Utilizing Gradient Boosting Machine Algorithms to Analyze the Causal Impact of Antimicrobial Agents on Recovery Times of fever and Hospital Lengths of Stay in Non-Critical Pediatric Patients with Acute Viral Respiratory Tract Infections

Abstract

Abstract Background The widespread prescribing of antibiotics in paediatric patients with upper respiratory tract infections (URTIs) exacerbates global concerns about antimicrobial resistance. The purpose of this study is to look at the effect of antibiotics on hospital stay duration and fever resolution in paediatric patients who have been diagnosed with viral infections using a comprehensive multiplex polymerase chain reaction (PCR) respiratory panel. Methods From October 2016 to December 2021, we performed a careful retrospective cohort analysis on paediatric patients confirmed with viral infections using nasopharyngeal aspirates at Imam Abdulrahman Bin Faisal Hospital. We accurately balanced the cohorts’ getting antibiotics versus those not receiving them using the Gradient Boosting Machine (GBM) technique for propensity score matching, ensuring a standardised mean difference threshold of 0.15. Results Human Rhinovirus/enterovirus was the most common (44.5%) among 238 paediatric patients, followed by respiratory syncytial virus (18.1%). In 8.4% of cases, significant co-infections were found, primarily involving HRV/EV and RSV. According to our survey-weighted linear regression model, antibiotic administration increased hospital LOS by an average of 2.19 days (p-value 0.00). Surprisingly, the presence of diarrhoea resulted in a 2.26-day decrease in LOS, while higher albumin levels resulted in a 0.40-day decrease in LOS. Neither fever nor CRP levels had a significant effect on LOS. The Kaplan-Meier curve comparing Time to Recovery from Fever for Abx0 (antibiotic-free) and Abx1 (antibiotic-received) groups indicated no significant difference in recovery length (p-value: 0.391), with a hazard ratio of 0.84 (CI: 0.57–1.2), confirming that antibiotic administration had no effect on recovery duration. Conclusions Our findings provide an insightful look at the common antibiotic prescribing practises for paediatric patients with acute viral respiratory infections. Although antibiotics did not speed up recovery, they were associated with longer hospital stays. The study emphasises the importance of clinicians exercising caution when giving antibiotics to paediatric patients with proven viral infections, especially when their clinical status is non-critical.