Differences among hospitals in Medicare patient mortality

Area of Science:

• Health Services Research
• Medical Quality Assessment
• Public Health

Background:

• Hospital mortality rates are a key indicator of healthcare performance.
• Variability in death rates among hospitals may reflect differences in patient severity, care quality, or other factors.
• Medicare data provides a large dataset for analyzing patient outcomes.

Purpose of the Study:

• To analyze variations in hospital death rates for Medicare patients (age ≥ 65) using fiscal year 1984 data.
• To compare different mortality measures and their impact on hospital rankings.
• To investigate geographic patterns in hospital death rates.

Main Methods:

• Utilized hospital discharge abstract data for acute care hospitals treating Medicare patients.
• Calculated four mortality rates: inpatient deaths, 30-day post-discharge deaths, and deaths within 30 days or 95th percentile length of stay post-admission.
• Adjusted for patient age, race, and sex to determine the probability of observed deaths.

Main Results:

• Significant inpatient death rate variations (p < .05) were found for 22 of 48 conditions and overall.
• Medical conditions contributed more to high-variation deaths than surgical conditions.
• Correlations between different mortality measures were moderate to high (Spearman's ρ = 0.54–0.99).
• Hospitals with low death probabilities showed significant geographic clustering (p < .01).

Conclusions:

• Current hospital discharge abstract data is insufficient to determine if variations in death rates are due to illness severity or care quality.
• The observed variability and geographic clustering suggest further investigation into hospital death rates is warranted.
• Hospital death rates may serve as a screening tool for potential quality issues, but careful selection of mortality measures is crucial.