The Biological Weapons Protocol as a

Health Care Intervention:

Cost-effectiveness and Cost-benefit

 

 

Lynn C. Klotz

Federation of American Scientists

Working Group on the Control of Biological Weapons

 

 

Introduction

 

The Biological Weapons Convention (BWC) of 1972 prohibits the development and stockpiling of biological weapons.  Because the BWC has no means to verify compliance with the Convention, it has not prevented the proliferation of biological weapons even in some countries that are parties to the Convention.  The international community recently had an unusual opportunity to strengthen the BWC by enacting a protocol, which an Ad Hoc Group of States Parties began negotiating in Geneva in 1996.  The negotiations resulted in a compromise proposed protocol in 2001, which requires investigations and on-site visits within the geographic boundaries of the States that are party to the protocol.[1]

 

While there is wide support among nations for a protocol, the United States has rejected the proposed protocol on grounds that it would be ineffective in preventing BW development and use and would compromise national security by revealing US defensive strategies for BW.  Unfortunately, the US has blocked enactment of any protocol for the time being.  The following summary of a longer analysis on the FAS website deals with effectiveness of the proposed protocol.

 

The analysis considers the protocol as a health care intervention, which helps prevent sickness and death from a biological weapons attack. This perspective allows us to use standard pharmacoeconomic procedures and benchmarks to analyze the cost-effectiveness and cost-benefits of the proposed protocol. 

 

Pharmacoeconomic analyses are widely used to look at the costs, benefits and effectiveness of health-care interventions to assess health care choices, weighing costs and outcomes of new drug therapies and prevention practices.  Each choice has its cost—and relative beneficial outcome.  Health care analysts use pharmacoeconomics to determine the best “health-care buy.”

 

Seen in the light of quantitative pharmacoeconomics, our analysis shows that a protocol that would provide only a small decrease in the probability of a BW attack on the US would still be highly cost-effective, because its cost is so low compared to the extremely high cost of a successful BW attack of moderate size.

 

Given the present terrorist threat, many believe that the risk of attack is substantial.  For a successful terrorist attack, it is more likely that weaponized BW agents will purposefully or inadvertently come from a State than be developed by terrorists on their own.

 

 

Data for the Analysis and Data Strategy[2]

 

To make a case for US support of the BWC protocol, assumptions regarding the data are conservative, that is, they tend to support the US government’s contention that the protocol is ineffective.  Thus, throughout the analysis, “conservative” means: where good data is not available or is uncertain, data values are chosen that tend to make supporting the protocol less favorable, financially and otherwise.

 

The analysis is carried out for a period of 25 years.

 

Cost to the US to support the protocol

 

The calculation of the cost of the protocol to the US is summarized in Table 1.  The total yearly cost to support the protocol was calculated to be $8.7 million.  It is the sum of three costs: the 22% US contribution to the estimated $30 million yearly operating cost of the international organization to administer the protocol[3], $2 million yearly operating costs of the US implementing organization, and a $60,000 per year cost to facilities to host visits from organization inspectors. 

 

The $30 million annual operating budget for the international organization is based on a staff of 233 persons.

 

For visits to facilities, costs were estimated based on the two-day visit limit set in the Chairman’s text of the protocol[4] and assuming that ten facility employees would be fully occupied by the Organization’s inspectors.  Again, this is a conservative estimate.  The protocol text states that inspectors should make every effort not interrupt the activities of the facility[5], so no cost for production loss in commercial facilities is included, as such loss would be rare.

 

The duties of the US implementing organization for the protocol include collecting declarations from facilities subject to visits, reporting transfers and other activities to the international organization, prepare reports, accompany inspectors on visits (maximum seven per year[6]), etc.  The operating cost of the implementing organization is estimated based on a full time staff of 20 people at $100,000 per year salary plus benefits plus overhead.  This staff is about one-tenth the size of the international organization staff, and so seems reasonable. 

 

The analysis is carried out for a period of 25 years, so the sum of the net present value (NPV) of the $8.7 million cost for each year is the total cost to the US.  In the pharmacoeconomic analysis, it is denoted by COSTS (see equations (1) and (2) below) and is equal to $122 million, using a 5% discount rate to calculate NPV.

 

Statistics for the scale of the attack

 

A BW attack on the US of moderate size is defined here as one in which 30,000 people are infected with 9,000 fatalities.  This attack size is used throughout the analysis.  A well-executed anthrax attack on a US city could cause over 100,000 fatalities[7], and a recent simulation of a smallpox attack on the US called “Dark Winter” estimated that several million people could be killed.[8]  Thus, the BW attack size of this analysis is conservative compared to what we could experience. 

For a smallpox attack, thirty percent of the victims could die.[9]  For an anthrax attack, a higher percentage of victims could die[10].  Since each death costs the US about $1.2 million (see the economic value of a single human life, calculated below), for an attack with an agent that causes greater than 30% fatalities the cost will rapidly rise with the number of fatalities.

 

It is assumed that the average age of victims of a BW attack is 42 years, slightly over the average age in the US of about 37 years[11].  Average age and average life span are needed to determine life-years-saved for cost-effectiveness analysis.

