As we approach the 90th anniversary of the devastating 1918 influenza pandemic, the ever-present influenza virus’s genetic jigsaw puzzle could be coming together again in a virulent way. As with other catastrophes, business planning for a pandemic is essential to both the preservation of our economy and the well-being of our society.
According to federal government estimates, pandemic influenza has the potential to cause 30% of the workforce to miss as much as three weeks of work, generating economic losses similar to a recession.Pandemic influenza could infect 90 million Americans; a severe strain could kill 2 million.1 Pandemic influenza would dramatically increase medical costs, but the greatest effect on most businesses will be lost productivity. For this reason, the authors created a human capital model that estimates lost work time as a driver of lost revenue and additional morbidity and mortality costs. Among other findings, our model verifies the government’s recession analogy. It also demonstrates how companies that successfully prepare for pandemic influenza can protect their revenue-generating ability and remain financially sound even when others in the same industries do not.
This article provides background on pandemic influenza and presents quantitative and qualitative considerations that businesses should keep in mind when planning for such a catastrophe. As others have noted, businesses can provide a crucial layer of defense and response for society as a whole.
Today’s pandemic—thus far, among birds
A pandemic influenza could occur if the highly pathogenic strain of H5N1 avian influenza A virus, currently endemic in wild birds, mutates to become highly contagious among people.2 Such a mutation would quickly spread worldwide given today’s air travel. The H5N1 influenza strain, first identified in humans in 1997, has thus far lacked the ability to spread easily from person to person. The 1997 episode in Hong Kong resulted in 18 hospitalizations and six deaths. As of March 1, 2007, there have been 277 human H5N1 cases reported and 167 deaths.3,4
click to enlarge
To date, nearly all human cases were caused by close contact with infected birds, although there is at least one case of probable human-to-human spread. Meanwhile, the H5N1 virus has spread in birds beyond Southeast Asia and China into Central Asia, Africa, and Europe; in recent years the total number of countries reporting avian infections has doubled every year (see Figure 1).
How precarious is the current influenza risk? The World Health Organization (WHO) believes we have reached a dangerous threshold. A new influenza subtype has caused the disease to appear in humans, but the strain has not yet evolved to where it can spread among humans. The as-yet infrequent spread of H5N1 to people triggered Phase 3 in WHO's six-phased system and marks the entry to a "pandemic alert period." Phases 4 and higher involve human-to-human transmissions. The current Phase 3 is the closest we have come to pandemic since the 1968 flu outbreak.6
click to enlarge
Any human-to-human transmissions are likely to affect people of all ages. In February 2007, WHO published results from its study of avian influenza cases for the first three years of recorded human infections.7 From Nov. 25, 2003, through Nov. 24, 2006, 10 countries reported 256 cases. The number of newly reported cases increased steadily over the 36-month period—from 45 cases during the first 12 months, to 93 cases during the next 12 months, to 188 cases during the last 12 months. The demographic characteristics of these cases are interesting.
- The influenza cases have been most prevalent among the young. The median age of confirmed cases was 18 years, more than half were people younger than 20 years, and almost 90% were people younger than 40 years. After adjusting for country demographics, WHO found incidence rates to be higher among younger persons.
- The median number of days from the reported onset of symptoms to hospitalization was four days. The median number of days from the reported onset of symptoms to death was nine days.
- The overall case fatality rate was 60%; it was highest among persons 10–19 years of age (76%) and lowest among persons age 50 and older (40%).
These statistics differ from seasonal influenza—ordinarily, older populations are more affected and the fatality rate is not nearly as high.
Our health system is not well-equipped to stop pandemic influenza, as demonstrated by the 2003 outbreak of SARS in Canada. SARS (severe acute respiratory syndrome), which was first identified in 2002, is a severe lower respiratory illness caused by a novel coronavirus. During an outbreak in late 2002 through early 2003, more than 8,000 SARS cases and nearly 800 deaths were reported.8 A mother returning home from Hong Kong brought SARS to Toronto, passing it to her son. The sick boy waited for more than 16 hours in a crowded emergency room, transmitting SARS to two other patients. Healthcare workers who failed to use proper procedures were also victims.9 Fortunately, SARS is not as contagious as pandemic influenza.
