The maternal morbidity crisis
Increased maternal morbidity and mortality in the United States is alarming and runs counter to trends in other high-income countries.¹ Annually, nearly 700 women in the U.S. die of pregnancy-related causes, two-thirds of which may be preventable.²
These deaths are the terminal event of severe maternal morbidities (SMM), which are about 100 times more common than maternal deaths.³
Numerous factors contribute to the rise in U.S. SMM rates, such as racial disparities in health status and health care access, increasing C-section rates, advanced maternal age, and growth in obesity and other preexisting comorbidities.³
These escalating trends in maternal morbidity would benefit from a data-driven response. The Centers for Disease Control and Prevention (CDC) used hospital discharge data from delivery episodes to create measures of SMM, including a composite measure encompassing 21 diagnosis/procedure groupings.
However, there is currently no comprehensive set of national SMM measures across the delivery episode and postpartum period to assess gaps.
Measurement framework to address gaps
Building on our experience developing a large set of key performance indicators (KPIs) focused on helping illuminate the issues and opportunities around the opioids crisis, OptumLabs collaborated with experts from Ariadne Labs and Optum to develop a set of KPIs to address the SMM gap.
This novel measure set uses claims of pregnant women from the delivery episode extending out to six weeks postpartum, with the goal of benchmarking performance and identifying opportunities for improvement.
We received input from national clinical and public health experts, including representatives from Centers for Disease Control and Prevention, academia and the California Maternal Quality Care Collaborative. This limited set of claims-based metrics builds on measures developed by the CDC.
Focusing on preventable SMMs, we developed 15 claims-based metrics in five clinical domains:
Eleven of these build on the CDC’s definitions of SMM.⁴ Each measure was evaluated over two time-periods: 1) the delivery inpatient stay (IP) and 2) the delivery inpatient stay + 6-week postpartum period.
The measures were stratified by delivery type (vaginal or C-section) and whether the issue was present on delivery hospital admission.
Severe Maternal Morbidities Measure Framework, 2016–2018
These ICD-10-based metrics included inpatient deliveries from March 2016 through March 2018 in women ages 12–55 enrolled in commercial insurance or Medicare Advantage (MA) plans. Not surprisingly, only a small number of delivery episodes (0.2%) occurred in women with MA coverage.
|Primary Measure / Composite Morbidities||Secondary Measures/Components|
Composite 1: Including congenital heart defect*
Composite 2: Excluding congenital heart defect*
|Embolism and cerebrovascular disorders*||
* Indicates a new OptumLabs KPI not derived from CDC metrics
How delivery experiences contributes to SMM
We evaluated 356,838 inpatient deliveries and 343,545 deliveries with inpatient stay + 6-week postpartum data.
The eclampsia rate during the inpatient stay was 16 per 10,000 deliveries. When we included the inpatient + postpartum period, that rate increased to 34 per 10,000 deliveries, suggesting the importance of monitoring blood pressure in women after discharge.
Hemorrhage and transfusion
The obstetric hemorrhage rate was 504 per 10,000 deliveries during the inpatient stay and 584 during the inpatient stay + postpartum period.
In the inpatient + postpartum period, the obstetric hemorrhage rate was higher for C-section deliveries than vaginal births (631 and 584, respectively).
For all deliveries, the majority of hemorrhage events occurred during the inpatient stay, and 24% were present on admission.
Overall, the transfusion rate was 70 per 10,000 in the inpatient delivery stay; 140 and 43 for C-section and vaginal births in the inpatient + postpartum period, respectively.
The sepsis rate was 11 per 10,000 in inpatient stay and rose to 57 per 10,000 in the inpatient + postpartum period, where 83% of sepsis cases occurred.
In evaluating over 2,000 women who developed sepsis, we identified ≥28 distinct patterns of sepsis onset. Twenty-three percent were readmissions with late sepsis (14% post C-section; 9% post vaginal delivery), suggesting missed opportunities for counselling or monitoring.
28 different patterns of sepsis onset during delivery IP stay to 42 days post-delivery admission
Cardiovascular SMM was 25 per 10,000 deliveries in inpatient delivery + postpartum and were 4 times more likely in C-section verses vaginal deliveries (49 vs. 13 respectively).
Although rare, underlying unrecognized cardiomyopathy is serious and potentially detectible through biomarker screening for brain natriuretic peptide or troponin levels.
Conclusions from building the KPI framework
This KPI framework demonstrates claims can be used to calculate SMM rates, identify patterns and variation and assess problems nationally.
To our knowledge, other existing SMM measurements rely on hospital discharge data that do not include the postpartum period or whether the morbidity events were present on admission.
Although we assessed morbidity events only extending out to six weeks after birth, our findings around the morbidities that occur in these six weeks underscore the importance of monitoring post-discharge to treat preventable conditions.
Comprehensive SMM assessment could be extended back to the prenatal period and out 12 months postpartum like the World Health Organization recommends.
We recognize that identifying SMMs that are present on admission is valuable in differentiating events that occurred during vs. prior to hospitalization so that intervention strategies can be devised for each situation.
These KPI measures and others that may be developed to span the fuller prenatal-postpartum period could be used to spur constructive investigation that may be causal such as provider preparedness, practice structure, hospital-specific issues and access barriers.
Driving dissemination to lower morbidity
OptumLabs has sought to use dissemination mechanisms to share these KPIs in addition to working on other maternal health projects in important and similar research directions.
OptumLabs and Ariadne Labs — with funding from Robert Wood Johnson Foundation’s Health Data for Action Grant program — are working on a study looking at patterns in prenatal care and which patient- and geographic-level factors are associated with receipt of guideline-concordant care.
In addition, a team from UC Health in San Diego is looking at antidepressant use patterns, specifically medication dosing and timing across the prenatal period and impact on both maternal and fetal outcomes.
Combined, all these analytic efforts spotlight our opportunity to help design strategies to reverse the U.S. crisis in maternal morbidity and mortality.
- Omid Ameli, Optum, MD, MPH, OptumLabs
- Christine E. Chaisson, MPH, OptumLabs
- Amber Weiseth, DNP, RNC-OB, Ariadne Labs
- Linda H. Genen, MD, MPH, Optum Population Health Solutions
1. Mann S, Hollier LM, McKay K, Brown H. What we can do about maternal mortality — and how to do it quickly. N Engl J Med. 2018 Nov 1;379:1689-1691.
2. Maternal mortality. Centers for Disease Control and Prevention. https://www.cdc.gov/reproductivehealth/maternal-mortality/index.html. Published September 4, 2019.
3. Severe maternal morbidity in the United States. Centers for Disease Control and Prevention. https://www.cdc.gov/reproductivehealth/maternalinfanthealth/severematernalmorbidity.html. Published January 31, 2020.
4. How does CDC identify severe maternal morbidity? Centers for Disease Control and Prevention. https://www.cdc.gov/reproductivehealth/maternalinfanthealth/smm/severe-morbidity-ICD.html. Published December 26, 2019.