Duke’s study of home-delivered meals after discharge found an association with fewer early readmissions among older adults at risk of food insecurity or malnutrition. The association weakened at 30 days. The comparison was observational, so it cannot establish that meal delivery caused the difference. The original study abstract
For care teams, the useful questions concern who actually received support, when it arrived and which follow-up period is being counted.
A referral did not always mean a delivery
The retrospective cohort included 380 eligible referrals involving 368 people aged 60 or older, from 621 unique referrals. These denominators describe different units: a person and a referral are not necessarily interchangeable.
The programme offered two weeks of meals after discharge. Researchers compared referred people who received meals with referred people who did not. Duke’s study account describes one meal daily and the programme’s setting at the health system.
That comparison is more specific than “meals versus usual care.” The groups were not assigned at random, and differences involved in receiving a delivery could also relate to subsequent readmission.
Follow-up changes the reported association
The abstract reports adjusted odds ratios for unplanned readmission, using a three-day landmark to exclude early events.
Adjusted readmission odds by follow-up period
| Period | Meal receipt versus no meal receipt: adjusted odds ratio | 95% confidence interval |
|---|---|---|
| 14 days | 0.31 | 0.14–0.69 |
| 21 days | 0.47 | 0.23–0.95 |
| 30 days | 0.82 | 0.44–1.53 |
The 30-day interval includes one, the value representing equal odds in this comparison. The table therefore does not support describing the same clear association at every time point. An odds ratio is also not automatically a percentage reduction in a person’s risk.
The authors’ timing analysis associated earlier delivery with fewer 14-day readmissions, with uncertainty at later periods. It supports further evaluation of timing, rather than a proven nine-day deadline for care.
The service pathway remains part of the research question
The distinction between being referred and receiving support is operationally important. Evaluating a programme requires both stages to be visible, along with the outcome window.
The original abstract and institutional account support this bounded report. The full journal paper was inaccessible; no full appraisal of adjustment or residual confounding is claimed. The result concerns a particular post-discharge programme, not an individualized diet recommendation or a guarantee that providing meals prevents readmission.