Delirium at the front door predicts a three-week admission

Design

Secondary analysis of a stepped-wedge, cluster randomised trial using logistic regression (NCT03300336)

Population

2,612 community-dwelling veterans with dementia hospitalised for at least two days at eight Veterans Health Administration hospitals, 2017-2019

Primary outcome

Patient-level factors associated with a prolonged hospitalisation of 21 days or more

Effect

Median stay 6 days, 5.3% stayed 21 days or more; social vulnerability OR 1.74 (1.16-2.61), Nosos burden 1.06 (1.01-1.10), delirium at admission 1.55 (1.07-2.26); greater frailty associated with lower odds

People with dementia stay in hospital longer, and longer stays cause the complications, deconditioning and deaths that make the next admission worse. This secondary analysis of a stepped-wedge cluster trial across eight Veterans Health Administration hospitals asked which factors present on admission identify who is going to be there in three weeks.

Among 2,612 community-dwelling people with dementia hospitalised for at least two days, the median stay was 6 days (IQR 4-9) and 138 (5.3%) stayed 21 days or more. Three admission-day factors predicted that: social vulnerability (odds ratio 1.74, 95% CI 1.16-2.61), chronic disease burden on the Nosos score (1.06 per unit, 1.01-1.10), and delirium present at admission (1.55, 1.07-2.26). Greater clinical frailty was associated with lower odds of a prolonged stay (overall p=0.006), which is counterintuitive and which the authors explicitly flag as needing further work rather than action - the obvious explanations include earlier discharge to institutional care and earlier death, neither of which is a good outcome.

The useful part is that all three predictors are knowable on day one, and two of them are not clinical. Social vulnerability being the strongest predictor says the discharge problem is often a housing, income or carer problem that nobody starts working on until the medical issues are resolved.

So the practical change is a timing one: start discharge planning for a person with dementia who arrives with delirium or a thin social situation on the day of admission, not when they are medically fit. In Indian practice, where discharge depends almost entirely on family capacity rather than on funded packages, the equivalent question on day one is who will be at home, and whether they can be.

• Screen for delirium at admission in every person with dementia - it predicts the stay, not just the day
• Ask on day one who the patient lives with and who can provide care after discharge
• Do not wait for medical stability to start discharge planning in this group
• Treat the frailty finding as unexplained; do not use it to reassure about a frail patient's trajectory
• Record the reason for any delay - most prolonged stays here were not driven by unresolved illness

The statistics, in plain English

An odds ratio of 1.74 for social vulnerability means the odds of a prolonged stay were about three-quarters higher, but the outcome is uncommon (5.3%), so the absolute increase is small - roughly from 5 to 8 per 100. The Nosos odds ratio of 1.06 is per unit of the score, so it only becomes substantial across a wide range of disease burden. This is a secondary analysis of a trial not designed to answer this question, and the associations are with factors recorded at admission, so they identify risk rather than explain causation. The inverse frailty association is the clearest sign that unmeasured competing outcomes are at work.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.