Severe Maternal Morbidity in California Hospitals: Performance Based on a Validated Multivariable Prediction Model

Area of Science:

• Maternal Health
• Healthcare Quality Improvement
• Public Health

Background:

• Severe maternal morbidity (SMM) is emerging as a critical indicator for assessing maternal healthcare quality.
• Evaluating hospital performance using standardized measures is essential for identifying areas of improvement.

Purpose of the Study:

• To assess California hospital performance using a standardized measure for severe maternal morbidity (SMM).
• To identify variations in SMM rates and hospital performance across different hospital types.

Main Methods:

• Logistic regression models were developed using California maternal hospital discharge data (2016-2017) to predict SMM, adjusted for clinical risk factors.
• Models were tested in 2018 data to evaluate hospital performance, calculating observed-to-expected (O/E) ratios for SMM, including and excluding blood transfusions.
• Hospitals were categorized by type (community, teaching, IDS, IDS teaching) and ranked by O/E ratio to identify over- and underperforming facilities.

Main Results:

• SMM rates (including/excluding transfusion) varied by hospital type: community (1.7%/0.9%), teaching (2.7%/1.5%), IDS (2.3%/1.2%), and IDS teaching (3.0%/1.6%).
• Higher-volume community hospitals showed a notable proportion of underperformers (20.7%/11.0%).
• Overall, 25.3% (14.9%) of California hospitals experienced significantly more SMM events than expected.

Conclusions:

• California hospital discharge data reveal substantial variation in standardized severe maternal morbidity.
• A standardized SMM measure can effectively guide and monitor quality improvement initiatives in maternal healthcare statewide.