When children are readmitted to the hospital after discharge, outcomes are worse in nearly every measurable way.
Readmitted children spend twice as many days in the hospital, incur higher hospital costs, and have higher mortality rates compared with those who are not readmitted.
While only 2.5% of children’s hospital patients are readmitted within 30 days, a new benchmarking tool could help reduce the number even further.
The Pediatric Readmissions Encounter Predictor (PREP) shows the likelihood of unplanned readmissions within 30 days for 324 diagnosis groups at four levels of severity.
That gives hospitals one of the most granular looks at readmission probabilities currently available. The predicated rates, calculated from median rates across the country, give children’s hospitals a reliable benchmark to compare their performance and discover areas for improvement.
“PREP shows children’s hospitals whether they’re doing better or worse than expected for kids with the same condition,” said Greg Attard, a data analyst at the Children’s Hospital Association who co-developed PREP and led the study testing and validating the model. “Because it shows predictions for all severity levels, hospitals can get really granular with their improvement efforts.”
Measuring against national benchmarks
Researchers used the 2019 National Readmissions Database to calculate expected probabilities of readmission for every diagnosis-severity combination within the All-Patient Refined Diagnosis Related Groups (APR-DRG) system.
After testing, PREP showed a 0.73 area under curve (AUC) score — the probability it can accurately assess a patient’s readmission risk level — indicating a solid statistical model.
The true power of PREP is showing hospitals where change can make the biggest difference.
A hospital treating children with pneumonia, for example, can compare its own readmissions against the national benchmark. If its rate is higher, that signals an opportunity to investigate and improve.
“These benchmarks help leaders ask the right questions,” Attard explained. “Why are our numbers higher than average? What are other hospitals doing differently? Where can we intervene earlier?”
Attard said hospitals could also apply PREP benchmarks to calculate observed-to-expected ratios, compare performance across subpopulations, and even track disparities by race and ethnicity.
PREP can also assist with planning and resource allocation during surges and seasons. For example, knowing that a high percentage of RSV patients with severity number 3 will be readmitted in 30 days enables the hospital to have more staff and equipment in place when those patients return.
“Children’s hospitals have limited resources, and they need to allocate them as best as possible,” Attard said. “Anything we can do to help them improve that is beneficial.”
While CHA’s Pediatric Health Information System (PHIS®) gives children’s hospitals apples-to-apples benchmarking in clinical and operational measures, PREP provides an added layer of specificity for readmissions.
Hospitals that use PHIS can apply PREP benchmarks directly to their own patient cohorts and compare actual readmissions against the PREP benchmark.
Impact on children’s health
Together, PHIS and PREP give hospitals a powerful way to evaluate quality and target improvements.
Quality improvement efforts informed by benchmarks can reduce avoidable hospital stays and improve the patient experience.
Over time, PREP can help children’s hospitals identify national priorities for collaborative work on conditions where many hospitals struggle and shared solutions are needed.
“Benchmarks are not the end goal,” Attard said. “They are a starting point for hospitals to make changes that reduce avoidable readmissions and improve the care experience for kids and families.”
Access the free Pediatric Readmissions Encounter Predictor.
