Falls Rate

Falls Rate per 1,000 Occupied Bed Days, Injurious Falls and the Cost of Inpatient Falls

A hospital falls rate divides the inpatient falls reported in a unit and period by the occupied bed days (patient days) in that unit and period, times 1,000; an injurious falls rate keeps that denominator and counts only falls that caused harm. The AHRQ Preventing Falls in Hospitals toolkit prefers the rate to a simple count, which rises and drops with ward occupancy whatever the quality of care. In England, falls were the most commonly reported patient safety incident in acute and community hospitals in 2015/16, and NHS Improvement put the cost to hospitals of reported inpatient falls at around £630 million a year, so the rate underpins both safety monitoring and costing.

What counts as a fall and what counts as a bed day

A falls rate is only as reliable as the agreement on what is counted. The AHRQ toolkit paraphrases the March 2012 National Database of Nursing Quality Indicators (NDNQI) definition, under which a patient fall is an unplanned descent to the floor with or without injury, and assisted falls are included. The toolkit warns that borderline cases, such as a patient eased to the floor by staff, are where definitions diverge, and that staff may under-report them if they fear their unit will compare badly.

The denominator is exposure time. The toolkit counts occupied beds at the same point each day, such as the midnight census, and sums the counts over the month to give occupied bed days.

Severity turns the falls rate into an injurious falls rate. NDNQI grades injury as none, minor, moderate, major or death; the National Reporting and Learning System in England uses no, low, moderate, severe harm and death, and the National Audit of Inpatient Falls recommends that a trust-wide patient safety group reporting to the board regularly review falls, and falls with moderate harm, severe harm and death, per 1,000 occupied bed days. The toolkit also suggests tracking repeat falls, because one patient who falls several times can move a unit's rate alone.

Calculating the rate per 1,000 occupied bed days

The falls rate is an event rate per unit of exposure time, scaled by 1,000 to give readable ward figures. The same formula serves for all falls and for injurious falls.

$$R = \frac{F}{B} \times 1000, \qquad R_{\text{inj}} = \frac{F_{\text{inj}}}{B} \times 1000$$

where $R$ is the falls rate per 1,000 occupied bed days, $F$ is the number of falls reported in the unit and period, $F_{\text{inj}}$ is the number of those falls that caused injury, $R_{\text{inj}}$ is the injurious falls rate and $B$ is the occupied bed days for the same unit and period.

The toolkit's example is a 30-bed unit with three falls in April and a daily census summing to 879 occupied bed days, out of a possible 30 × 30 = 900. Recomputed, 3 ÷ 879 = 0.003413, and 0.003413 × 1,000 = 3.41 falls per 1,000 occupied bed days; the toolkit rounds the intermediate figure and reports 3.4. One more fall would have moved the rate to 4 ÷ 879 × 1,000 = 4.55; the toolkit advises following the trend because rates fluctuate from month to month. It recommends monthly rates on a run chart and names u-charts and g-type charts (statistical process control).

Worked example of two wards with the same bed days

All figures here are illustrative. Two wards each record 10,000 occupied bed days in a year. Ward A, an acute medical ward, admits 1,000 patients with a mean stay of 10 days; 62 of them fall, producing 80 falls, 24 of them injurious. Ward B, a rehabilitation ward, admits 500 patients with a mean stay of 20 days; 60 of them fall, producing 90 falls, 18 of them injurious.

The rates per bed day are close, yet a patient admitted to Ward B is about twice as likely to fall during the stay, mainly because stays are twice as long. None of these figures shows which ward gives safer care, because rehabilitation encourages mobility and the patient mix differs.

If Ward A introduced a prevention programme costing £30,000 a year that lowered its falls rate from 8.0 to 6.4, a rate ratio of 0.80, with occupancy unchanged, the falls prevented would be as follows.

$$\Delta F = \frac{(R_0 - R_1) \times B}{1000} = \frac{(8.0 - 6.4) \times 10000}{1000} = 16$$

where $\Delta F$ is the number of falls prevented in a year, $R_0$ and $R_1$ are the falls rates without and with the programme, and $B$ is the annual occupied bed days. Using NHS Improvement's average of £2,600 per fall as the unit cost, the costs avoided are 16 × 2,600 = £41,600, so on these assumptions the costs avoided exceed the programme cost, although most of the avoided cost is the value of bed days released, not cash, and NHS Improvement describes its figure as indicative and not for local decisions without local length-of-stay data. If the rate fell by only 10%, 8 falls would be prevented, avoided costs would be 8 × 2,600 = £20,800, and the net cost of £9,200 (30,000 minus 20,800) would give 9,200 ÷ 8 = £1,150 per fall prevented. A cost per fall prevented has no agreed benchmark, so comparison with a cost-effectiveness threshold needs falls translated into QALYs and a cost per QALY.

