Predicting the Impact of a Flu Epidemic
By Eric Toner, M.D., December 1, 2005
CDC’s Updated Modeling Tool: FluSurge 2.0
On November 7, 2005 the CDC released a new version of its FluSurge program, which is an Excel spreadsheet designed to project the impact of a flu epidemic on hospitals. The FluSurge program calculates the impact on hospital capacity over the course of an epidemic, based on user-supplied data on population and medical assets, including the number of hospital beds, ICU beds and ventilators. Calculations are based on assumptions regarding the percentage of flu patients that would require these services.
The first version of FluSurge made calculations based on data from the mild pandemics of 1957 and 1968, and those assumptions could not be altered, which meant the program could not be used to model a more severe pandemic. The new version 2.0 allows users to alter the assumed relative number of patients requiring hospitalization, intensive care, and mechanical ventilation. Length of stay and number of deaths can also be adjusted. With these changes, FluSurge v.2 can now be used to model the impact of various pandemic scenarios on hospitals.
Modeling Provides Vivid Illustration of the Potential Impact of a Severe Epidemic
For planning purposes, the recent HHS Pandemic Influenza Plan posits both a moderate 1957- and 1968-like scenario and a severe 1918-like scenario. The difference between the two is an 11-fold increase in the number of patients requiring hospitalization. Using default variables based on metropolitan Atlanta demographics, FluSurge vividly illustrates the demands that hospitals will face in responding to each type of scenario.
Applying the default assumptions to HHS’s 1968-like scenario for an 8 week epidemic with a 25% attack rate, the model predicts that 28% of all hospital beds and 77% of all ICU beds will be occupied by flu patients by week 5. Flu patients will also use 42% of existing ventilators. While certainly challenging, an effective response to those types of surge demands seems within the realm of possibility using strategies such as cancellation of elective admissions and procedures, conversion of outpatient areas, and doubling up patients in single rooms. The FluSurge program does not allow for adjustments on that level of granularity, but it does allow for adjustments to the number of staffed beds so as to project the impact of staff absenteeism. Significant absenteeism would make any surge response much more difficult.
However, a very different picture emerges for a 1918-like scenario. For such a severe epidemic, the HHS plan assumed 9.9 million hospitalizations nationwide, which equates to 126,776 in metropolitan Atlanta. Based on those assumptions, FluSurge projects that at the peak of the epidemic, flu patients will occupy 2.5x the number of hospital beds now in existence, and critically ill flu patients will occupy 7x the number of ICU beds and 4x the number of ventilators now in existence. These types of demands will never be met given current resources and approaches to planning and preparation.
Planning for a Severe Epidemic Requires a New Approach
Clearly, planning for a severe epidemic requires a very different kind of thinking. If the number of ventilators needed is off by a factor of 4, then expending significant resources to achieve marginal improvements in the critical care surge capacity is a waste of effort and money. Planning resources may be better spent, for example, on figuring out how to provide hydration, antibiotics, oxygen, analgesia and other supportive care on a very large scale. This approach poses difficult challenges indeed, but they are quite different from the challenges inherent in efforts to increase ICU capacity. More importantly, preparing for one type of scenario does not necessarily equal preparing for the other, and may in fact prove detrimental.
Since no one can predict the severity of the next pandemic, planning must address both types of epidemic, which means that hospitals must start considering tiered levels of response. Unfortunately, since funding and manpower for preparedness are limited, hard choices will have to be made when it comes to allocation of scarce resources and the need to plan for both types of flu epidemic scenario.