Abstract
BACKGROUND: Acute rheumatic fever is an immune-mediated condition triggered by Streptococcus pyogenes sore throat and possibly skin infection, with a substantial burden in resource-limited settings. Clinical decision rules (CDRs) are commonly used to guide antibiotic treatment of sore throat based on signs and symptoms, but their diagnostic accuracy varies by study and setting. This work aimed to assess the accuracy of multiple CDRs in Fiji to diagnose S. pyogenes sore throat. METHODS: We conducted a prospective diagnostic accuracy study at two primary healthcare centres in Suva, Fiji, enrolling children aged 5-15 years presenting with sore throat. Clinical features were assessed, and two throat swabs were collected from each participant for S. pyogenes detection using culture and a point-of-care nucleic acid amplification test (NAAT). Six CDRs were evaluated against NAAT and culture as reference standards. FINDINGS: Of 250 participants, S. pyogenes was detected among 31.7% (95% CI: 26.0-37.9) by NAAT and 10.4% (95% CI: 7.6-15.8) by culture. The Fiji CDR demonstrated high sensitivity (98.7%, 95% CI: 93.1-100 vs. NAAT; 100%, 95% CI: 86.8-100 vs. culture) but very low specificity (4.7% (95% CI: 2.1-9.1) vs. NAAT; 4.0% (95% CI: 1.9-7.5) vs. culture). All CDRs had poor discriminatory power (area under receiver operating characteristic curve: 0.48-0.55). INTERPRETATION: CDRs cannot accurately diagnose S. pyogenes sore throat in this tropical setting where rheumatic fever is common. There appears to be a high burden of S. pyogenes sore throat in Fiji, apparently underestimated when traditional culture-based methods are used. Although NAAT testing offers higher sensitivity than culture, the costs remain high. There is an urgent need for accurate, affordable diagnostics to guide sore throat management in resource-limited settings. FUNDING: This project was funded by a New Zealand Aid Programme grant, awarded to Cure Kids (NZ research charity). Funds covered all costs pertaining to the study, including research personnel, data collection, patient recruitment and analysis.