Abstract Purpose Respiratory infections are the most common reason for admission to paediatric intensive care units (PICU). Most patients with lower respiratory tract infection (LRTI) receive broad-spectrum antimicrobials, despite low rates of bacterial culture confirmation. Here, we evaluated a molecular diagnostic test for LRTI to inform the better use of antimicrobials. Methods The Rapid Assay for Sick Children with Acute Lung infection Study was a single-centre, prospective, observational cohort study of mechanically ventilated children (> 37/40 weeks corrected gestation to 18 years) with suspected community acquired or ventilator-associated LRTI. We evaluated the use of a 52-pathogen custom TaqMan Array Card (TAC) to identify pathogens in non-bronchoscopic bronchoalveolar lavage (mini-BAL) samples. TAC results were compared to routine microbiology testing. Primary study outcomes were sensitivity and specificity of TAC, and time to result. Results We enrolled 100 patients, all of whom were tested with TAC and 91 of whom had matching culture samples. TAC had a sensitivity of 89.5% (95% confidence interval (CI₉₅) 66.9–98.7) and specificity of 97.9% (CI₉₅ 97.2–98.5) compared to routine bacterial and fungal culture. TAC took a median 25.8 h (IQR 9.1–29.8 h) from sample collection to result. Culture was significantly slower: median 110.4 h (IQR 85.2–141.6 h) for a positive result and median 69.4 h (IQR 52.8–78.6) for a negative result. Conclusions TAC is a reliable and rapid adjunct diagnostic approach for LRTI in critically ill children, with the potential to aid early rationalisation of antimicrobial therapy.