Falls rate and the proportion of patients who fall

The Cochrane review by Cameron and colleagues defines the rate of falls as the total number of falls per unit of person time that falls were monitored, compared with a rate ratio, and the risk of falling as the number of people who fell once or more, compared with a risk ratio. The proportion of patients who fall is therefore a cumulative incidence over stays of differing length, so it mixes daily risk with length of stay (see the worked example), while the falls rate counts every event.

The two can move differently. In care facilities, in residents with low vitamin D levels, the review found that supplementation probably reduces the rate of falls (rate ratio 0.72, 95% CI 0.55 to 0.95) but probably makes little or no difference to the risk of falling (risk ratio 0.92, 95% CI 0.76 to 1.12). Costs attach to each fall, so an economic model built on fallers alone understates the events avoided when repeat falls are common.

Statistical analysis of falls rates

Falls are counts of recurrent events, since most patients never fall and a few fall several times. Robertson and colleagues compared Andersen-Gill and marginal Cox models with negative binomial regression in two trials of a home exercise programme; these methods allow for frequent, recurrent events with a non-normal distribution and adjust for each participant's follow-up time. The Cox models' assumptions were violated in one trial, and the authors recommended negative binomial regression, which was also easier to use. A Poisson regression with bed days as the exposure also gives an incidence rate ratio but assumes the variance equals the mean; repeat fallers make counts overdispersed, so its confidence intervals are too narrow unless robust standard errors are used.

Six hospital trials in the Cochrane review tested multifactorial interventions, five of them cluster randomised; pooling the five that reported rate data, the review found these interventions may reduce the rate of falls (rate ratio 0.80, 95% CI 0.64 to 1.01; low-quality evidence). In the FallSafe evaluation by Healey and colleagues, reported fall rates dropped in the intervention units (adjusted rate ratio 0.75, 95% CI 0.68 to 0.84), injurious fall rates showed no significant change (0.86, 95% CI 0.71 to 1.03), and the share of staff certain that falls had been reported rose from 60% to 77%. Reported falls fell even as reporting improved; the injurious falls estimate was imprecise, which matches the AHRQ warning that few injurious falls make change hard to detect.

Why hospital falls rates are hard to compare

The AHRQ toolkit states that there are no national benchmarks, partly because hospitals treat different patients, and that rates calculated in different ways cannot be compared. A cited NDNQI study found rates from 1.30 falls per 1,000 patient days in intensive care units to 4.54 on medical and 7.15 on rehabilitation units.

The National Audit of Inpatient Falls published locally reported falls rates per 1,000 occupied bed days in 2015 but stopped in its 2017 report. It cited National Quality Board guidance that incident data should never be presented as actual incidents, reported that triangulation with the National Hip Fracture Database suggested local numbers were unreliable as incidence data, and noted that rates reflect patient demographics and specialist services such as rehabilitation or dementia care. Fair comparison needs at least stratification by unit type; risk adjustment or an observed-to-expected ratio needs risk factors for every patient on the unit.

What inpatient falls cost hospitals

NHS Improvement's 2017 report counted around 250,000 falls reported to the National Reporting and Learning System in 2015/16 across acute, mental health and community hospitals, and estimated an average cost of £2,600 per fall and a total of around £630 million a year. The extra stay assumed for falls with no or low harm (8.2 and 9 days) was the most uncertain input; cutting it by half or a quarter gave totals of £354 million and £494 million. Wider health and social care costs are excluded.

Morello and colleagues measured the burden directly in a cost-of-illness study of 12 acute wards in six Australian hospitals, with costs linked for three. Of 27,026 admissions, 966 (3.6%) involved at least one fall. After adjustment, patients who fell stayed 8 days longer (95% CI 5.8 to 10.4) and incurred additional hospital costs of AUD 6,669 (95% CI 3,888 to 9,450) at 2013 prices, and the authors concluded that programmes should target all falls, not only injurious ones.

