Clinical Decision Support Tool Predicts Resource Utilization

Using predictive analytics technology, clinical decision support tools can help organizations assess resource utilization for elective surgeries during the COVID-19 pandemic, revealed a study published in JAMA Network Open.

With the onset of the COVID-19 pandemic, many hospitals and health systems were faced with the challenge of planning for strains on resources – including fluctuating numbers of available beds, personal protective equipment and ventilators, and staff shortages.

“As hospitals prepared for possible surges of infected patients requiring admission and possible intensive care (ICU) stay, entire institutions and health systems took stock of their resources to meet an uncertain demand,” researchers stated.

“National specialty societies examined the lessons from their counterparts abroad and, along with federal and state agencies, issued guidelines for the evaluation of the urgency of procedures.”

During the initial wave of infections, many health systems severely restricted elective surgical procedures. As the surge of infections begins to slow in some parts of the country, hospitals are starting to address the backlog of operations that were postponed. However, organizations still need to prepare for potential surges in cases and renewed demand for hospital resources.

While clinical decision support systems are prevalent in several areas of care delivery, the team noted that there are few surgical clinical decision support tools to predict resource utilization.

“Most work has been specialty and/or procedure specific or has addressed questions of comparative effectiveness,” researchers stated.

“The role of predictive analytics in improving operating room efficiency has been touted for day-of-staffing and resource-allocation decisions. However, there remains a need for a CDS that could be used across all specialties to evaluate the demand for resources imposed by increasing case volumes.”

The group set out to see whether they could use historic data to develop a clinical decision support tool for resource utilization during the COVID-19 pandemic. Researchers built a predictive model using retrospective data from the EHRs of patients undergoing elective procedures from January 1, 2017 to March 1, 2020.

For each patient, researchers sought to predict four outcomes: overall hospital length of stay, ICU length of stay, mechanical ventilator requirement, and discharge to a skilled nursing facility.

The results showed that overall, the models performed very well. The strongest models were found for predicting need for ICU, with an area under the curve (AUC) of 0.94, and need for ventilators, with an AUC of 0.92. Predictions were weaker but still accurate for discharge to a skilled nursing facility (AUC 0.84) and long versus short ICU stay (AUC 0.76).

The most predictive variables were demographic, service utilization, and procedural factors, researchers noted. Significantly, clinical information was not one of the top predictors, showing the ability to use easily accessible information available at the time of scheduling to develop a comprehensive clinical decision support tool.

Researchers also integrated the predictive model into a dashboard that shows scheduled cases over the next month and is refreshed every morning. The interactive nature of the dashboard enables decision-makers to assess subsamples based on week, location, service line, or procedure grouping.

The team expects that this clinical decision support tool can help organizations identify patients at highest risk to avoid exceeding hospital capacity.

“We believe that this tool, in conjunction with knowledge of currently available resources—including overall hospital and ICU beds, ventilator use, and area SNF capacity—will help leadership decide whether a case should proceed or be postponed,” researchers stated.

“The clinical decision support tool is applicable across our 3 hospitals, all surgical specialties, and all patient age groups. While different specialties likely have different risk factors, a machine learning approach allows the model to find the appropriate degree of heterogeneity.”

The study was limited in that the clinical decision support system was developed using information generated in a single hospital system. While this may limit the tool’s validity in other health settings, the team noted that using similar predictor variables should enable the tool to perform comparably in similar settings.

The team implemented the clinical decision support tool and associated dashboards in their own institution on June 17, 2020. Since deploying the tool, researchers have made modifications to speed up the data flow process and allow users to assess the performance of the risk model.

The study’s findings show that clinical decision support tools, combined with advanced analytics algorithms, could help organizations stay ahead of potential surges in healthcare utilization.

“The framework that we present for building a formalized CDS tool for surgical resource utilization, in conjunction with the workflow of integrating EHR data with a dashboard, is highly replicable,” researchers concluded.

“As institutions attempt to chart a safe and sustainable path to a new normal, this work illustrates how surgical teams and hospital leadership can do so in a data-driven way, generating a learning health care environment.”