Recently, the Congressional Budget Office (CBO) and the U.S. Department of Health and Human Services (HHS) published seminal reports on the effects of a pandemic influenza outbreak.10 While the focus of the CBO report was on the macroeconomic effects of pandemic influenza, the focus of the HHS report was on pre-pandemic planning.11 Because only three widespread influenza outbreaks occurred during the 20th century, modelers have scant information with which to develop infection and fatality rates. Both reports look to the 1918 outbreak for “severe” infection and fatality rates, and to the 1957 and 1968 outbreaks for a "moderate" outbreak.
The CBO assumed that under a severe outbreak, 30% of all individuals would be infected, and that among those infected, 2.5% of the cases would be fatal. Survivors would miss three weeks of work because of illness, fear of becoming sick, or because they were caring for family or friends. The CBO assumed that, under a moderate outbreak, 25% of all individuals would be infected and the case fatality rate would be between 0.1% and 0.2%. Survivors of a moderate outbreak would miss 25% of the amount of time lost under a severe outbreak.
The CBO estimates that real GDP would drop by 4.25% over the year following a severe pandemic influenza, with approximately half of the drop because of employees’ inability to work and half to decreases in consumer demand. Under the moderate pandemic scenario, economic effects might be masked by normal variations in economic activity.
General business advice
Planning for pandemic influenza should be an important part of an organization’s risk management and disaster recovery plans. The Business Roundtable, an association of American CEOs, has made explicit the connection between pandemic influenza and homeland security: “Should a pandemic influenza outbreak occur, chief executive officers would play a key role in protecting their employees’ health and safety as well as mitigating the impact on business. As with all homeland security preparations, the key task is planning.12 Yet according to a survey conducted between mid-April and mid-May 2006, The Conference Board found that 44% of employers with sales between $500 million and $1 billion did not have any plans in place to address the effect of a pandemic influenza.13
click to enlarge
Even modest steps could minimize adverse impacts on employers. Business continuity plans that map essential links in the employer's operations could help control the disruption caused by worker absence. A reserve of prophylactic treatments could slow disease spread across the employer's workforce. The U.S. Department of Homeland Security provides extensive preparedness plans for businesses.14
Economic effects on particular businesses
The authors' model considers the following factors, which businesses can consider in their own planning:
Worker absence as a driver of reduced revenue: Most U.S. companies have revenue that is about twice as high as wages. In other words, every dollar of lost wages produces two dollars in lost revenue. Manufacturing, with high raw material costs, has a higher multiplier. Profit as a percentage of revenue varies from company to company.
Workforce composition affects likely absences—and the damage they cause: In pandemic influenza, workers with dependents are more likely to miss work because they may stay home to care for ill family members. Different worker types (executives, managers, sales, production, etc.) have different family compositions and wage levels.
Scenarios: The CBO/HHS severe and moderate scenarios define infection rate, morbidity, and mortality, and are based on historical data.
Benefit designs: Of course medical benefit spending goes up, which varies by scenario. Depending on the scenario, life insurance benefits and replacement costs could become important.
Use of antiviral prophylaxis: Antiviral prophylactics add to cost, but if they dramatically reduce infection rates as many expect, the pandemic influenza damage will be much lower.
Applying company specifics to the well-established scenarios gives enterprise risk managers a foundation on which to develop plans.
Treatment, prevention, and preparedness
Pandemic influenza does not yet exist and its clinical characteristics are unknown, but experts expect treatment will be the same as for seasonal influenza, ranging from symptomatic treatment to antivirals to outpatient medical care, hospitalization, ICU care, and mechanical ventilation. If a pandemic were to occur, antiviral drugs could limit the spread of the infection through aggressive prophylaxis and antivirals could reduce the severity of cases. Current experience with seasonal influenza shows that, to be most effective, antiviral drugs used as prophylaxis need to be administered within 48 hours of exposure to the infection. In May 2006, the HHS set a goal to stockpile enough antiviral treatments for at least 25% of the U.S. population, or 75 million individuals. In a January 2007 report to Congress, the HHS reported that by fiscal year 2008, the U.S. federal government will have completed the purchase of 50 million antiviral treatment courses and subsidized the states and other entities with the purchase of another 31 million treatment courses (see Figure 2).16
click to enlarge
In a pandemic influenza outbreak, drug companies could convert their seasonal influenza vaccine programs to produce a vaccine for the pandemic influenza strain. The current seasonal influenza vaccine process requires about six to nine months from start to finish and that process can’t start until the pandemic influenza strain has been identified—which means the pandemic has begun. Emerging technologies may dramatically shorten that process. However, vaccines take about two weeks to become effective after inoculation.17 Even if vaccination technology leaps forward, a basic strategy for limiting the impact of pandemic influenza must involve delaying the spread of the disease. Delaying infections means reducing the concentration of economic and human impact—and buying time until a vaccine that prevents the infection can be developed and disseminated. Based on historical data, epidemiologists expect two or more waves of the virus; not all susceptible people will catch the virus at the same time, so delaying tactics would seem to be effective.
