Appendix

    While complex epidemiological models help us learn about how epidemics develop, they are unlikely to accurately predict the future. The same obviously applies to the very simple model used to generate this book’s casualty estimates and demand for health resources during the course of a major influenza epidemic in Canada. These estimates are the foundation of the scenario presented in this book and so deserve explanation in this appendix. Of course, any mistakes are the author’s.
    National pandemic planning emphasizes a response to predictions based on mild to moderate pandemic scenarios consistent with the last two pandemics. The FluSurge 2.0 model——developed by the U.S. Centers for Disease Control to predict, week by week, the surge in flu cases during a pandemic——uses defaults for the attack, hospitalization and mortality rate based on the 1968 Hong Kong Flu pandemic.
    Since this book presents a severe pandemic scenario such as the Spanish Flu pandemic, which is entirely possible in our future, casualty estimates and demand for health resources must be adjusted based on reasonable assumptions.
    The first major variables to be explored are:

1) population
2) infection rate, or how many people will become infected
3) attack rate, or how many people will become clinically ill because of infection
4) what portion of the clinically ill population will seek outpatient care
5) what portion of the clinically ill population will become hospitalized
6) what portion of the clinically ill population will die

    In this book, it is assumed that a pandemic occurs in 2-3 waves and that the first wave has been documented as having ended in Canada. The Canadian Pandemic Influenza Plan for the Health Sector, published by Health Canada in 2006, assumes up to 70 percent of the public could become infected during the entire pandemic, with 15-35 percent of the population becoming clinically ill and the rest experiencing asymptomatic infection. Most clinically ill cases, or 25 percent of the population——if assuming 35 percent become infected over the entire pandemic, will occur during the first wave.
    We have now made assumptions regarding 1) population, 2) infection rate and 3) attack rate. Given these assumptions, in Canada, with its population of 31,612,897 (2006 Census), about 15.8 million would become infected. Of these, about 7.9 million would have become infected but show no symptoms and about 7.9 million would become clinically ill.
    If the pandemic has a mild or moderate severity, the Plan assumes up to 50 percent of clinically ill cases will seek outpatient care, 1 percent will be hospitalized and recover, and 0.4 percent will die. In Canada, this would translate to about 3.95 million people seeking outpatient care, about 79,000 hospitalized and recovering, and about 31,600 dying during the first wave of the pandemic. For a severe pandemic, and in the absence of interventions, however, the Plan states: “Recently, projections have been made based on a more ‘severe’ scenario. In the severe scenario it is estimated that 2 percent of clinical cases will die and 10 percent will require hospitalization for management of their illness.” This book sets the hospitalization rate at 10 percent. The difference between 1 percent and 10 percent in terms of hospitalizations would likely be catastrophic to the Canadian health system.
    We have now made assumptions regarding 4) what portion of the clinically ill population will seek outpatient care and 5) what portion of the clinically ill population will become hospitalized. Given previous assumptions, we now have about 7.9 million clinically ill cases in Canada. Of these, about 3.95 million would seek outpatient care and about 790,000 would be hospitalized due to complications such as pneumonia.
    The author also adjusted the 2 percent mortality rate, cited in the Plan as a possibility in a severe pandemic scenario, to 2.5 percent, as the virus in this book is modeled on the Spanish Flu of 1918-19, which presented an estimated case fatality rate of >2.5 percent. Now we have an assumption for 6) what portion of the clinically ill population will die. In Canada, this would translate to nearly 200,000 fatalities, less than 1 percent of the total population.
    The next key question is: Will the health system be able to respond effectively to this demand given available resources?
    The health system can be crudely described as resources consisting of skilled people and equipment. If we know the number of flu cases requiring different types of resources, we can then compare these numbers to the available amounts of the resources. As a simple example, if 10 people get sick and need aspirin, and the health system only has five aspirin and can’t get any more, then demand for aspirin is 200 percent of the health system’s aspirin capacity. In the case of the Avian Flu, critical resources include 1) acute care beds, 2) Intensive Care Unit beds and 3) beds equipped with mechanical ventilators.
    First, we must understand that obviously the epidemic will not occur all at once, but over time. The total number of cases that will occur must be distributed over the anticipated duration of the epidemic. In this book, this is accomplished using the FluSurge 2.0 model, which allows the user to set the epidemic duration for six, eight or 12 weeks. This book arbitrarily assumes the epidemic in any given region will last eight weeks. If it were to last six weeks, the health system would become much worse off because the total number of cases would have to be treated in a shorter amount of time, increasing pressure. If it lasted 12 weeks, the situation would become much better.
    Some pandemic plans assume that clinically ill people requiring hospitalization during a pandemic will require a bed (a “bed” being the equipment and personnel needed to care for an individual requiring critical care who would be occupying it) for an average of five days. They further assume that 15 percent of people needing hospitalization will require care in an Intensity Care Unit for an average of 10 days, and 7.5 percent of them will require a ventilator bed for an average of 10 days. Finally, they assume that 70 percent of all flu-related deaths will occur in hospitals.
    This book accepts these assumptions except for length of hospital stays; because of the severity of the Avian Flu virus in this book, clinically ill admitted to hospitals will require an acute care bed for an average of seven days, an ICU bed for an average of 14 days, and a ventilator bed for an average of 14 days.
    By distributing these cases over eight weeks and comparing them to available resources——acute care beds, ICU beds and ventilator beds——we can see whether demand exceeds available resources, and by how much. For population numbers, we can use 2006 Census figures. For resource numbers, we can use figures published by provincial health systems citing 2004-2005 inventories. The author located three pandemic response plans in which figures for total acute care, ICU and ventilator beds are listed: the Calgary Health Region and the provinces of Ontario and Nova Scotia.
    Consider a simple model for Ontario as an example, as shown in Table A-1. According to the Ontario Health Plan for an Influenza Pandemic (fourth edition, July 2007), in 2004, the province maintained 17,116 acute care beds, 1,510 ICU beds and 1,096 ventilator-supported beds. Based on the province’s population in the 2006 Census, a 25 percent attack rate, and 10 percent hospitalization rate, during the Pandemic nearly 320,000 clinically ill will seek admission to hospitals in the province. These numbers are distributed over the course of an eight-week epidemic, with cases gradually building until they peak and then start to decline.
    In week 1 in Ontario, 77.5 percent of hospital admissions, or about 15,000 people, will require an acute care bed, consuming 86 percent of hospital acute care bed capacity. Fifteen percent of hospital admissions, or nearly 2,900 people, will require an ICU bed, demanding 189 percent of hospital capacity. And 7.5 percent, or about 1,400 people, will require a ventilator bed, demanding 130 percent of hospital capacity. This means, roughly, that theoretically everybody needing a hospital bed can access one, but only about one out of two people who need an ICU bed can get one and about three out of four people who need a ventilator bed can get one.

