Lessons learned from identifying clusters of severe acute respiratory infections with influenza sentinel surveillance, Bangladesh, 2009-2020

Md Ariful Islam,Md Zakiul Hassan,Mohammad Abdul Aleem,Zubair Akhtar,Sukanta Chowdhury,Mustafizur Rahman,Mohammed Ziaur Rahman,Md Kaousar Ahmmed,Syeda Mah-E-Muneer,A S M Alamgir,Shah Niaz Rubaid Anwar,Ahmed Nawsher Alam,Tahmina Shirin,Mahmudur Rahman,William W Davis,Joshua A Mott,Eduardo Azziz-Baumgartner,Fahmida Chowdhury
DOI: https://doi.org/10.1111/irv.13201
2023-09-22
Abstract:Background: We explored whether hospital-based surveillance is useful in detecting severe acute respiratory infection (SARI) clusters and how often these events result in outbreak investigation and community mitigation. Methods: During May 2009-December 2020, physicians at 14 sentinel hospitals prospectively identified SARI clusters (i.e., ≥2 SARI cases who developed symptoms ≤10 days of each other and lived <30 min walk or <3 km from each other). Oropharyngeal and nasopharyngeal swabs were tested for influenza and other respiratory viruses by real-time reverse transcriptase-polymerase chain reaction (rRT-PCR). We describe the demographic of persons within clusters, laboratory results, and outbreak investigations. Results: Field staff identified 464 clusters comprising 1427 SARI cases (range 0-13 clusters per month). Sixty percent of clusters had three, 23% had two, and 17% had ≥4 cases. Their median age was 2 years (inter-quartile range [IQR] 0.4-25) and 63% were male. Laboratory results were available for the 464 clusters with a median of 9 days (IQR = 6-13 days) after cluster identification. Less than one in five clusters had cases that tested positive for the same virus: respiratory syncytial virus (RSV) in 58 (13%), influenza viruses in 24 (5%), human metapneumovirus (HMPV) in five (1%), human parainfluenza virus (HPIV) in three (0.6%), adenovirus in two (0.4%). While 102/464 (22%) had poultry exposure, none tested positive for influenza A (H5N1) or A (H7N9). None of the 464 clusters led to field deployments for outbreak response. Conclusions: For 11 years, none of the hundreds of identified clusters led to an emergency response. The value of this event-based surveillance might be improved by seeking larger clusters, with stronger epidemiologic ties or decedents.
What problem does this paper attempt to address?