RESUMEN
INTRODUCTION: Cluster-associated transmission has contributed to the majority of COVID-19 cases in Malaysia. Although widely used, the performance of the World Health Organization (WHO) case definition for suspected COVID19 in environments with high numbers of such cases has not been reported. MATERIALS AND METHODS: All suspected cases of COVID-19 that self-presented to hospitals or were cluster screened from 1st April to 31st May 2020 were included. Positive SARS-CoV-2 rRT-PCR was used as the diagnostic reference for COVID-19. RESULTS: 540 individuals with suspected COVID-19 were recruited. Two-third of patients were identified through contact screening, while the rest presented sporadically. Overall COVID-19 positivity rate was 59.4% (321/540) which was higher in the cluster screened group (85.6% vs. 11.6%, p<0.001). Overall, cluster-screened COVID-19 cases were significantly younger, had fewer comorbidities and were less likely to be symptomatic than those present sporadically. Mortality was significantly lower in the cluster-screened COVID-19 cases (0.3% vs. 4.5%, p<0.05). A third of all chest radiographs in confirmed COVID-19 cases were abnormal, with consolidation, ground-glass opacities or both predominating in the peripheral lower zones. The WHO suspected case definition for COVID-19 accurately classified 35.4% of all COVID-19 patients, a rate not improved by the addition of baseline radiographic data. Misclassification rate was higher among the cluster-associated cases (80.6%) compared to sporadic cases (35.3%). CONCLUSION: COVID-19 cases in Malaysia identified by active tracing of community cluster outbreaks had lower mortality rate. The WHO suspected COVID-19 performed poorly in this setting even when chest radiographic information was available, a finding that has implications for future spikes of the disease in countries with similar transmission characteristics.