Table A-1. The epidemic in Ontario without interventions.

    This situation deteriorates over time due to two factors. First, the number of cases will grow until they peak in week 4 and begin to decline. Second, ICU beds and ventilator beds will be required for two weeks, tying up these resources.
    By the peak of Ontario’s epidemic in week 4, about 46,500 people will require acute care beds, demanding 273 percent of hospital capacity, about 16,000 people will require ICU beds, demanding 1071 percent of hospital capacity, and about 8,000 people will require ventilator beds, demanding 738 percent of hospital capacity.
    The final key question is: Could government intervene in some way to reduce hospital demand and/or mortality? The simple answer is yes——but by how much?
    There are three ways government can intervene: First, it can increase available hospital capacity from emergency equipment stockpiles and conscript workers to operate these resources. Second, it can promote social distancing by closing down schools and non-essential businesses, public education and banning public gatherings. Third, it can stage pharmaceutical interventions in the form of antiviral drugs and a vaccine.
    Hospital resources are likely to increase during a pandemic, as shown in the updated Ontario model in Table A-2. The national emergency stockpile may increase local capacities by 25-50 percent during a pandemic. Assuming a perfectly equitable distribution that increases all hospital capacity in Canada (including ICU and ventilator capacity, not just acute care beds) by 25 percent, a factor of 1.25 is applied to all of these resources around week 3 of the epidemic. Another increase in assets is likely to occur due to the establishment of alternative care sites, which is called for in various pandemic response plans. However, it is unknown what impact these additional sites will have on health system capacity. In this book, the author arbitrarily assumes an additional increase by a factor of 0.25 for acute care beds, but no increase in number of ventilator and ICU beds, applied around week 4 of the epidemic.
    (Another boost in assets may occur from health systems seeking to purchase any ventilators that their manufacturers might have in inventory; this effect could not be estimated and is not included in this book. A final boost in assets may be based on the fact that the resource numbers for Ontario are from 2004 and the numbers for Nova Scotia are from 2005, and related to 2006 population numbers; as population increased through 2006, these resources may also have increased.)

Table A-2. The epidemic in Ontario with an increase in hospital capacity due to utilization of the national emergency stockpile and deployment of alternative healthcare sites.

