Mills W, Thomas K. "Feasibility of conducting qualitative research with persons living with dementia and their caregivers during a home-delivered meals pilot trial." Pilot and Feasibility Studies 9(1):65, 2023, doi:10.1186/s40814-023-01302-5, accessed 2026-11-17. Link ↗
- Bunker JN, Hilgeman MM, McCreedy E, Gadbois E, Thomas KS. "Evaluating the Implementation Fidelity of a Pilot Pragmatic Randomized Clinical Trial Comparing Daily-Delivered Meals to Mailed Frozen Meals." Journal of Applied Gerontology 43(11):1605–1610, 2024, doi:10.1177/07334648241248269 (full text read through library access; record), accessed 2026-11-17. Link ↗
- Bannon SM, Elizalde Carrillo R, McCage S, Shahab MH, Allison GO, Aminpour A, Levinson C, Presciutti AM, Fishel A. "New horizons in cooking and mealtime interventions for mild cognitive impairment and early dementia: a scoping review." The Journal of Nutrition, Health & Aging 30(12):101009, 2026, doi:10.1016/j.jnha.2026.101009, accessed 2026-11-17. Link ↗
- GE Appliances, "Smart Microwave - Scan-To-Cook Feature" (support page), accessed 2026-11-17. Link ↗
- Live Well Collaborative, "COA Meals & More" (project page), accessed 2026-11-17. Link ↗
- ClinicalTrials.gov, "Deliver-EE: Evaluating Effects of Meal Delivery" (NCT05357261, Brown University), accessed 2026-11-17. Link ↗
- Balkan E, Gadbois EA, Tucher EL, Bernard KP, Thomas KS. "Home-Delivered Nutrition Services for Older Adults Under the Older Americans Act." JAMA Network Open 8(9):e2534747, 2025, doi:10.1001/jamanetworkopen.2025.34747, accessed 2026-11-17. Link ↗
- Administration for Community Living, "Modernizing Home Delivered Meals Through Frozen Meals and a Web-Based App – Indiana" (2018 grantee capstone), accessed 2026-11-17. Link ↗
- Grimsby Live, "New 'pill box' style meal organiser created to help elderly relatives eat" (2024-07-22), accessed 2026-11-17. Link ↗
Revision notes
This brief gets checked against its sources and contains the following revisions. Future revisions get added here.
Worth knowing
- the main evidence is one pilot trial, which was not large enough to show a difference in nursing-home placement
- the storage problem rests on one caregiver's account
Full check
Sources read
- Thomas et al., home-delivered meals pilot trial in self-reported dementia (JAMA Network Open, 2023): full text
- Gadbois et al., interviews with recipients and caregivers in the trial (Pilot and Feasibility Studies, 2023): full text
- Bunker et al., delivery fidelity in the trial (Journal of Applied Gerontology, 2024): full text (read through library access)
- Bannon et al., scoping review of cooking and mealtime support in early dementia (2026): full text
- Balkan et al., interview study from the larger Deliver-EE trial (JAMA Network Open, 2025): full text
- ClinicalTrials.gov, Deliver-EE trial record: full text
- Brown University, Deliver-EE trial page: full text
- GE Appliances, scan-to-cook support page: full text
- Live Well Collaborative, COA Meals & More packaging project: full text
- Administration for Community Living, Indiana frozen-meal programme report: full text
- Grimsby Live, meal 'pill box' organiser (2024): full text
What changed
- Not found in the sources, replaced with what they say: the Indiana programme was said to screen clients' freezers when they joined; its report shows an assessment, used at the start and end, that asks whether the client has a standard-size refrigerator, and the page and the study-design entry point now say that; the meal 'pill box' organiser was called a publicity piece; the article says it was made to highlight the issue of missed meals, and the entry point now says so.
- Stated more strongly than the source: the page said the interviews reached only higher-functioning recipients; the authors saw signs that higher-functioning people chose to take part, but the recipients' screening scores ranged widely and four were likely impaired; the page now gives both, and says caregivers often preferred to be interviewed in place of recipients; chilled 'daily' meals were said to have needed reheating; the trial says they may have needed it.
- Numbers: the 6.0-point difference in nursing-home admission was called unadjusted, which the trial does not say; it is now the difference between the two groups in the share admitted.
- Title: the long title now gives the number interviewed, nine recipients and seven caregivers, instead of calling the recipients higher-functioning (display title unchanged).
Left open
None noted in this check.
Separate check of specific claims: A later check changed no fact on the page.
Earlier full check: Seven details were corrected: the trial's dates now run into the following year, the delivery-fidelity finding now notes the weak points in the daily-delivery group, the scan-to-cook app sends directions to the phone, not the microwave, the Indiana report is described as a federal grantee report with a satisfaction survey, and the fidelity paper is now marked as read in full.
Separate check of specific claims: A larger trial of the same comparison in older adults in general, an interview study from inside it, an Indiana programme that asked clients about their freezers and a meal 'pill box' organiser were added.
Separate check of specific claims: The entry points were reworked: the no-facility one now measures the storage problem and drafts an arrival notice, the object one is tested with older adults using empty boxes and dummy trays with no food served and nothing heated, and people with dementia appear only in an entry point where students design the study, with an ethics-review line.
Earlier full check: The long title no longer says recipients described storing and reheating as hard (the quoted recipients managed reheating), the storage problem is credited to one caregiver, the delivery size was corrected to two boxes, 20 meals in all, every other week, and a packaging redesign project and the trial's difference between groups were added.