Saturday, April 22, 2017

PSAF Is The New Pandemic Severity Index




















#12,403

In this, our first close look at the revised 2017 CDC/HHS Community Pandemic Mitigation Plan, published yesterday in the MMWR, we look at the new gauge of pandemic intensity; the PSAF (Pandemic Severity Assessment Framework). 

The Pandemic Severity Index (see graphic below) was adopted in the 2007 Community Strategy for Pandemic Influenza Mitigation plan as a way to quantify the likely impact of any pandemic outbreak.  It was based on the initial CFR (Case Fatality Ratio) of the virus, and was modeled in many respects after the 5 category Saffir-Simpson wind scale used for hurricanes. 


While a familiar format to most Americans, it ran into some of the same problems during the 2009 pandemic that the Saffir-Simpson scale has run into with Hurricane Katrina in New Orleans, Superstorm Sandy in New York, and Hurricane Mathew along Florida's east coast.

A single metric (be it CFR or wind speed) doesn't always accurately predict the impact of a pandemic or a hurricane.

In 2009, early reports (where the most seriously ill are most likely to be identified) suggested an elevated case fatality rate. Not unexpectedly, people were taking that number, and multiplying it times 30% of the population, and coming up with horrendous death tolls (see Categorically Speaking).
And just as there can be a huge difference in damage between a Cat 3 hurricane hitting Miami (as Wilma did in 2005), and a Cat 3 hitting New Orleans (as Katrina did the same year), what may turn out to be a CAT 1 pandemic in Ottumwa, Iowa could well end up being a CAT 2+ pandemic in Mumbai, India.
A one size-fits-all rating, based on a single (easily misjudged) metric, can go quickly awry.  Add in the fact that pandemic viruses are constantly evolving, and what might start out as a mild pandemic could strengthen over time, while a severe pandemic might weaken greatly after the opening weeks or months.

What is needed is a more comprehensive and encompassing method of assessing a pandemic virus and predicting its likely impact. To that end, we have the PSAF.  

Pandemic Severity Assessment Framework (PSAF)
 
Assessing Pandemic Severity and Health Impact


When a novel influenza virus emerges that can spread easily and efficiently and cause a pandemic, CDC and partners must gauge its projected impact and recommend rapid action to reduce virus transmission, protect vulnerable population groups, and minimize societal disruption (5). Historically, the severity of influenza pandemics has been estimated by calculating case-fatality ratios.§§ However, as we learned during the 2009 H1N1 pandemic (Box 1), case-fatality ratios may be difficult to measure early in a pandemic because of care-seeking behavior and testing practices (i.e., not everyone will seek care for their illness, and not everyone will be tested and diagnosed with pandemic influenza). As a result, severe and fatal cases may be more likely to be reported, creating a bias.


Due to such limitations, reliance on any single measure of viral transmission or clinical outcomes is unlikely to provide an accurate estimate of the potential impact of an emerging pandemic. CDC has, therefore, developed a new assessment framework that uses multiple clinical and epidemiologic indicators to create a comprehensive picture of the potential impact of an emerging pandemic (3). As indicated in Tables 5 and 6, the Pandemic Severity Assessment Framework (PSAF) estimates pandemic severity (or health impact) by synthesizing multiple measurements of:

  •  Viral transmissibility, including school, workplace, and/or community attack rates, secondary household attack rates, school and/or workplace absenteeism rates, and rates of emergency department and outpatient visits for ILI.
  •  Clinical severity, including case-fatality ratios, case-hospitalization ratios, and deaths-hospitalizations ratios.
Additional PSAF data may be obtained by characterizing genetic markers in a pandemic virus and by conducting animal studies on its transmissibility and virulence.





No matter how well designed the algorithm, getting good numbers out of a formula requires plugging `good' numbers in. And getting those numbers - particularly through the `fog of flu' common in the early days of a pandemic - may not be possible.

But once reasonably accurate data becomes available, this method ought to provide us with a much better idea of what we are facing and must prepare for.

No comments: