Pandemic H1N1 Influenza
The Center for Communicable Disease Dynamics – and its predecessor, a MIDAS research group at HSPH -- took an active part in analysis and response to the 2009 H1N1 influenza pandemic, working closely with CDC, local health departments, and academic collaborators around the world. Dr. Lipsitch, served on the PCAST Working Group on 2009 H1N1 Influenza and on CDC’s Team B, and members of the CCDD both at Harvard and in Hong Kong provided advice to local, state and national government agencies throughout the pandemic.
Key findings to date from projects led by or including CCDD investigators (at HSPH, and CCDD collaborators in Hong Kong, the Netherlands and elsewhere) include the following:
Severity. The case-fatality ratio for H1N1 was estimated as lying between approximately 1/2000 and 1/14,000 symptomatic cases, and corresponding estimates of the case-hospitalization and case-ICU ratios were provided. We also assisted CDC in developing multiplier models to estimate symptomatic cases and deaths from hospitalizations. Travel data were used to estimate that case numbers in Mexico were 2 or more orders of magnitude larger than confirmed case counts, suggesting that severity there was substantially overestimated.
Transmissibility. Using the US line list the basic reproduction number for H1N1 in the early part of the US epidemic was on the order of 1.7 to 1.8. This was slightly higher than most other estimates, reflecting perhaps the use of a different data set, different adjustments for changing ascertainment (which was a focus of our analysis), local variability in transmissibility, or other factors. Work on the same question is currently in review for Hong Kong. We have also been working on several analyses – using seasonal influenza data -- of viral shedding and potential impacts of interventions as a function of timing with respect to viral shedding (12).
Antivirals. Just prior to the pandemic, a strategy was proposed to minimize the risk of antiviral resistance spreading using sequential or simultaneous multiple drug treatment at the population level, and we collaborated with Australian colleagues to translate this into policy recommendations. A cost-benefit model was constructed and analyzed to show that predispensing of antivirals to individuals at high risk of complications would under very broad assumptions likely reduce total mortality in the pandemic.
Analysis of other interventions. In the period when severity was unclear and unusual measures were under consideration to reduce morbidity and mortality, a mathematical model for the use of passive immunotherapy was constructed and analyzed to assess the conditions under which this could be a successful population-wide strategy.
School closure. School closure in Hong Kong was shown to have had a large effect in reducing H1N1 transmission.
Intervention targeting. In another piece of work undertaken prior to the pandemic and finished during it, it was shown that maximal reductions in transmission can be achieved by targeting interventions to groups with the highest incidence rate or highest product of incidence rate and force of infection, depending on the type of intervention.
Seasonality. In another study not initially targeted to pandemic work, it was shown that absolute humidity, previously implicated by laboratory studies as a cause of flu seasonality, could explain the epidemiologic patterns of seasonality in flu for the last several decades in the United States. Using this work and some real-time estimates of transmission in the United States following the October pandemic peak, we estimated in response to a BARDA query that the winter uptick in transmissibility would be unlikely to result in a “second wave” of transmission.
Policy analysis. In the first weeks of the pandemic, we defined the likely mismatch in timing between clear evidence of severity/transmissibility and the need to make decisions concerning pandemic response. We described a system of surveillance that could implement the WHO medically attended infection. CCDD members participated in a WHO working group report on the key data needs for surveillance in summer recommendations to cease counting cases, suggesting a combination of syndromic surveillance and targeted, random viral testing to define the incidence of symptomatic, the pandemic (15). Dr. Lipsitch co-authored the PCAST H1N1 working group report to the President, issued in August 2009.
CCDD investigators : Marc Lipsitch, Ed Goldstein, Marcello Pagano, Laura White, Justin O’Hagan, Joel Miller Ted Cohen, Joseph Wu, Jeffrey Shaman, Jacco Wallinga, Steve Riley, Ben Cowling, Martin Lajous
Many collaborators at CDC, New York City Department of Heal Ben Cowling th and Mental Hygiene, Milwaukee Health Department, Hong Kong University, MRC Biostatistics Unit, and elsewhere.