HAI Scoreboard: Hospital-Acquired Infection Rates

Every year, about 1 in 31 hospitalized patients acquires an infection they didn't have when they were admitted. This post visualizes CMS hospital-acquired infection data across ~4,789 US hospitals, exploring where hospitals fall on national benchmarks and what that means for patients choosing where to receive care.

Around 4,789 US hospitals report hospital-acquired infection (HAI) data to CMS. Only 211 hospital-infection pairs (less than 1%) are rated "Worse than the National Benchmark" — but the data reveals real variation between hospitals. MRSA bacteremia shows the lowest rates (most hospitals perform well) while C. difficile is the most prevalent HAI by patient volume. The most important finding: your hospital's infection rating is searchable, and it varies enough to matter.

Every year, about 1 in 31 hospitalized patients in the US acquires an infection they didn't have when they were admitted. These healthcare-associated infections (HAIs) — infections contracted in hospitals, nursing facilities, or during outpatient procedures — account for roughly 100,000 deaths annually and billions of dollars in excess healthcare costs.

CMS collects and publishes HAI data for all Medicare-participating hospitals through its Hospital Compare program. Hospitals report six major infection types: Central Line-Associated Bloodstream Infections (CLABSI), Catheter-Associated Urinary Tract Infections (CAUTI), two types of Surgical Site Infections (SSI), MRSA bacteremia, and Clostridium difficile (C. diff). Each is measured using the Standardized Infection Ratio (SIR): the number of observed infections divided by the number predicted based on national baseline rates and patient mix. A SIR of 1.0 means your hospital's rate exactly matches the national average. Below 1.0 is better; above 1.0 is worse.

This post explores where hospitals fall on those benchmarks — and what that means for patients choosing where to receive care.

For each of the six infection types, this chart shows what percentage of hospitals are rated Better, the Same as, or Worse than the national benchmark.

Several patterns emerge:

• "Not Available" dominates. The largest share of hospital-infection combinations lacks a valid SIR rating — often because the hospital had too few procedures or patient-days to produce a statistically reliable estimate. A hospital with no SIR is not necessarily a safe hospital; it may simply be too small to be measured.
• Most rated hospitals are within the normal range. Among hospitals with a valid benchmark rating, the majority fall into "No Different than National Benchmark."
• "Worse" is rare but real. Hospitals rated worse than the national benchmark are a small minority, but they exist across all six infection types. CAUTI (urinary tract infections) and C. diff show the most "Worse" ratings in absolute terms, partly because they are the most common HAIs by volume.
• SSI categories show the most "Better" ratings. Surgical site infection rates have improved substantially over the past decade as surgical technique and antibiotic prophylaxis have improved; many hospitals now perform better than historical baselines.

Percentage of hospital-infection pairs rated "Worse than the National Benchmark," by state.

The geographic pattern is noisy — HAI rates are hospital-specific, and state averages can be driven by a handful of hospitals in each direction. That said, a few patterns are consistent with other quality measures:

• States with higher overall hospital quality ratings (from CMS Hospital General Information) tend to have lower rates of HAI "Worse" designations.
• Some states show elevated rates of worse-than-benchmark infections, which may reflect older facilities, different patient populations, or reporting differences.
• The "Not Available" category (excluded from this percentage) is higher in rural states with many small critical-access hospitals, which may make state comparisons misleading.

Each point is a hospital. The x axis shows its C. diff SIR; the y axis shows its MRSA SIR. The reference lines at 1.0 mark the national average. Color reflects the hospital's overall benchmark rating across all six infection types.

MRSA and C. diff are often cited as the most consequential HAIs because they are drug-resistant or treatment-difficult, and they spread easily in healthcare settings. A few observations:

• Most hospitals cluster near the origin (SIR < 1 for both), meaning the average hospital performs better than the historical baseline — a sign of genuine improvement in infection prevention over the past decade.
• Outliers in the upper-right are concerning. Hospitals with high SIR scores for both MRSA and C. diff simultaneously face a systemic challenge — it isn't random variation in one metric, but a pattern suggesting infection prevention gaps.
• Green dots (Better overall) tend to cluster below the median for both infections. Hospitals that perform better than benchmark across the board tend to have MRSA and C. diff rates below the national average too — suggesting that overall infection prevention culture matters more than individual measure management.

These 20 hospitals have the highest average SIR scores across all reported infection types (minimum 3 reported measures). These are hospitals where observed infections consistently exceed the predicted rate.

Several important caveats for interpreting this list:

• Patient mix affects SIRs. The Standardized Infection Ratio is designed to adjust for patient complexity, procedure volume, and type of ICU — but no adjustment is perfect. Hospitals treating the most complex, immunocompromised patients may show elevated SIRs even with good care.
• Small hospitals can have noisy SIRs. A hospital with 5 predicted infections and 8 observed has a SIR of 1.6 — but that's based on small numbers. CMS has minimum thresholds, but some outliers still reflect statistical noise.
• This list is a starting point, not a verdict. CMS's full Hospital Compare tool allows patients to see individual infection ratings plus the detailed confidence intervals and footnotes that explain data limitations.

The takeaway: before a planned procedure, it's worth checking your hospital's HAI ratings — not to make a binary good/bad judgment, but to have an informed conversation with your care team about infection prevention practices.

The HAI data is public and searchable through CMS's Care Compare tool. For planned surgeries or procedures, patients can:

1. Look up their hospital's infection ratings at medicare.gov/care-compare — search by hospital name, see each HAI measure.
2. Ask about specific prevention protocols. For surgical cases: does the hospital use SCIP (Surgical Care Improvement Project) protocols? For MRSA: is pre-admission MRSA screening standard practice?
3. Ask about CLABSI bundles. The evidence-based "central line bundle" (hand hygiene, barrier precautions, chlorhexidine skin antisepsis, optimal catheter site selection, daily review of catheter necessity) has been shown to reduce CLABSI rates by up to 70%. Hospitals that follow it consistently should be able to confirm this.
4. Consider volume. Hospitals that perform more of a given procedure tend to have lower complication rates, including infection rates. A center that does 400 hip replacements per year has likely optimized its infection prevention workflow more than one that does 40.

The data exists to make these conversations possible. Hospitals that perform well on HAI metrics tend to have a culture of measurement and improvement — which correlates with better outcomes broadly.

All data from the CMS Healthcare-Associated Infections - Hospital dataset, current as of April 2024–March 2025.

• Hospitals covered: 4,789 unique hospitals with at least one HAI measure reported.
• Measures: Six SIR measures — HAI_1SIR (CLABSI), HAI_2SIR (CAUTI), HAI_3SIR (SSI Colon), HAI_4SIR (SSI Abdominal Hysterectomy), HAI_5SIR (MRSA), HAI_6SIR (C. diff).
• Benchmark comparison: CMS assigns each hospital-measure pair to one of four categories based on whether the SIR is statistically Better, No Different, or Worse than the national benchmark, or Not Available (insufficient data).
• State map: Percentage of hospital-measure pairs rated "Worse than the National Benchmark" among all rated pairs (excluding "Not Available"). Only 2-character state codes included (territories excluded).
• Scatter chart: Restricted to hospitals with valid MRSA and C. diff SIR scores; SIR values above 10 excluded as outliers affecting chart scale.
• Worst-performing hospitals: Average SIR across all reported measures; requires ≥3 measures with a numeric SIR. Color coding based on the number of "Worse than National Benchmark" designations across all six measures.