- Design
- Systematic review and meta-analysis of interventional studies
- Population
- 16 studies, 7 pooled, of sensor-based fall prevention in hospitals
- Primary outcome
- Fall rates, number of fallers and fall-related injuries
- Effect
- Fall rate IRR 0.77 (95% CI 0.38 to 1.58); injuries IRR 0.98 (0.81 to 1.18)
A systematic review included 16 studies of sensor-based fall prevention systems in hospitals, randomised and quasi-experimental, with seven in the meta-analysis.
Pooled results showed no significant effect on fall rates (incidence rate ratio 0.77, 95% CI 0.38 to 1.58; I² 92%), on the number of patients who fell (OR 1.24, 0.76 to 2.02) or on fall-related injuries (IRR 0.98, 0.81 to 1.18). Machine-learning-based systems were linked to lower fall rates (IRR 0.50, 0.32 to 0.79), as was mandated implementation (IRR 0.51, 0.37 to 0.70), though not to fewer patients falling. Certainty of evidence was very low to low.
Heterogeneity was high and subgroup findings come from very few studies. The practical point is that buying alarm technology does not substitute for the basics of hospital fall prevention, and the subgroup signals deserve proper trials before any investment.
- Sensors did not clearly reduce falls, fallers or injuries when studies were pooled.
- The machine-learning subgroup looked better, from very few studies.
- Keep the basics: toileting rounds, footwear, mobility aids and medication review.
- Ask for local fall and injury data before buying new alarm systems.
Why it matters
It questions a common hospital investment before it is adopted at scale.
Don't overread it
Very low to low certainty from seven pooled studies; subgroup findings are exploratory.
The statistics, in plain English
An I² of 92% means the studies disagreed greatly, so the pooled rate ratio of 0.77 is not a reliable single estimate. A subgroup result from a handful of studies is a hint, not a finding.
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.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for geriatrics, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free