Adverse Event Measures
Adverse event measures are OASIS-derived quality indicators that flag rare, potentially preventable negative outcomes, such as emergent care for an injury caused by a fall or a substantial decline in function during the episode. Unlike publicly reported outcome measures, they are designed as internal warning signals: each flagged case is meant to trigger a chart review, not a rate comparison.
What adverse event measures capture
Adverse event measures identify episodes where OASIS data at transfer or discharge indicates something went wrong that good care might have prevented. Classic examples include emergent care for injury caused by a fall, emergent care for wound infection or deteriorating wound status, substantial decline in activities of daily living, and development of a urinary tract infection. These originated in CMS's Outcome-Based Quality Monitoring (OBQM) reporting, the companion to Outcome-Based Quality Improvement (OBQI), and agencies access them through their quality reports in iQIES, the CMS system that replaced CASPER for home health reporting.
How they differ from outcome and process measures
Outcome measures, such as improvement in ambulation, describe how often patients get better and are risk adjusted for fair comparison, with several publicly reported on Care Compare. Process measures describe whether specific care steps happened, like timely initiation of care. Adverse event measures are different in kind: they capture low-frequency bad events, and because the events are rare, the rates are statistically noisy, especially for smaller agencies. CMS built them for case-level investigation rather than benchmarking. A single quarter's uptick may mean nothing; a flagged case that reveals a missed fall risk intervention means a great deal.
How to use adverse event reports
A practical review cycle looks like this:
- Pull adverse event reports from iQIES on a regular schedule, at least quarterly
- Identify each flagged episode and conduct a focused chart review
- Verify OASIS accuracy first, since a data entry error can create a phantom event
- For confirmed events, determine whether the care plan addressed the relevant risk
- Feed recurring themes into QAPI as performance improvement projects
The review should involve both QA staff and a clinician who can judge whether the event was plausibly preventable given the patient's condition.
Common pitfalls
The biggest mistake is treating adverse event rates like outcome rates. With small denominators, one or two events can swing a percentage wildly, so a rate that doubled may reflect two patients, not a systemic failure. The opposite error is dismissing every flag as noise and never opening the charts. Also watch for documentation-driven artifacts: inaccurate OASIS responses at transfer or discharge can both create false events and hide real ones, which is why accuracy review comes before root cause analysis. Finally, do not ignore unavoidable events entirely; even a non-preventable fall can reveal a home safety or teaching gap worth closing.
Frequently asked questions
Are adverse event measures publicly reported on Care Compare?
No. Adverse event measures are internal quality surveillance tools available to agencies through their CMS reports. Public reporting on Care Compare draws on a separate set of OASIS-based, claims-based, and HHCAHPS measures. That said, the same underlying events can show up indirectly in publicly reported hospitalization measures.
What triggers an adverse event flag?
Specific OASIS response patterns at transfer or discharge, such as emergent care for an injury caused by a fall or data indicating substantial functional decline during the episode. Because the flags come from assessment data, OASIS accuracy directly determines whether your adverse event reports reflect reality.
Our adverse event rate jumped this quarter. Should we panic?
Not before reviewing the cases. Adverse events are rare, so small agencies can see large percentage swings from one or two episodes. Review each flagged chart, confirm the OASIS data is accurate, and judge preventability case by case. A pattern across multiple confirmed events warrants a QAPI project; a single event warrants a case-level fix.