Influence of Social Media on an Acute Respiratory Outbreak at a High Profile Party Venue— Los Angeles, February 2011

BACKGROUND: On February 11, 2011, public health officials received an inquiry about an outbreak of legionellosis reported through social media by attendees of a conference held February 1–3. Several attendees indicated a legionellosis diagnosis. We investigated to characterize illnesses, identify etiology, and determine associated exposures.

METHODS: We reviewed blogs, Twitter^®, and Facebook^®, and emailed a SurveyMonkey^® survey to all conference attendees from 30 countries and Venue A hostesses and staff to assess symptoms, illness onset, and exposures. A case was defined as fever and ≥1 symptom (headache, cough, dyspnea, myalgia, vomiting or diarrhea) with onset after February 1 and within 10 days of last exposure to Venue A. When available, respiratory, blood, and urine specimens from 45 (30%) ill attendees were tested for Legionella and other respiratory pathogens using polymerase chain reaction (PCR), culture, urinary antigen, and serology. Environmental samples from conference venues were tested for Legionella using standard methodology (Gaia et al, 2005).

RESULTS: Seventy-nine persons self-reported non-specific illness through social media. Of 715 conference attendees, 439 responded, 123 met the case definition; 69 (56%) reported illness onset on February 5, suggesting a point-source outbreak. The relative risk for illness associated with attending a Venue A party on February 3 was 3.8 (95% confidence interval: 2.0–7.5). Out of an estimated 180 hostesses, 47 were surveyed, of whom 22 (46%) met the case definition. Of 67 staff, 41 were interviewed, of whom 3 (7%) met the case definition. Four attendees tested PCR-positive for influenza A; 3 were identified as influenza A (H1N1) pdm09 (pH1N1). Among 16 cases who submitted paired sera, 12 (75%) cases were positive for pH1N1 antibody (including 2 PCR-positive cases) and 3 (19%) showed a fourfold rise in pH1N1 antibody titer indicative of recent infection. Multiple Legionella pneumophila isolates grew from Venue A whirlpool spa (4 serogroups 1 and 3) and the pool filter (serogroup 6); however, no cases of legionellosis were confirmed among attendees.

CONCLUSIONS: Venue A was associated with increased risk for self-reported illness. Legionella exposure is supported by environmental but not biological testing. Laboratory and epidemiological evidence supports pH1N1 as the most likely outbreak etiology. Social media helped identify this outbreak among geographically dispersed persons and facilitated data collection but may have influenced recall and participation among attendees and the public health investigation.