Understanding the super-spreading events of SARS in Singapore.

INTRODUCTION: It has been noted that SARS transmission is characterised by a few super-spreading events (SSEs) giving rise to a disproportionate number of secondary cases. Clinical and environmental features surrounding the index cases involved were compared with cases in non- SSEs. MATERIALS AND METHODS: Data on 231 cases of probable SARS admitted to Tan Tock Seng Hospital (TTSH) were used. Index cases directly causing 10 or more secondary cases were classified as having been involved in SSEs; all others were defined as non-SSEs. RESULTS: Only 5 cases were involved in SSEs; all 5 were isolated on day 5 of illness or later, and spent at least a brief period in a non-isolation ward; in contrast, amongst the 226 non-SSE cases, only 40.7% and 4.0% were isolated late and admitted to non-isolation wards respectively, and only 3.1% had both these environmental features present; the differences were highly significant (P = 0.012, P <0.001 and P <0.001 by Fisher's Exact test). When compared to 7 non-SSE cases with delayed isolation and an admission to non-isolation wards, SSEs were more likely to have co-morbid disease or require ICU care at time of isolation (P = 0.045 for both factors). CONCLUSION: SSEs were likely due to a conglomeration of environmental factors of delayed isolation and admission to a non-isolation ward, coupled with severe disease stage at time of isolation.