AI-based patient monitoring for fall prevention in stroke patients: a pilot study at a Malaysian acute stroke unit.
Monica Danial, Chee Toong Chow, Meng Hui Lim and 3 others
PMID 41121354WHAT IT FOUND
An AI monitoring system detected all bed exits in 30 stroke patients, triggering alerts in 1,439 events.
Staff responded to half. One fall occurred, compared to six in the same ward the previous year. No control group was used.
Key findings
01The AI system detected 100% of bed-exit events with a false alarm rate of 4.66% across 1,439 alerts.
02One fall was recorded during the study, compared to six falls in the same ward in 2023.
03Staff responded to 50.9% of the alerts generated by the system.
STILL TO COME
How it was doneWhat they found
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
No control group or randomization was used, so the reduction in falls cannot be definitively attributed to the AI system rather than other changes in care or chance. The sample size was small (n=30) and consent was difficult to obtain, limiting generalizability. The comparison is against historical data from a previous year, which introduces potential confounding factors such as changes in patient acuity or staffing. The study did not assess outcomes like injury severity, length of stay, or cost-effectiveness. Staff responded to only about half of the alerts, raising questions about alert fatigue or workflow integration.
Declared interests
The study was funded by SmartPeep Pte Ltd, the company that developed the AI system being evaluated.
The easy way to misread this
Do not conclude that the AI system caused the reduction in falls. The study lacked a control group and compared results to historical data from a different year, meaning other factors like changes in staffing or patient mix could explain the difference.
Summarised by AI from the full paper, without a clinician reviewing it. Check it against the source before it changes what you do. Read it on PubMed →