STRATIFY Risk Assessment Tool: Scoring, Evidence, and Limits

Between 700,000 and 1 million patients fall in US hospitals every year, by the Agency for Healthcare Research and Quality’s count, and Medicare stopped reimbursing injurious inpatient falls back in 2008. Screening starts, on thousands of wards, with the STRATIFY risk assessment tool.

The tool has a precise pedigree. David Oliver and colleagues built it at St Thomas’ Hospital in London and published it in the BMJ in October 1997, naming it the St Thomas’s Risk Assessment Tool In Falling elderly inpatients. Five bedside questions produce a score from zero to five.

STRATIFY Risk Assessment Tool: Key Takeaways
STRATIFY scores five bedside items, one point each; totals of 2 or more flag high fall risk, per the original 1997 BMJ study.
Expect real-world accuracy near 67 percent sensitivity and 61 percent specificity, well below the original cohort’s 93 and 88.
US hospitals log 700,000 to 1 million inpatient falls a year, and CMS treats serious fall injury as a never event.
NICE guideline NG249 (January 2025) shifts hospitals to multifactorial assessment; run STRATIFY as triage into it, never as the whole program.
Rescore after every fall, ward transfer, and medication or mobility change, and put that cadence in written policy.
Trend STRATIFY tallies monthly beside fall rates per 1,000 bed-days so the screen behaves like a control with an owner and escalation path.

The catch sits in the evidence. The original study reported 93 percent sensitivity, later validations never came close, and the January 2025 NICE guideline moved hospitals away from prediction scores altogether. Used with those limits in view, the five items still make a fast, teachable front door to falls prevention.

Inside the STRATIFY Risk Assessment Tool: Five Items and One Threshold

Each STRATIFY item asks a yes-or-no question a nurse can answer at the bedside without instruments. One point per positive answer, and a total of 2 or more flags the patient as high risk in the original protocol. Scoring a patient takes under a minute once the risk assessment process is familiar.

STRATIFY item	What it asks	Points
Falls history	Did the patient present with a fall, or fall since admission?	0 or 1
Agitation	Is the patient agitated or acutely confused?	0 or 1
Vision	Is the patient visually impaired to the point everyday function suffers?	0 or 1
Toileting	Does the patient need especially frequent toileting?	0 or 1
Transfer and mobility	Do combined transfer and mobility scores land on 3 or 4?	0 or 1

The STRATIFY Risk Assessment Tool at a Glance

Figure 1. The STRATIFY risk assessment tool: five items, a two-point threshold, and sub-minute scoring explain three decades of bedside popularity.

Agitation and toileting need come from nursing observation and can change shift to shift. The mobility item borrows the transfer and mobility scores from the Barthel index, and only combined scores of three or four earn the point. Everything else comes straight from risk identification during admission.

The Falls Problem the STRATIFY Risk Assessment Tool Exists to Solve

Inpatient falls run at three to five per 1,000 bed-days, which compounds to the 700,000 to 1 million falls a year the AHRQ Patient Safety Network reports. One injurious fall adds more than $14,600 in costs and 6.3 days of stay. The Centers for Medicare and Medicaid Services counts serious fall injuries among its never events.

The Falls Numbers Driving STRATIFY Adoption

Figure 2. Four numbers from AHRQ frame the exposure every falls screening program answers to.

Regulators keep the pressure on. The Joint Commission Center for Transforming Healthcare made falls one of its flagship improvement projects after member hospitals logged thousands of injurious falls, and ECRI listed falls among its top patient safety concerns as recently as 2024. A documented screen, built on a conventional risk assessment method, is the minimum defensible position.

Figure	What it says	Named source
700,000 to 1M	Inpatient falls in US hospitals each year	AHRQ Patient Safety Network
3-5	Falls per 1,000 bed-days	AHRQ Patient Safety Network
$14,600+	Added cost of one injurious fall	AHRQ PSNet falls primer
6.3 days	Added length of stay after a fall	AHRQ PSNet falls primer
2008	Year CMS stopped paying for injurious inpatient falls	CMS hospital-acquired conditions rule

Scoring the STRATIFY Risk Assessment Tool: A Worked Example

The arithmetic is easiest to see on a concrete admission. A 78-year-old arrives after a fall at home, presents settled on the ward, wears fully correcting glasses, needs toileting help every two hours, and takes one-person assistance to transfer. Each item gets scored exactly as written, with no partial credit.

STRATIFY item	Finding for this patient	Points
Falls history	Admitted after a fall at home	1
Agitation	Settled and oriented on the ward	0
Vision	Glasses fully correct everyday function	0
Toileting	Needs assistance every two hours	1
Transfer and mobility	One-person assist; combined score of 3	1
Total	Crosses the high-risk threshold of 2	3

A total of 3 clears the threshold of 2, so the patient is flagged high risk and the ward moves to interventions: bed in low position, toileting rounds, mobility referral. Documentation matters as much as arithmetic, and a quantitative score beats adjectives in any later record review. Log the tally in the shared scoring template so trends survive audits.

