nuttapong punna/istock via Getty

How Risk Stratification Can Help Combat Preterm Birth Disparities

Parkland Center for Clinical Innovation is identifying at-risk pregnancies and addressing premature birth disparities through analytics-based risk stratification.

In 2021, preterm births, which occur when an infant is born before 37 weeks gestational age, affected roughly one out of every 10 infants born in the US. The rate of preterm births rose 4 percent in a single year from 10.1 percent in 2020 to 10.5 percent in 2021, according to statistics from the Centers for Disease Control and Prevention (CDC).  

That same year, African American women experienced preterm birth rates of 14.8 percent, which was approximately 50 percent higher than those experienced by Hispanic or white women at 10.2 and 9.5 percent, respectively. 

Socioeconomic factors, such as race and income, are known to be associated with preterm birth, alongside additional risk factors such as prior preterm birth, tobacco or substance use, and carrying more than one baby. 

The CDC recommends strategies such as seeking prenatal care, avoiding alcohol and drugs, quitting smoking, and waiting at least 18 months between pregnancies to help prevent preterm birth, but disparities in preterm birth rates have led some healthcare organizations to pursue data- and even artificial intelligence (AI)-driven maternal and infant health approaches to promote health equity. 

Parkland Center for Clinical Innovation (PCCI) and Parkland Hospital in Dallas, Texas have developed one such program, which uses social determinants of health (SDOH) to identify at-risk pregnant patients and enable early interventions to help reduce preterm births. 

Yolande Pengetnze, MD, a pediatrician and vice president of Clinical Leadership at PCCI, spoke with HealthITAnalytics about how her organization’s risk prediction model and text message-based patient education program helped target clinical and population-level interventions to address preterm births. PCCI’s approach led to reductions in preterm delivery rates by up to 20 percent and an eight percent increase in prenatal doctor visits since the program’s launch in 2018. 

IDENTIFYING DISPARITIES AND RISKS FOR PRETERM BIRTHS 

When discussing disparities in preterm birth rates, Pengetnze highlighted the importance of understanding the risks associated with preterm birth. 

“Children who are born prematurely are born immature. That means their lungs, their brains, their digestive system, their livers, [and] their kidneys do not function optimally. Their immune systems do not function optimally,” she explained. 

The CDC recommends that if possible, infants should be delivered after 39 weeks of pregnancy have been completed, as key growth and organ development happen up until that time. At 37 weeks, an infant’s brain, lungs, and liver are not fully developed, but with proper medical care, babies born after that point may have minimal or no complications. 

“Typically, children who are born between 35 and 37 weeks could be okay. They can be discharged home. But before 35 weeks, they have one of the highest risks of having complications, and before 32 weeks, even higher,” Pengetnze said. 

“The younger they are, the higher the risk of complications including cerebral palsy, developmental delay, respiratory problems that might require them to be put on a ventilator, digestive gut issues, feeding issues that might require them to be fed through IVs, infections, and a whole host of issues including vision and hearing issues down the line.” 

These outcomes afflict those infants that survive, but a significant number do not. 

Data from CDC’s National Center for Health Statistics indicate that approximately 16 percent of infant deaths are caused by preterm birth and low birth weight, underscoring the need for effective preterm birth prevention strategies. 

Developing these strategies, however, requires accurate risk assessment and stratification across populations. To do so, Pengetnze indicated that identifying whether a pregnant patient meets the criteria for known clinical risk factors of preterm birth is critical. 

“There are two major risk factors, [the first of] which is that someone who has had a history of spontaneous preterm delivery is at higher risk of having another one,” she stated. “But if it's a first-time pregnancy, or in other subsequent pregnancies, that history may be lacking.” 

The second known risk factor is related to cervical length during pregnancy, as a short cervix is associated with higher risk for preterm delivery, Pengetnze explained. 

However, she indicated that these two criteria are only present in 20 percent of preterm births, while the remaining 80 percent appear to be caused by additional risk factors. 

“That's where the challenge lies in being able to identify those patients who are at high risk, because by not having one, two, or three strong criteria, we need to figure out a way of calculating that additional risk and adding that up to identify those folks who are going to end up having a preterm delivery,” Pengetnze said. 

An additional hurdle lies in the timeliness of preterm birth prediction. Patients need to be identified early enough in their pregnancies so that care teams can work to intervene and prevent a preterm delivery, but this requires that patient risk be stratified within the first or second trimester. This type of analysis can pose a major challenge, especially for patients in high-risk groups who often do not receive prenatal care before the second trimester. 

Despite this, Pengetnze emphasized that such issues are not insurmountable. 

“Some of these risk factors are modifiable. Some may not be modifiable, but they're actionable. They will point us to the right place to work with these pregnant women. Those are truly the challenges: identifying the risk factors that are driving risk in these subgroups and identifying those high-risk women in a timely manner to give us time to [intervene],” she stated. 

CONSOLIDATING RISK FACTORS AND PREDICTING OUTCOMES 

To modify these risk factors and make them actionable, Pengetnze pointed out that a risk stratification approach must combine risk profiles effectively for each patient. 

