Quick answer
Fall tracking should distinguish the underlying event, reports about that event and residents who experienced one or more events. Preserve duplicate reports without counting the fall twice. Define the observation period, included locations and resident-day method before calculating a rate, and keep local quality-improvement measures separate from formal MDS or public-reporting specifications.
Give reports and events different identifiers
A fall may produce an incident form, a nursing note and a later witness statement. Those records can describe one event. Another fall involving the same resident later that day is a separate event. A counting process that deduplicates only by resident and date can lose that recurrence.
The AHRQ Falls Management Program measurement chapter distinguishes falls, residents who fall, residents with two or more falls and falls with serious injury. It emphasizes consistent tracking and warns that improved reporting can initially increase observed counts. Use that bounded measurement context; the local reconciliation below is an original operational example.
Immediate resident assessment, care, notifications and required reporting follow approved procedures. Data cleanup must not hold those actions. The purpose here is to make the resulting quality information reproducible.
Reconcile a small incident ledger
Assume this fictional facility uses a calendar month and has already established its clinical fall definition. Its reviewer receives six reports:
| Report | Resident | Described event | Reconciliation |
|---|---|---|---|
| R1 | A | May 3, 08:10, room | Event F1 |
| R2 | A | May 3, 08:10, room | Additional account of F1 |
| R3 | A | May 3, 20:40, dining area | Separate event F2 |
| R4 | B | May 9, 14:20, corridor | Event F3 |
| R5 | C | May 31, 23:45, room | Event F4; entered June 1 |
| R6 | B | May 9, time initially uncertain | Reviewer confirms additional account of F3 |
There are six reports, four distinct falls and three residents who fell. One resident, A, had two falls. The late entry R5 still describes a May event under this example’s event-date convention. Saving the report-entry date remains necessary to explain why an earlier May extract contained only three events.
R6 cannot be merged merely because B already had a report that day. The example assumes a reviewer confirmed the relationship from actual source details. When identity remains uncertain, preserve the uncertainty and show the provisional count’s limitation.
State exactly what each denominator means
For illustration, the facility defines a local event rate as confirmed falls occurring within its included setting during the month divided by included resident-days, multiplied by 1,000. It documents how admission, discharge and temporary absence days contribute. This is an explicitly chosen local measure, not an asserted CMS formula.
Suppose its reviewed census extract contains 2,400 included resident-days. Four events produce 4 ÷ 2,400 × 1,000 = 1.67 falls per 1,000 resident-days, rounded for display. Using the six reports would produce 2.50 and overstate event frequency.
For a separate local question, suppose 96 unique residents were present during the month. Three residents with a fall produce 3 ÷ 96 = 3.125%, displayed as 3.13%. That resident percentage and the event rate answer different questions. Neither denominator is average daily census, the end-of-month census or the number of incident forms.
Document whether repeat admissions identify the same person in the unique-resident count. Preserve facility boundaries when local medical-record numbers repeat. A regional total requires compatible definitions and underlying counts before aggregation.
Keep location and period exclusions inspectable
A hospital event reported after return may be important to clinical care while falling outside a locally defined in-facility event numerator. Preserve the event and its setting; record why it is included or excluded for this particular measure. Never delete clinically relevant history simply to reconcile a dashboard.
For formal assessment work, use the applicable instructions rather than importing the local monthly rate. The CMS RAI manual page identifies the current manual and versions. The qualified MDS team determines which event information applies to its item and observation period.
Maintain a cutoff and revision history. The preliminary May count, the June discovery of R5 and the revised May count should remain distinguishable. Without that history, a reviewer cannot tell whether the trend changed because care changed, reporting improved or an old period was corrected.
Require a buyer demonstration that exposes bad counting
Provide the six reports and the reviewed census fixture to a software supplier. Ask the ordinary quality user to reconstruct the four events, open each source report and identify A’s recurrence. R1 and R2 should remain readable after they are linked. F2 must survive despite sharing A’s event date.
Then enter R5 after the original cutoff and rerun the report. Retain the prior output and reconcile the revision. Deliberately remove one day of census input; the application should expose incomplete denominator coverage rather than present an apparently final rate.
Save the event crosswalk, inclusion decisions, census reconciliation, formula, reporting version and actual exports as acceptance evidence. Resolve discrepancies at the record level. A chart with a declining line is useful only when staff can explain what was counted and whether the comparison period was built the same way.
Questions & answers
Should a duplicate incident report be deleted?
Preserve its authorship and information, link it to the underlying event and document the reconciliation decision. Removing a duplicate from an event count does not mean erasing the report.
Are the number of falls and the number of residents who fell interchangeable?
No. A resident can have multiple distinct falls in the period. Report event counts, unique residents and residents with recurrent events separately.
Can a local fall rate be presented as an MDS quality measure?
Only after applying and verifying that measure’s actual specification. A local resident-day rate uses its own defined scope and denominator and should be labeled accordingly.
Primary sources
- AHRQ Falls Management Program: Data Collection and Analysis Using TRIPS
- CMS RAI manual version and current downloads
Use the cited source and its applicable version when making a facility-specific decision. These guides explain operational evaluation methods; they are not individualized clinical, coding, or legal advice.