What Validation Studies Say About the STRATIFY Risk Assessment Tool

Oliver’s own numbers set a high bar. The 1997 BMJ development study reported 93 percent sensitivity and 88 percent specificity in the local validation cohort, figures that made STRATIFY the default falls screen across UK and Commonwealth hospitals within a decade.

Replication told a humbler story. A systematic review of fall risk scales pooled STRATIFY performance at 67.2 percent sensitivity and 61.2 percent specificity with an area under the curve of 0.69, and Billington’s 2012 meta-analysis reached similarly modest conclusions about the threshold of 2. An Italian retrospective study questioned the tool for its inpatient mix outright.

STRATIFY Accuracy Across Two Decades of Validation

Figure 3. The original cohort’s 93 and 88 percent never replicated; pooled estimates run roughly 25 points lower.

A 2023 retrospective cohort at a tertiary hospital in Colombia, published in Scientific Reports, still found the scale significantly associated with falls among hospitalized adults. Setting drives the results: geriatric wards resemble the original cohort, while stroke units and younger populations sit far outside it.

Study	Setting	Sensitivity	Specificity
Oliver et al., BMJ 1997 (local)	Elderly inpatients, London	93%	88%
Pooled systematic review, 2021	Mixed hospital settings	67.2%	61.2%
Billington et al., 2012 meta-analysis	Threshold of 2, multiple cohorts	Modest	Modest
Colombia cohort, 2023	Tertiary hospital adults	Significant association reported	AUC-based

STRATIFY vs Morse Fall Scale: Picking a Falls Screen

Hospitals choosing a falls screen usually shortlist the Morse Fall Scale beside STRATIFY. Morse weighs six items to a 125-point scale, takes longer to complete, and in one re-evaluation on obstetrics and gynecology wards reached an area under the curve of 0.772 at the standard cutoff of 40. Neither instrument scores well enough to run unaccompanied.

STRATIFY vs Morse: Discrimination Compared

Figure 4. The STRATIFY risk assessment tool and Morse both beat chance; neither approaches diagnostic certainty, and setting shifts every number shown.

Dimension	STRATIFY	Morse Fall Scale
Items scored	5, one point each	6, weighted to 125 points
High-risk threshold	2 or more	40 to 45 by hospital policy
Time to score	Under 1 minute	1 to 3 minutes
Origin	Oliver et al., BMJ, 1997	Janice Morse, Canada, 1989
Pooled discrimination	AUC near 0.69	AUC 0.772 in one 2024 re-evaluation

A widely cited review of inpatient falls prevention science places both scales roughly level with structured nurse judgment, which reframes the buying decision. Pick the screen your staff will complete consistently, wire it into a 5×5-style scoring discipline, and spend the saved effort on interventions rather than instrument debates.

Where the STRATIFY Risk Assessment Tool Fits After NICE NG249

In January 2025, NICE published guideline NG249, which asks hospitals to run a multifactorial falls risk assessment for every inpatient aged 65 and over and for higher-risk patients aged 50 to 64. Prediction scores lost their central role because the assessment targets modifiable factors instead of fixed ones.

STRATIFY item	NG249 modifiable factor	Ward intervention it triggers
Falls history	Falls history, causes and consequences	Post-fall review; address the cause found
Agitation	Cognitive impairment	Delirium screen; medication review
Vision	Visual impairment	Optometry referral; lighting and spectacles check
Toileting	Continence problems	Scheduled toileting rounds; continence plan
Transfer and mobility	Postural instability and mobility problems	Physiotherapy referral; mobility aids at reach

The two frameworks fit together better than the headlines suggest. Every STRATIFY item maps onto an NG249 modifiable factor, so a positive screen can trigger the deeper assessment and the interventions the NICE quality standard already expects. UK guidance travels, too: US fall programs built on AHRQ toolkits take the multifactorial route as well.

Running the STRATIFY Risk Assessment Tool on a Live Ward

Rollout succeeds on cadence and ownership before anything else. Score at admission, rescore after any fall, any transfer between wards, and any change in medication or mobility, and set rescoring frequency in policy so it survives staff turnover. An NIHR realist review of acute hospital practice found assessment routines drift fast without named owners.