“Our strategy is always to build a model that takes into consideration different risk factors from different walks of life,” she explained. “We want to be able to capture medical risk factors, health services utilization risk factors, behavioral risk factors, social risk factors and put them together so that we can understand the whole person and intervene at different points in the patient's life.” 

To do so, PCCI’s risk model uses multi-variable regression, which allows care teams to consider multiple risk factors. Pengetnze indicated that this approach is better for premature birth risk stratification than a rule-based model, which would classify individual risk based on one or two risk factors. 

“[Using a multi-variable approach,] we can actually capture small increments in a risk factor and add them together to identify which patients are at higher risk based on those small increments,” she said.  

Risk factors are generated from various data types and sources, including claims data, health services utilization, mental and behavioral health information, demographics, and SDOH data from databases such as the American Community Survey. CDC data is also pulled for additional insights into the socioeconomic context in which patients live.  

“We put all those together into these multi-variable regression models, and we optimize them to predict the outcome that we are interested in. In this case, we predicted the risk that the pregnant woman would end up having a delivery at or before 35 weeks of completed gestational age,” Pengetnze said. 

LEVERAGING RISK-BASED EDUCATION, ENGAGEMENT, AND SOCIAL INTERVENTIONS AT SCALE

Once a patient’s risk is identified, interventions are tailored to that patient as best as possible.  

“The way we've approached our intervention is risk-based,” Pengetnze noted. “We risk stratify these women into four categories: those who are very high risk, high risk, medium risk, and low risk.” 

She explained that PCCI’s preterm birth prevention intervention modalities are centered around education and patient engagement. Through these methods, the organization focuses on encouraging prenatal visit attendance, teaching patients about the risks of preterm birth and signs of preterm labor, and providing resources for smoking, alcohol, and drug use cessation. 

To support these efforts, PCCI has designed a text messaging-based program to help engage pregnant patients.  

Patients in the program receive text messages multiple times per week depending on their risk level. Low-risk patients receive messages containing general information about pregnancy two times per week, while higher-risk patients may receive general education about pregnancy in addition to information about preterm birth up to five times per week. 

Some examples of texts the program sends out are: ‘ask your doctor about the signs of labor at your next doctor’s visit. It is helpful to know the signs of labor to know when to start preparing,’ and ‘remember to take your prenatal vitamins every day! Prenatal vitamins help you and baby get all of the vitamins and minerals you need for healthy growth.’ 

“That risk-driven approach seems to be fairly well-received by these moms, who are very satisfied in the program,” Pengetnze indicated. “They remain on the program all the way through the end of the pregnancy and two months after delivery, [and] they seem to appreciate it. Over 75% of women are highly satisfied with the program as it is.” 

In addition to focusing on patient education and engagement, PCCI also works to activate these patients’ providers, she explained. Identifying preterm birth risk factors and creating patient risk profiles is part of preventing preterm births, but without care teams who are informed about a patient’s risk and can help tailor program interventions, that work would be largely ineffective. 

PCCI hopes to deploy a third level of interventions aimed at tackling social needs that pregnant patients are faced with, which could help alleviate stress and improve wellbeing. 

For now, though, Pengetnze reported that PCCI’s approach has seen some significant successes. The program is going into its fifth year of tracking its impact on prenatal care attendance, preterm birth rates, and patient perceptions of how it affected their pregnancies.  

To evaluate prenatal care attendance, PCCI analyzed the outcomes of patients that were enrolled in the program versus those that were of similar risk, but not enrolled.  

“When we compare those two groups, women that participate in the program had a higher prenatal care attendance ranging from 8 percent to 15 percent depending on the year of evaluation,” Pengetnze said. “Then, looking at preterm delivery [rates], we saw a 2 percent absolute drop, which corresponds to about a 20 percent relative drop in preterm deliveries among pregnant women that passed in the program versus those that did not. 

She further explained that in general, these trends show improvements in the severity of preterm delivery among those participating in the program versus not. This indicates that those who participate in the program, even when they have a preterm delivery, do so later and deliver infants of a later gestational age. 

Later delivery results in newborns who are more mature than those delivered by patients not participating in the program, which Pengetnze highlighted can have significant impacts on health outcomes for preterm infants.  

In terms of patients’ perception of and satisfaction with their participation in the program, she stated that 75 to 80 percent of patients report learning something from the program, that their participation helped them be prepared for their pregnancy, and that they would recommend the program to friends and family. 

Pengetnze further indicated that a program like PCCI’s is scalable, and encouraged other healthcare organizations to explore how a similar approach could help reduce preterm births in their own patient populations through a strong focus on addressing disparities. 

“Preterm birth is preventable… There are ways that we can work through the different risk factors to help bring pregnancies to term,” she stated. “Given the cost to society, the cost to the children, to their families, not only the financial cost but the societal cost, in general, I think [preventing preterm birth] is an effort that is worth putting into place.” 

Next Steps

Dig Deeper on Population health management