315. Use of Social Network Analysis to Potentially Identify Patients Most at Risk Early in a Norovirus Outbreak
Session: Poster Abstract Session: Assessing and Reducing Infection Risk
Friday, October 21, 2011
Room: Poster Hall B1
  • IDSA noro2.pdf (1.4 MB)
  • Background: Norovirus is a cause of epidemic diarrhoea and vomiting in hospitals. Recent guidelines have emphasised the importance of organisational preparedness for outbreaks. However, a rationale basis for outbreak control is lacking.

    Methods: All suspected cases of viral gastroenteritis occurring over 3 months in Oxfordshire acute hospitals were assessed. Outbreaks were identified if all 4 modified Kaplan criteria were fulfilled: median illness duration 12-60h; median incubation period 24-48h; >50% with vomiting AND staff and patients affected. Prior anonymised hospital movement data from hospital patients were used to construct a social network. Complete potential disease transmission paths and patient connectivity were calculated. Cases within outbreaks were divided daily into two groups according to their connectivity (median network centrality). Standard models of epidemic growth were fitted to these two groups and their properties compared.

    Results: 10 of 15 suspected outbreaks were identified between 29th December 2009 and 12th February 2010. A total of 188 cases presented with diarrhoea and/or vomiting. Of 79 faecal samples tested, 46 (58%) were norovirus GI/GII positive by PCR. Of these, 43/46 (93%) were sequence typed and all were within the GII.4 lineage. The 10 outbreaks were temporally clustered into 3 waves. Within each of the 3 outbreak waves the median (IQR) connectedness – the number of other hospital patients each case was linked to by prior shared ward exposure - was 99 (64-119), 27 (12-35) and 77 (32-103), respectively. Epidemic curves showed that outbreaks peaked in the most highly connected cases 4 days earlier than the less connected patients (median time to maximum 8 versus 4 days respectively), suggesting these patients could be the source of ongoing transmission.

    Conclusion: Social network analysis of hospital patients can identify highly connected patients who could be targeted for infection control to minimise outbreaks.

    Subject Category: N. Hospital-acquired and surgical infections, infection control, and health outcomes including general public health and health services research

    Tse Hua Nicholas Wong, MB BS , A. Sarah Walker, PhD, Lily O'Connor, Kate Dingle, Tim Peto, MB BS, DPhil, Derrick Crook, MB BCh and John Finney, NIHR Oxford Biomedical Research Centre, Oxford, United Kingdom


    T. H. N. Wong, None

    A. S. Walker, None

    L. O'Connor, None

    K. Dingle, None

    T. Peto, Optimer Pharmaceuticals: Scientific Advisor, Consulting fee

    D. Crook, Optimer Pharmaceuticals: Investigator and Scientific Advisor, Consulting fee and Institution received per-case funding to support trial patient expenses

    J. Finney, None

    Findings in the abstracts are embargoed until 12:01 a.m. EST Thursday, Oct. 20 with the exception of research findings presented at IDSA press conferences.