Task	Owner	Cadence
Score new admissions	Admitting nurse	Within 6 hours of arrival
Rescore after trigger events	Bedside nurse	Same shift as the trigger
Audit scoring consistency	Ward manager	Monthly sample of 10 charts
Trend scores against fall rates	Quality committee	Monthly dashboard review
Review thresholds and policy	Falls committee	Annually and after any serious event

We push clients to treat any screening score as a key risk indicator, with thresholds, owners, and escalation written down. A STRATIFY tally that nobody trends ward by ward stays a form. Charted monthly beside fall-rate KPIs, it starts working like a control.

STRATIFY Risk Assessment Tool: Your Questions Answered

What is the STRATIFY risk assessment tool?

STRATIFY is a five-item falls screening tool developed by David Oliver’s team at St Thomas’ Hospital and published in the BMJ in 1997. Nurses score falls history, agitation, visual impairment, toileting needs, and transfer or mobility difficulty, one point each. Totals of 2 or more flag high fall risk.

What score on the STRATIFY risk assessment tool means high risk?

A score of 2 or more out of 5 marks a patient as high risk under the original protocol. Some hospitals lower the threshold to 1 for frail populations, trading specificity for sensitivity, and any local change belongs in written policy with the risk assessment template staff actually use.

How accurate is the STRATIFY risk assessment tool?

Accuracy depends heavily on the ward. The original 1997 cohort showed 93 percent sensitivity and 88 percent specificity, while pooled estimates across later validations sit near 67 percent sensitivity, 61 percent specificity, and an area under the curve of 0.69. Treat it as a screen that opens a fuller assessment.

Is the STRATIFY risk assessment tool still recommended in 2026?

Guidance has moved toward multifactorial assessment. NICE guideline NG249, published in January 2025, directs hospitals to assess modifiable risk factors for all inpatients 65 and over, displacing standalone prediction scores. STRATIFY survives as a fast triage step that routes patients into the deeper multifactorial workup.

How long does the STRATIFY risk assessment tool take to complete?

Under a minute for a nurse who knows the patient, which is the tool’s strongest selling point. All five items come from observation or the admission record, no equipment or gait testing required. Build it into admission workflows and scenario-based drills so scoring stays consistent across shifts.

Can the STRATIFY risk assessment tool be used outside hospitals?

Validation evidence concentrates in acute and geriatric inpatient settings, and performance drops when the tool travels. Nursing homes saw mixed results in early studies, and community programs lean on the CDC’s STEADI initiative instead. Match any screen to the population it was validated in before adopting it.

What Goes Wrong With STRATIFY Scoring and the Fixes That Work

Most STRATIFY failures are use failures, and they repeat so reliably that auditors can predict them from the ward roster. Six show up in nearly every record review we read; the remedies cost a policy edit and a training huddle, and none waits on procurement.

Pitfall	Root cause	Remedy
Score once at admission, never again	No trigger list in policy	Rescore after falls, transfers, and medication changes
Every elderly patient scores 2+	Threshold too blunt for the ward mix	Pair the screen with multifactorial assessment, per NG249
Scores disagree between nurses	Items interpreted, never trained	Ten-minute calibration drills with sample vignettes
High score, no intervention follows	Screen disconnected from care planning	Auto-trigger toileting rounds and mobility referral at 2+
Original 93 percent sensitivity quoted in policy	Evidence never updated	Cite pooled figures; set expectations at screen level
Forms missing at audit or litigation	No retention or indexing rule	File scores with the admission record, dated and signed

Where Falls Risk Screening Goes After STRATIFY

Expect prediction scores to keep losing ground to embedded analytics. Medication lists, mobility notes, and prior falls sit in the electronic record waiting to be scored automatically, and hospital systems are testing continuous monitoring that flags bed-exit movement in real time. The five STRATIFY items increasingly read like inputs feeding a model, with the paper form as fallback.

Watch the NG249 ripple spread through 2026 and 2027. Commissioners and accreditors tend to import NICE positions within a few audit cycles, and US systems tracking the ongoing prevention journey already grade multifactorial assessment over screening scores. Budget owners should fund intervention capacity, the true bottleneck, ahead of new tools.

Litigation and payment policy will keep the floor firm. Serious fall injuries remain on the never event list, and plaintiff attorneys routinely request screening records in inpatient fall claims. A completed STRATIFY record with dates attached answers those requests early, exactly what a healthy risk management lifecycle is built to do.

None of the coming changes retires the bedside fundamentals. Whatever platform scores the risk, someone still has to fix the lighting, schedule the toileting round, and review the sedatives, the critical components of any risk assessment since long before 1997. Teams that master those will absorb every new tool comfortably.

 

Sharpen Your STRATIFY Risk Assessment Tool Rollout With Risk Publishing

A falls screen should change what happens on the next shift, and that is the standard we hold client programs to. Risk Publishing benchmarks your STRATIFY setup against NG249 and AHRQ expectations, then rewrites thresholds, escalation, and audit cadence into one page a charge nurse can run. Start with our services