On 22 October 2007, a case of suspected measles in an unvaccinated two-year-old was notified to Thames Valley Health Protection Unit (TVHPU) by North East and North Central London Health Protection Unit (NE&NCL HPU). The child, having had a rash for four days and infectious, had attended a social event in the Thames Valley area on 21 October with over 400 guests, mostly Bangladeshi. The child was apparently running all around the hall during the party, thus potentially putting all guests at risk. In view of the large number of contacts, diagnostic confirmation was urgently sought. The source of infection in the index case was unknown and there was no suggestion or evidence that this was linked to any other outbreak.

Confirmation of measles by PCR was reported to TVHPU on 24 October. This prompted a contact tracing exercise to identify ‘at-risk’ persons among the wedding guests in order to minimise secondary spread. Current guidelines in the United Kingdom state that susceptible contacts should be vaccinated with Measles, Mumps, Rubella (MMR) vaccine within three days of exposure and Human Normal Immunoglobulin (HNIG) should be offered to unvaccinated pregnant women and the immuno-compromised up to six days following exposure [1,2]. Due to delays in contacting the organiser of the event, the list of guests with their contact numbers was obtained one day after receiving the laboratory confirmation, i.e. three days after measles exposure of the guests. This delay meant that the window of opportunity for administering MMR vaccination to susceptible contacts within the recommended three days was lost.

Results of the contact tracing
A list of 28 families resident in Thames Valley and their telephone numbers was provided by the event organiser on 25 October. Information on MMR vaccination, past measles infection, immune status and pregnancy (if applicable) was obtained from 21 out of 28 families by the TVHPU team (five staff members) within two hours of receipt of the list. For the remaining seven contacts, an appropriate message was left on answer machines to make contact whenever possible. The presence of a Bengali speaker in the team was fortuitous, as two contacts had very limited ability to communication in English.

Contact tracing revealed that 44 adults, including three pregnant women and 50 children resident in Thames Valley attended the event (n = 94). None of the children and adults were immuno-compromised. Many adults were unaware of their MMR vaccination status or past measles infection. They were therefore advised to check their General Practitioner’s record. Among the three pregnant women, two had no history of MMR vaccination or past measles infection. HNIG was arranged for both women on the fifth day following the exposure. Of the 50 children, four (age range: 18 months to 9 years) had not had MMR vaccination, as well as two eight-month-old children who were not due for MMR until five months later. MMR vaccination was strongly recommended for these six children for long term protection, although that this would not prevent measles if they were already incubating the virus.

On day five after the event (Friday), active contact tracing was considered no longer appropriate, as HNIG could not be received from the single central stock in London the same day and General Practices, for administering HNIG, were closed on Saturdays (see Figure 1). A letter providing information on the symptoms of measles along with the contact details of the relevant Health Protection Unit was cascaded by the NE&NCL HPU to all guests via the parents of the index case. No secondary cases have been reported to date and the incubation period for secondary cases has passed.

Discussion
Several lessons were identified during and following this exercise. Firstly, the need for better coordination: there was an avoidable delay of one day between diagnostic confirmation and obtaining the guest list from the party organisers, despite the NE&NCL HPU alerting TVHPU of this incident on Day 1 after the event. Action could and should have been initiated to obtain the guest list and their contact details while awaiting the test results. Had this delay been minimised, susceptible contacts could have been vaccinated within the first three days post exposure. However, once TVHPU received the list, prompt action resulted in the two pregnant women identified as ‘at risk’ being adequately protected with HNIG and six children being advised to receive MMR vaccine.

Secondly, the lack of immediate access to HNIG led to the slightly premature end of active contact tracing on day five after the event. The feasibility of having local/regional stocks of Immunoglobulin for quick access and the logistical issues in administering HNIG over the weekends need to be explored and clarified in order to reduce the likelihood of missed opportunities in future incidents.

The third lesson concerned access to interpreter/translator services: in the United Kingdom (UK), interpreter services for HPUs have traditionally been provided via the local Primary Care Trusts but without any formal agreement or policy for such use. The investigation highlighted the need for HPUs and public health systems elsewhere to have formally agreed policies and procedures for accessing interpreter services. This need is increasing in a globalised world with multi-ethnic and multi-lingual communities.

This incident provided an opportunity to promote the benefits of vaccination to parents and raise awareness of recent increased measles incidence. A number of European countries have recently reported an increase in the incidence of measles and the need for continued efforts in measles control and vaccination awareness [3,4].

Contact tracing following exposure in big social events is justified because of the high transmissibility of measles when introduced into susceptible populations [5]. Indeed, some studies have found vaccination rates in minority groups to be lower than that of the general population, suggesting higher susceptibility to infection following exposure [6,7]. However minority ethnic communities are not a homogeneous group and national level data on MMR vaccination rates by ethnic group are not available in the UK. Immunisation coverage of more than 92% is considered necessary to prevent measles outbreaks in a given population through ‘herd immunity’ [8]. In the TVHPU cohort of contacts where telephone contact was successful, 44 of the 48 eligible children (92%) were already vaccinated against measles. The vaccination status of children in families that were not contactable is unknown, thus this proportion could be lower or indeed higher. Had the high vaccination rates in the Bangladeshi community been known prior to contact tracing, a different course of action may have ensued as large numbers of secondary cases were unlikely, supported by the lack of subsequent notifications to the TVHPU. Contact tracing raises a number of practical issues, particularly those of coordination and staffing resources needed within a tight timescale. The availability of vaccination coverage data by ethnic group enable better targeting of response to disease exposure as well as highlighting and responding to inequalities in vaccine coverage among various ethnic groups.

Acknowledgements
We wish to thank colleagues in North East & North Central London HPU for their cooperation in the investigation.