    As shown in the updated model in Table A-2, these measures effectively increase hospital capacity but it is still not enough to meet demand. Ontario would not have enough acute care beds for five weeks out of the eight-week epidemic. It would not have enough ICU or ventilator beds for the entire epidemic, costing lives.
    Now that government has increased the supply of hospital resources, the next question is whether it can reduce demand through interventions. Non-pharmaceutical interventions typically promote social distancing, isolation and good hygiene. An example of a social distancing intervention is to ban public gatherings and close schools and non-essential businesses, as a significant number of disease transmissions occur at schools, workplaces and public venues. An example of isolation is encouraging families to stay at home if they are caring for somebody sick in the home, as a significant number of disease transmissions occur inside the home. And an example of good hygiene is a media program discouraging shaking hands and encouraging frequent hand-washing. These programs can be effective, but are most effective if implemented early in an epidemic.
    This book assumes that in a democratic country concerned about its economic health and stability, it will be difficult for government to gather sufficient political will to go beyond the simplest measures early in the epidemic——closing schools and the largest public venues, promoting hygiene, and encouraging voluntary home quarantine.(79) It also assumes that it will be unlikely that government will be able to enact the social distancing measures early enough in the epidemic to make a significant difference in an environment of scientific uncertainty. However, later in the epidemic, government is likely to take stronger measures. In any case, social distancing measures are considered unlikely to reduce the total number of cases, or the overall attack rate; instead, their utility is in reducing peak cases, which can prolong an epidemic but ease pressure on the health system.
    As shown in the updated model in Table A-3, this book assumes that the school closings, enacted in weeks 3-4, would reduce peak cases by 10 percent, and that business closings, enacted in weeks 3-5, would reduce peak cases by another 10 percent. This reduces the peak of the epidemic but has the tradeoff of making the epidemic longer by about two weeks, from eight to 10 weeks.
    In the updated model, peak reductions in cases in weeks 4-5 reduce pressure on the health system, but do not solve the problem; Ontario would not have enough acute care beds for five weeks out of the 10-week epidemic, or enough ICU or ventilator beds for the first eight weeks. At eight weeks, the epidemic is virtually over; new cases are dropping fast. However, the combination of a greater supply of and less demand for ICU beds and ventilators saves lives.
    Health systems would have one significant weapon in its arsenal to fight the epidemic, however——pharmaceutical interventions. If a vaccine used as a prophylaxis, it could dramatically reduce the number of clinically ill cases. However, due to production times it will not begin distribution until the first wave is essentially over. That leaves antiviral drugs in Canada’s National Antiviral Stockpile, established in 2004 with a goal of building a stockpile of 16 million doses of two patented antiviral drugs proven effective against influenza. It was later reported that the Stockpile would be increased to 55 million does by 2008.

Table A-3. The epidemic in Ontario with an increase in hospital capacity in addition to a reduction in demand due to government social distancing interventions.

     Antiviral drugs can be used two ways. First, as a prophylaxis, which could significantly reduce the number of clinically ill if sufficient quantities were available. Since the Stockpile as defined above cannot cover the entire population of Canada, this is not possible. Second, it can be used as a treatment for clinically ill to reduce mortality.
    This book assumes that this occurs, that the Flu virus is sensitive to the drugs, and that the drugs are effective at reducing mortality by 50 percent; actual results during a real pandemic may of course be different. The stockpile would almost be large enough to cover the segment of the population needing it during the first wave of our fictional virus at 15 doses per individual.
    Assuming perfect distribution of these drugs (which is unlikely during an actual epidemic), mortality is reduced from 2.5 percent of clinically ill to 1.25 percent, as shown in the updated mortality in Table A-4. The adjusted mortality rate would reduce Ontario’s mortality 50 percent, resulting in nearly 40,000 deaths. Canada’s mortality would also be reduced 50 percent, resulting in about 100,000 deaths.
    Of course, this is a simple model, leaving some questions unanswered. For example: What impact would antiviral drugs have on length of stay in acute care, ICU and ventilator beds? Would the health system have the resources to increase acute care capacity by as much as 50 percent? And perhaps most important: Instead of assuming a flat mortality rate for all clinically ill cases, specifically what impact would the presence or lack of ICU beds and ventilators have on mortality? Since our hypothetical Flu attacks the lower lungs, many patients will end up needing mechnical ventilation to breathe, and without them, their chances are slim.
    Even using basic assumptions and a simple model, however, one could argue that a severe pandemic——as defined by The Canadian Pandemic Influenza Plan for the Health Sector with up to 10 percent of clinically ill people requiring hospitalization——would overwhelm Canada’s health systems and result in mass fatalities, social disruption and severe economic shocks even with additional capacity, antiviral drug and social distancing interventions.

Table A-4. The epidemic in Ontario with an increase in hospital capacity, reduction in demand due to social distancing, and distribution of antiviral drugs as a treatment for majority of clinically ill cases.

This site requires Flashplayer. Download it for free here.

©2008 Future Shock Books, a division of ZING Communications, Inc.