Parkland Community Health Plan uses tech to prevent preterm births
Parkland Community Health Plan, Parkland Center for Clinical Innovation and Parkland Hospital have found a way to prevent dangerous preterm births by using machine learning and smart phone texting.
Dallas-based Parkland Community Health Plan, Parkland Center for Clinical Innovation and Parkland Hospital have found a way to prevent dangerous preterm births by using machine learning and smart phone texting.
The direct engagement with at-risk expecting women resulted in a 24 percent increase in prenatal visit attendance; a 27 percent reduction in preterm births under 35 weeks; and a 54 percent reduction in baby costs for the first year of life, according to early findings from a year-long pilot study conducted by PCHP and PCCI.
“That’s a cost savings of more than $1 million in just this initial cohort,” says Steve Miff, president and CEO of PCCI, noting that the optimal results required minimal engagement from the clinical teams.
“PCCI’s mission is to help support those in our under-served communities, and those economically challenged are more likely to experience preterm births and be overwhelmed by the huge risks and costs that take place when a baby is born early,” he says.
For the pilot, PCHP and PCCI determined which pregnant women were at risk, prompting them via smart phone texts to get the care they needed. The texts included doctor appointment reminders, nutrition tips, what to expect notifications and other patient-specific messages. Miff says texts were used because the women were most receptive to this form of communication.
PCHP and PCCI launched the Preterm Birth Prevention Program in March 2018, leveraging machine learning-driven predictive models that gather data from medical and pharmacy claims, eligibility data, community and census data to drive early identification of at-risk pregnant women. PCHP and PCCI risk-stratified 26,000 pregnancies, with 819 of the pregnant women enrolling in the text messaging program, the organizations said.
“It all started with one patient and their need for truly comprehensive care, which wasn’t possible in the current healthcare environment,” says PCCI, which spun off from Parkland Health System in 2012. Setting out to connect communities, PCCI built a data bridge between hospitals and communities to best care for underserved patients and was the first company to deploy such a program and its underlying technology across Dallas.
The Preterm Birth Prevention Program relies on using social determinants of health (SDOH) with measurable impact and advanced data analytics to prevent preterm births. “We used a deep understanding of where these expecting mothers are coming from,” Miff says, who attributes the success of the program to developing and building an algorithm based on the neighborhood block in which a woman lives, not the zip code designation. “Zip codes weren’t specific enough,” he says.
“The amazing part about this is we built this model purely on Medicaid claims,” Miff says, which have a 90-day lag— a lot of time to lose in a pregnancy—but despite that, the program was still able to achieve success.
Over the next couple years, PCHP and PCCI plan to enhance the prediction accuracy of the analytics by using data from electronic medical records. The program also plans to factor mental health into the predictive modeling.
In 2017, preterm births affected 10 percent of babies born in the U.S. in 2017, according to the Centers for Disease Control and Prevention, while in 2015, preterm birth and low birth weight accounted for about 17 percent of infant deaths.
The direct engagement with at-risk expecting women resulted in a 24 percent increase in prenatal visit attendance; a 27 percent reduction in preterm births under 35 weeks; and a 54 percent reduction in baby costs for the first year of life, according to early findings from a year-long pilot study conducted by PCHP and PCCI.
“That’s a cost savings of more than $1 million in just this initial cohort,” says Steve Miff, president and CEO of PCCI, noting that the optimal results required minimal engagement from the clinical teams.
“PCCI’s mission is to help support those in our under-served communities, and those economically challenged are more likely to experience preterm births and be overwhelmed by the huge risks and costs that take place when a baby is born early,” he says.
For the pilot, PCHP and PCCI determined which pregnant women were at risk, prompting them via smart phone texts to get the care they needed. The texts included doctor appointment reminders, nutrition tips, what to expect notifications and other patient-specific messages. Miff says texts were used because the women were most receptive to this form of communication.
PCHP and PCCI launched the Preterm Birth Prevention Program in March 2018, leveraging machine learning-driven predictive models that gather data from medical and pharmacy claims, eligibility data, community and census data to drive early identification of at-risk pregnant women. PCHP and PCCI risk-stratified 26,000 pregnancies, with 819 of the pregnant women enrolling in the text messaging program, the organizations said.
“It all started with one patient and their need for truly comprehensive care, which wasn’t possible in the current healthcare environment,” says PCCI, which spun off from Parkland Health System in 2012. Setting out to connect communities, PCCI built a data bridge between hospitals and communities to best care for underserved patients and was the first company to deploy such a program and its underlying technology across Dallas.
The Preterm Birth Prevention Program relies on using social determinants of health (SDOH) with measurable impact and advanced data analytics to prevent preterm births. “We used a deep understanding of where these expecting mothers are coming from,” Miff says, who attributes the success of the program to developing and building an algorithm based on the neighborhood block in which a woman lives, not the zip code designation. “Zip codes weren’t specific enough,” he says.
“The amazing part about this is we built this model purely on Medicaid claims,” Miff says, which have a 90-day lag— a lot of time to lose in a pregnancy—but despite that, the program was still able to achieve success.
Over the next couple years, PCHP and PCCI plan to enhance the prediction accuracy of the analytics by using data from electronic medical records. The program also plans to factor mental health into the predictive modeling.
In 2017, preterm births affected 10 percent of babies born in the U.S. in 2017, according to the Centers for Disease Control and Prevention, while in 2015, preterm birth and low birth weight accounted for about 17 percent of infant deaths.
More for you
Loading data for hdm_tax_topic #better-outcomes...