Even if about one million women have miscarriages each year, in most cases where miscarriages occur, there are few complications related to complications of pregnancy. Experts told ProPublica that the need to explore this stage is urgent given the national abortion ban undermines the way maternal health care is done.
Although most early miscarriages have no complications to resolve, patients with heavy bleeding may suffer from unsuitable treatments – including a procedure called dilation and curation, D&C, now tangled in legislation that prohibits miscarriage. As the woman narrates being left to shed dangerous blood, Propublica tells the story of a mother dying at a Houston hospital while seeking miscarriage care, journalists seek a way to gain a broader understanding of what is happening in the state.
We consulted dozens of researchers and clinicians to develop our methodology and learn how to view early miscarriage outcomes in the emergency room.
Our latest analysis of hospital discharge data in Texas found that the number of blood transfusions during emergency room visits increased by 54% in August 2022 after the state made abortion a felony.
More than a dozen experts told ProPublica that the number of emergency room employment during the emergency room also increased by 25%, suggesting that women may return to poorly ill hospitals.
Experts say this spike is a disturbing indicator of delay in care.
Experts say the most effective way to prevent severe blood loss during miscarriage is D&C, which uses suction to remove remaining tissue, thereby closing the uterus. This procedure is also used to terminate pregnancy.
Dr. Elliott Main, a maternal bleeding specialist and former medical director of the California Maternal Quality Care Collaboration, said doctors working under the abortion ban are now postponing the prolonged interventions that disrupt patients — “until they actually bleed.”
These findings add to Propublica’s growing report, suggesting that maternal outcomes have become worse after the state’s abortion ban. In February, we published an analysis of two-year pregnancy loss hospitalization and found that the state’s ban on post-abortion sepsis has increased by more than 50%. The study focused only on hospitalizations in Texas hospitals. However, many clinicians and researchers we spoke with told us that the focus will limit our perception of miscarriage care in early pregnancy. In shorter emergency department visits, most people with pregnancy complications may appear rather than hospitalization.
This method lists our steps to study early miscarriage outcomes for the emergency department to help experts and interested readers understand our approach and its limitations.
Confirm the first three years of emergency visits
We purchased seven years of inpatient and outpatient patient encounter records from the Texas Department of Health Services Hospital and Outpatient Surgery Center. These records contain data used to access, as well as information about encounters, including recorded diagnostic and execution procedures, as well as demographic information and billing data for some patients.
We will analyze the access to diagnose pregnancy loss in inpatient and outpatient datasets. We followed a methodology where maternal researchers have used for years to identify “miscarriage outcomes” – examples of pregnancy loss in less than 20 weeks, including diagnoses such as ectopic pregnancy and miscarriage. Researchers have usually identified these cases to exclude them from indicators that evaluate complications of delivery. By contrast, we focused only on the analysis to those encounters with pregnancy loss diagnosis. Medical experts suggest that it is possible that more women self-manage abortions at home. Because self-managed abortions can appear like spontaneous abortion, we cannot distinguish these patients in the data.
We also restrict analysis to emergency room visits or emergency room starts with hospitalization. The state’s outpatient data also includes data for encounters and bedridden surgery centers for outpatient procedures, and we do not include focus on emerging hospital care. Ultimately, our analysis focused on 35,500 pregnancy visits that entered the hospital through the emergency room each year, excluding a few (about 1,400 times a year) hospitalizations without starting in the emergency room.
To limit our analysis to pregnancy loss in the first three months, we looked for diagnostic codes indicating the week of pregnancy. If multiple weekly gestation codes were recorded during the accommodation process, we adopted the latest code. We excluded any line with a weekly gestational code of 13 weeks or more, which marked the beginning of the second trimester. The vast majority (78%) of emergency treatment pregnancy loss codes indicate week-of-girl pregnancy or no week-of-girl diagnosis code. We included these visits in the first pregnancy period. Clinicians tell us that a pregnant patient came to the emergency room during the second trimester and it is unlikely that a doctor will be appointed to determine the gestational age. Since pregnancy losses in the second or three months of pregnancy are more severe, and because it is easier to establish gestational age in pregnancy, it is likely that emergency room doctors will be able to establish gestational age during treatment in this case.
We then filtered the access lists for patients who were female and between 10 and 54 years old to rule out potential errors. This deleted 2,692 visits, or 1.1% of all visits we determined.
The number of first-time hospitalizations in the emergency room was relatively stable, during Covid-19-19. In 2022, the first year after the state passed a six-week abortion ban, the number of encounters increased by 11%. In 2023, a year after state abortions were criminalized, they rose again, up 25% from pre-hybrid levels.
While we can determine the increase in visits, we are unable to identify patients within the scope of visits, which means we cannot say how many of these visits represent the same person who returned to the emergency room multiple times to cause the same pregnancy loss. Live births in Texas have increased since the state banned miscarriages — about 2.7% in 2022, down slightly in 2023 compared to the pre-cycle average. However, this growth in birth-but by extension, the increase in pregnancy does not explain the rate of change in emergency visits, which is far beyond.
Clinicians also told us that the threshold for diagnosing pregnancy loss increases after the state prohibits abortions. To evaluate how many relevant visits our analysis might ignore, and whether more access is missing after changes in hospital policy, we looked for access without a pregnancy loss code, but diagnosed “threatening abortion” or “early pregnancy bleeding” indicating uterine cramps or early pregnancy bleeding. Since clinicians told us that these diagnoses may range from light spots to obvious bleeding, and that bleeding in pregnancy is common and does not always indicate ongoing miscarriage, we did not include these visits in the main analysis. However, we also determined that visits to these codes increased by 23% – from an annual average of 70,936 to 2023, to 87,431 in 2023.
Identify blood transfusions
Next, we identified pregnancy loss through blood transfusion, which usually indicates a risk of blood loss.
For our hospitalization dataset, procedures performed during hospitalization were recorded as ICD-10-PCS codes, and we identified access to blood transfusions using a list of codes defined by the Centers for Disease Control and Prevention. The outpatient dataset using the current program terminology code has only one code for blood transfusion (36430).
Prior to Covid-19, there were an average of 840 pregnancy emergency room visits and blood transfusions each year. In 2022, the first year after the state passed its first abortion ban, blood transfusions rose to 1,076, a 28% increase from the previous year. By 2023, the number climbed to 1,290 in the first year after abortion was criminalized, a 54% increase before Bixing. In 2023 blood transfusions, this is 450 more than the previous spin average.
Even with the increase in the number of visits to the emergency room, the percentage of visits to blood transfusions increased, from 2.5% in the pre-grocery year to 2.8% in 2022, while 3% in 2023 also increased – suggesting that the increase in blood transfusions may not be explained by the increase encountered only.
Experts reviewing ProPublica data want to know whether the increase in blood transfusion may be driven by more women experiencing ectopic or molar pregnancy complications, which are rare and infeasible pregnancy, in which case the likelihood of blood transfusion is much higher than spontaneous crimes. The data has not been resolved. The increase in the number of blood transfusions lost in pregnancy is even higher when we exclude ectopic and molar pregnancy diagnosis – by 2023, it has risen by 61%.
To understand whether the number of blood transfusions in other mothers’ visits increased during the same period, we also used a federal approach to study blood transfusions in delivery events to identify birth complications. In hospital births, the number of transfusions increased by 6.7% in 2022, while the number of transfusions increased by 9.9% in 2023, while the average of the pre-Hing average increased by a margin of more than a smaller margin, much smaller than the increase in hospitalizations expected for pregnancy loss.
Sophie Chou contributed data reports.