The Department of Homeland Security identifies key tactics that could be adopted as part of a pandemic preparation plan, and each measure is applicable to most businesses:
Isolation: Separation of persons with specific infectious illnesses in their homes, in hospitals, or in designated healthcare facilities.
Quarantine: Separation and restriction of the movement of people who, while not yet ill, have potentially been exposed to an infectious agent.
Social distancing: Within the workplace, social distancing measures could take the form of modifying the frequency and type of face-to-face employee encounters (e.g., placing moratoriums on hand-shaking, substituting teleconferences for face-to-face meetings, staggering breaks, posting infection control guidelines), establishing flexible work hours or worksites (e.g., telecommuting), promoting social distancing between employees and customers to maintain three feet of spatial separation between individuals, and implementing strategies that request and enable employees with influenza to stay home at the first sign of symptoms.
Closing places of assembly: Voluntary or mandatory closure of public places, including churches, schools, and theaters.
"Snow days" and/or furloughing nonessential workers: Voluntary or mandatory closure of all nonessential businesses and/or furloughing all nonessential workers. Of course the employer should consider extending paid time off.
Changes in movement patterns: Restricting movement at the U.S. border, instituting reductions in the transportation sector, and applying quarantine protocols. Business and discretionary travel would be severely curtailed, with a rippling effect on the petroleum, airline, railroad, trucking, postal, and delivery industries, to name a few.
click to enlarge
Two antiviral drugs, zanamivir and oseltamivir, have indications as prophylaxis against seasonal influenza; they can greatly reduce the likelihood that an unvaccinated individual will become infected. While the characteristics of pandemic influenza are unknown, some organizations are stockpiling these drugs to protect their workforces from pandemic influenza.
At the outset, we posed the question: "How realistic a threat is an influenza pandemic?" Admittedly, it is impossible to predict the "if, when, what, where, and how." However, recent history has taught us that low-probability, high-cost events certainly do occur. Whether the disasters are attributed to nature (e.g., earthquakes, hurricanes) or man (e.g., terrorist attacks), it is prudent to be aware of the possible consequences and take appropriate preparatory action. Understanding the nature of the risk—and preparing for it—might determine whether we avoid the devastation of an event like the 1918 outbreak or relive it.
PHILIP S. BORBA is a principal and senior consultant in the economics consulting practice of Milliman. He has worked extensively on the development and analysis of large claims databases for property-casualty insurance industry applications. Philip has also completed several data assembly and analysis projects for workers’ compensation, private passenger and commercial automobile, homeowners, general liability, and crop insurance. He has completed claim-cost projects for research organizations, employer associations, and property-casualty insurers.
KATHRYN FITCH is a principal and healthcare management consultant in the New York office of Milliman. Her expertise is in the intersection of disease processes, financing health benefits, and managing care. Recent projects have included evaluating disease management outcomes for commercial, Medicare, and Medicaid populations; a variety of population-based cardiovascular risk studies; program assessments for care management processes; and inpatient process improvement at several hospitals focused on improved denial management and length-of-stay reduction. Kate’s clinical background includes extensive experience as a registered nurse in emergency, adult inpatient, and ambulatory care units.
BRUCE PYENSON is a principal and healthcare consulting actuary in the New York office of Milliman. His practice focuses on integrating financial, clinical, and health benefits knowledge to serve a wide variety of business and public policy clients. Recent client projects have included evaluating the cost/benefit to employers of cancer screening and smoking cessation programs, research on the cost and survival of Medicare hospice patients, evaluating Medicare demonstration projects, and benefit choice modeling for a large employer.