Understanding and Improving Pediatric Discharge Antimicrobial Prescriptions 

Patients at LPCH have been commonly prescribed antimicrobials upon discharge, and the Antimicrobial Stewardship committee (ASP), composed of ID physicians and pharmacists at LPCH, retrospectively reviews the prescriptions. Since 2021, ASP identified 20% of antimicrobial prescriptions ordered during the transition of care as suboptimal. There is some discrepancy in drug choice, dose, frequency, duration, and formulation between what was ordered and what is recommended by ASP. Suboptimal prescriptions lead to poor patient outcomes. For example, recently, a 6-year-old boy was discharged on two incompatible IV medications, which led to clogging of his central line and resulted in readmission and surgical removal of the line. Such outcomes are unfavorable outcomes for both the patient and the hospital.

To understand the drivers of suboptimal prescriptions, the SURF team analyzed data of over 5000 prescriptions since January 2021 and engaged with stakeholders at LPCH, including physicians, care teams, and pharmacists. The duration was found to be the most common discrepancy, followed by dose and frequency. By indication, prescriptions for perioperative prophylaxis were commonly found to be suboptimal. These discrepancies can be attributed to the differences among clinical judgment of providers.

Among all antimicrobials, Augmentin  was found to be one of the most commonly prescribed medications, with the highest rate of discrepancy at approximately 40%. Multiple formulations, confusing guidelines, and  inconsistencies in the prescribing process between inpatient and outpatient settings cause many of these discrepancies. Improving Augmentin prescriptions is important to eliminate avoidable GI side effects and prevent treatment failure. Recommended formulation, dose, and frequency of Augmentin for specific indications is clearly defined in guidelines. Automated calculation of the same using an algorithm can eliminate these discrepancies.

Focusing on Augmentin as a proof of concept, the team developed an algorithm to calculate recommended formulation, dose, and frequency based on the indication for use. The algorithm compared the results with historical order data and tagged discrepancies in the orders. The results of the algorithm verification were validated against the human audit data. The algorithm was found to be 90% accurate, producing the exact same results as the human audit 74% of the time. Further, the algorithm also helped identify errors in human audit documentation in 16% of the prescriptions. In 10% of the prescriptions, clinical factors involved in the human audit process could not be captured by the algorithm, which will be further explored.

The team is now focused on developing a quick, easy-to-use tool that can determine the recommended formulation, dose, and frequency to be prescribed based on the indication being treated and patient weight. The tool is envisioned to provide real-time automated decision support to prescribers, improving the efficiency and accuracy of prescriptions for all antimicrobials. The tool has immense potential for improving all pediatric prescriptions at transitions of care across the nation as well in other care settings. Optimized prescribing will result in better hospital resource utilization and patient outcomes.