Cost-effectiveness of preventing inpatient falls

Haines and colleagues evaluated a patient education programme using a randomised trial of 1,206 acute and rehabilitation inpatients, from the health service perspective with the hospital stay as the time horizon. Among cognitively intact patients it cost AUD 294 per fall prevented and AUD 526 per faller prevented (2008 values); the programme had a 52% probability of being both more effective and less costly than usual care. Decision tree analysis showed it would save money where at least 4.0% of cognitively intact patients would otherwise fall, so the ward's baseline proportion of cognitively intact patients who fall, a risk measure and not the falls rate, decides whether the programme pays for itself.

NHS Improvement estimated that multifactorial assessment and intervention of the FallSafe kind could reduce falls by 25 to 30%, but its savings are gross figures that exclude implementation and running costs. It also judged zero falls unrealistic, because hospitals trade patient mobility against fall risk, some falls are part of rehabilitation, and eliminating all risk would cost too much. NICE guideline NG249 (2025), which replaced CG161, recommends comprehensive falls assessment and management for hospital inpatients aged 65 and over, or 50 to 64 at higher risk, and advises against falls risk prediction tools.

Common misreadings of a falls rate

Most errors treat a reporting measure as an exact count of events. The AHRQ toolkit notes that rates often rise when an improvement programme starts, because staff report falls they previously missed.

1. Reporting read as safety. A lower reported rate may reflect weaker reporting, and a higher one better reporting.
2. Injurious falls alone. Few events make change hard to detect.
3. Deaths taken as complete. NHS Improvement notes that deaths after falls are likely to be underestimated in incident reports.

Sources

• Agency for Healthcare Research and Quality. Preventing Falls in Hospitals: A Toolkit for Improving Quality of Care. Section 5: How do you measure fall rates and fall prevention practices? AHRQ Publication No. 13-0015-EF. Rockville, MD: AHRQ; 2013 (content last reviewed January 2013). https://www.ahrq.gov/patient-safety/settings/hospital/fall-prevention/toolkit/measure-fall-rates.html
• Cameron ID, Dyer SM, Panagoda CE, Murray GR, Hill KD, Cumming RG, Kerse N. Interventions for preventing falls in older people in care facilities and hospitals. Cochrane Database of Systematic Reviews. 2018;9:CD005465. https://doi.org/10.1002/14651858.CD005465.pub4
• Haines TP, Hill AM, Hill KD, Brauer SG, Hoffmann T, Etherton-Beer C, McPhail SM. Cost effectiveness of patient education for the prevention of falls in hospital: economic evaluation from a randomized controlled trial. BMC Medicine. 2013;11:135. https://doi.org/10.1186/1741-7015-11-135
• Healey F, Lowe D, Darowski A, Windsor J, Treml J, Byrne L, Husk J, Phipps J. Falls prevention in hospitals and mental health units: an extended evaluation of the FallSafe quality improvement project. Age and Ageing. 2014;43(4):484-491. https://doi.org/10.1093/ageing/aft190
• Morello RT, Barker AL, Watts JJ, Haines T, Zavarsek SS, Hill KD, et al. The extra resource burden of in-hospital falls: a cost of falls study. Medical Journal of Australia. 2015;203(9):367. https://doi.org/10.5694/mja15.00296
• National Institute for Health and Care Excellence. Falls: assessment and prevention in older people and in people 50 and over at higher risk (NG249). London: NICE; 2025. https://www.nice.org.uk/guidance/ng249
• NHS Improvement. The incidence and costs of inpatient falls in hospitals. London: NHS Improvement; July 2017. https://web.archive.org/web/20201229074130/https://improvement.nhs.uk/documents/1471/Falls_report_July2017.v2.pdf
• Robertson MC, Campbell AJ, Herbison P. Statistical analysis of efficacy in falls prevention trials. Journals of Gerontology, Series A. 2005;60(4):530-534. https://doi.org/10.1093/gerona/60.4.530
• Royal College of Physicians. National Audit of Inpatient Falls: audit report 2017. London: RCP; 2017. https://www.hqip.org.uk/wp-content/uploads/2018/02/national-audit-of-inpatient-falls-audit-report-2017.pdf