Rationality of Empirical Antibiotic Prescribing and Its Association with Length of Hospital Stay in Patients with Severe COVID-19

Excessive empirical antibiotic prescribing during the COVID-19 pandemic raised concerns regarding antimicrobial resistance, particularly in hospitals with limited microbiological diagnostic capacity. This study aimed to describe empirical antibiotic prescribing patterns, evaluate the rationality of antibiotic use, and examine its association with length of hospital stay among patients with severe-to-critical COVID-19 at a private secondary-care hospital in Batam City, Indonesia. A retrospective observational study was conducted using the medical records of 120 patients admitted between April and June 2022. Antibiotic rationality was evaluated according to drug selection, dose, route of administration, dosing interval, and treatment duration using the local Hospital Antibiotic Use Guideline, the Indonesian COVID-19 Management Guideline, and relevant international guidelines. Length of stay was compared between patients receiving antivirals with antibiotics and those receiving antivirals without antibiotics, followed by multivariable linear regression. Of the 120 patients, 99 received antibiotics and 21 received no antibiotics. The most frequently prescribed regimen was levofloxacin combined with meropenem (20.83%), followed by levofloxacin monotherapy (18.33%). Among antibiotic recipients, appropriateness was 64.6% for drug selection, 79.8% for dose, 96.0% for route of administration, 94.9% for dosing interval, and 100% for treatment duration. Patients receiving antivirals with antibiotics had a shorter mean length of stay than those receiving antivirals without antibiotics (9.2 ± 2.4 versus 12.8 ± 3.1 days; p = 0.002). After adjustment for age, sex, comorbidity, oxygen-support level, and antiviral type, antibiotic co-administration remained associated with a 3.6-day shorter hospital stay (adjusted β = −3.6; 95% CI, −4.8 to −2.3; p 0.001). However, this association should not be interpreted as evidence of a causal therapeutic benefit because of the retrospective design, residual confounding, informative censoring, and the absence of microbiological confirmation. These findings support strengthening pharmacist-led prescription review, guideline-based antibiotic audits, renal-dose adjustment, and antimicrobial stewardship in hospitals with limited microbiological facilities.