Three separate measurement programs, two star rating groups, one dataset.
The Complications and Deaths program is one of the most data-rich in hospital quality reporting. It draws from three distinct measurement approaches — each asking a different question about risk, outcomes, and safety. Together they populate two of the five groups that make up the Overall Hospital Star Rating.
The first approach looks at mortality: how often patients die within 30 days of admission for specific conditions and procedures. The second looks at in-hospital safety: how often patients experience serious, potentially preventable complications during their stay, using a composite of AHRQ Patient Safety Indicators. The third isolates surgical complications specifically for elective hip and knee replacements — one of the most common high-volume procedures in Medicare.
Results across all three approaches are expressed the same way: better than, no different than, or worse than the national rate — based on risk-adjusted statistical comparison with confidence intervals.
Mortality, safety, and surgical outcomes — each measured differently.
These measures estimate how often patients die within 30 days of admission — from any cause — for six specific conditions and procedures, plus a hospital-wide rate. CMS measures death at 30 days rather than in-hospital death because length of stay varies widely across patients and hospitals, making in-hospital death an unreliable comparison point.
Rates are risk-standardized to account for differences in how sick patients were at admission. Each hospital's rate is compared to a national rate to determine if its performance is better than, no different than, or worse than expected. Hospitals with too few cases to reliably compare are flagged rather than rated.
Important: when these same mortality conditions appear in the HVBP dataset, CMS reports them as survival rates, not death rates — the inversion is intentional and worth knowing when comparing across datasets.
- Heart attack (acute myocardial infarction, AMI)
- Heart failure
- Pneumonia
- Chronic obstructive pulmonary disease (COPD)
- Coronary artery bypass graft surgery (CABG)
- Stroke
- Hospital-wide mortality (all eligible admissions)
CMS Patient Safety Indicators are a set of measures developed by the Agency for Healthcare Research and Quality (AHRQ) that flag potentially avoidable complications occurring during inpatient hospital care. Unlike mortality — which looks at outcomes after discharge — PSIs measure what happened while the patient was in the building.
CMS publicly reports two PSI measures. The first is PSI-4: the death rate among surgical inpatients who experienced serious treatable complications — a measure of whether complications, once they occurred, were managed successfully. The second is PSI-90: a composite of eleven component indicators, each measuring a different type of preventable complication or iatrogenic event.
PSIs apply only to Medicare fee-for-service patients discharged from hospitals paid under the Inpatient Prospective Payment System (IPPS). They are calculated from Medicare claims data, not from hospital-reported data — which means they cannot be gamed by what a hospital chooses to document.
Total hip arthroplasty (THA) and total knee arthroplasty (TKA) are among the most commonly performed elective procedures in Medicare. Because they are high-volume and relatively standardized, they are well-suited to complication measurement — a hospital performing 200 hip replacements a year can be compared to another hospital with statistical confidence.
This measure tracks the rate at which patients experience at least one of eight specific complications within defined time windows after their surgery. The eight complications span from infections and mechanical failures to cardiac events and death. Importantly, this is a separate measure from the PSI-90 composite — a patient who experiences a hip/knee complication may also appear in PSI counts, but the measures are calculated independently.
- Heart attack — during admission or within 7 days
- Pneumonia — during admission or within 7 days
- Sepsis or septic shock — during admission or within 7 days
- Surgical site bleeding — during admission or within 30 days
- Pulmonary embolism — during admission or within 30 days
- Death — during admission or within 30 days
- Mechanical complications — during admission or within 90 days
- Periprosthetic joint infection or wound infection — within 90 days
Every measure produces the same three-category comparison.
All three programs use the same display convention on Care Compare and in the underlying data. A hospital's risk-adjusted rate is compared statistically to the national rate. The comparison produces a confidence interval — if the interval is entirely below the national rate, the hospital is performing better. If it straddles the national rate, performance is no different. If it is entirely above, performance is worse.
Claims-based, risk-adjusted, and updated annually.
All three measure programs in this dataset are calculated from Medicare fee-for-service claims — not from data that hospitals submit directly. This is an important distinction: CMS constructs these measures from billing records, applying its own risk models. Hospitals cannot choose which patients to include or exclude.
Collection windows vary by program. Mortality measures use 36 months of data; PSI measures use 24 months; hip/knee complication measures use 36 months. All three are refreshed once per year. The longer collection windows are necessary to produce stable estimates — particularly for conditions like CABG mortality, where even large hospitals may have relatively few eligible cases per year.
The data is published as part of the CMS Provider Data Catalog in the Complications_and_Deaths–Hospital.csv file, with a separate file for PSI-4 and PSI-90 decimal-precision scores used in HACRP calculations.
- CMS Hospital Downloadable Database Data Dictionary, January 2026 — CMS Provider Data Catalog
- AHRQ Patient Safety Indicators Technical Specifications — AHRQ Quality Indicators
- Mortality Measures Methodology Reports — QualityNet.cms.gov
- THA/TKA Complication Measure Methodology Report — QualityNet.cms.gov
- Overall Hospital Quality Star Ratings Methodology — QualityNet.cms.gov