In frail inpatients, comorbidity counts stopped predicting loss of function

Design

Retrospective multicentre cohort

Population

1,299 adults aged 65 or over admitted to acute hospitals in Japan

Primary outcome

Hospital-associated disability (Barthel Index fall of 5 or more)

Effect

Charlson score OR 1.44 (95% CI 1.20 to 1.73) non-frail; 1.11 mildly frail; 0.98 (0.90 to 1.06) frail

This multicentre retrospective cohort followed 1,299 adults aged 65 or over admitted to acute hospitals in Japan, measuring comorbidity with the Charlson index and frailty with the Clinical Frailty Scale. Hospital-associated disability was a fall of at least 5 points in the Barthel Index during admission.

In non-frail patients, each step up in comorbidity score was linked to higher odds of disability (OR 1.44, 95% CI 1.20 to 1.73). In mildly frail patients the link weakened (OR 1.11), and in frail patients it disappeared (OR 0.98). The difference between groups was statistically clear.

The practical point is that once a patient is frail, counting diagnoses adds little to predicting who will lose function in hospital. Frailty itself should drive early mobilisation, nutrition and discharge planning.

• Record a Clinical Frailty Scale score on admission for every patient aged 65 or over.
• Use frailty, not only the number of diagnoses, to identify patients at risk of losing function in hospital.
• Start mobilisation and nutrition plans early in frail patients.
• Measure function, such as the Barthel Index, at admission and discharge.

Why it matters

Risk tools built on disease counts may miss the patients most likely to leave hospital less able than they came in.

Don't overread it

This is an observational Japanese cohort; it shows a change in prediction, not that any intervention prevents disability.

The statistics, in plain English

An odds ratio of 1.44 in non-frail patients means each one-point rise in Charlson score was linked to 44% higher odds of disability. In frail patients the odds ratio of 0.98 with an interval spanning 1 means no association.

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.