 

The essential data for the analysis are summarized in Table 2.

 

The economic value of human lives

 

The value of human lives and lives saved by a health-care intervention can be expressed in dollars.  In the US, a human life may be valued at $34,000 per year, which is the gross national product per year (GNP) per capita—one measure of the value of a human life in economic terms.  For an average victim’s age of 42 years and with a life expectancy of 77 years[12], each life saved through prevention of a BW attack results in 77-42=35 life-years-saved.  Thus, the average dollar value of each life saved from preventing a BW attack is 35x$34,000=$1.19 million.  This number is used in the cost-benefit analysis. 

 

Direct and indirect costs of the attack.

 

The direct cost of an attack is taken to be the cost of medical treatment of victims and is estimated to be $535 million, based on therapy costs for a number of serious diseases.[13]  How this number is calculated is summarized in Table 3. 

 

However, there are not enough hospital beds at any one location to house 30,000 patients.  For example, the whole state of Massachusetts has only a total of about 20,000 hospital beds,[14] some of which would be occupied by patients who could not be moved..  Then, at first blush it would appear that the cost of treating BW attack victims must be much less than estimated since only a fraction of those infected can be hospitalized, even though some victims could be transported to empty beds in other locations. 

 

It is likely, however, more victims will die if not hospitalized, and human life has economic value.  On average every life saved is worth about $1.19 million (see above), so if 449 additional victims die ($535 million/$1.19 million) which represents an 5.0% increase in deaths (449/9,000) from inadequate medical care, the cost to the US from these additional fatalities equals the whole direct cost of hospitalization.  The anticipation of 5.0% increase in fatalities from inability to hospitalize victims seems reasonable, so the cost to the US for hospitalizing some victims along with additional fatalities among those not hospitalized is likely similar to the cost of hospitalizing all the victims.  We therefore feel comfortable with our estimate based on hospitalizing all the victims, as if that were possible. 

 

For this analysis, indirect cost of an attack is equal to the direct costs for medical care of victims.  This is a serious underestimate as the indirect cost will likely be several times the direct costs.  Indirect costs would include prophylactic antibiotics or vaccines for those that are exposed, but not infected; lost work days of non-infected people who are ordered to stay at home; the cost of quarantining those who are infected or possibly infected; costs of environmental clean-up; costs related to mental or physical care of panicked unaffected individuals; criminal investigation costs, and costs of emergency and military responses to the attack.  Thus, the total cost estimate ($535 million x 2) of about $1 billion, for the BW attack of moderate size is very conservative.

 

A third type of cost is the cost of fatalities. At an economic value of $1.19 million per person (see above), a BW attack with 9,000 fatalities would cost the US $10.7 billion, an amount that also would greatly exceed the direct costs.

  

Data for the likelihood of an attack and protocol effectiveness

 

The conservative probability of 0.01 per year for an attack on the US is chosen for the analysis.  In other words, this low probability translates to a 50% chance of no BW attack of moderate size on the US in the next 69 years (years=log[0.5]/log[1-0.01]).

 

Effectiveness of the protocol is represented simply by the reduction of the probability for a BW attack[15].  For a weakly effective protocol, we assume this reduction to be 10%; that is, the probability of a BW attack becomes 0.009 per year (0.01 x [1-0.1]).  Said another way, a weakly effective protocol would prevent only one attack in ten.  This is the value for reduction in attack probability used in the analysis.  In the analysis, effectiveness of the protocol is represented entirely by the reduction of the probability for a BW attack.

 

The data on likelihood of an attack and protocol effectiveness are summarized in Table 4.  All attack-associated data is weighted by the probability of attack without the protocol or with the protocol in place, as appropriate.

 

Observations from the data

 

Compared to the small financial cost to the US to support the protocol, $8.7 million per year ($122 million NPV for 25 years), a moderate-sized attack on the US would be extremely expensive, minimally $1 billion for an attack, not even including the $10.7 billion economic value of lost human lives.  This logic of this extreme situation is almost akin to Pascal’s logic in his well-known wager for belief in God: “Let us weigh the gain and the loss in wagering that God is… If you gain, you gain all; if you lose, you lose nothing.  Wager, then, without hesitation that He is.”[16]  As in Pascal’s argument, there is much to gain and little to lose in supporting the protocol. 

 

The simple pharmacoeconomic analysis (see below) “transforms” political arguments into a discussion of probabilities, which may help clarify discussion about effectiveness of the protocol.

 

 

A Primer on Pharmacoeconomics[17]

 

Two common pharmacoeconomic measures, cost-effectiveness and cost-benefit, were employed for our analysis.  Cost-effectiveness (CE) is the dollar cost of the intervention divided by the life-years-saved (lys) from the intervention:

 

            CE = COSTS/lys                                                                                             (1)

 

For the BWC protocol, COSTS is net present value (NPV)[18] cost of the protocol to the US over the years that the protocol is in place; lys is the life-years saved from the prevention of a biological weapons attack due to the protocol. 

 

What constitutes a cost-effective or good health-care buy?  Health-care interventions with CE less than the GNP per capita, $34,000 per lys, are cost-effective and so are a good health-care buy.  (In certain cases, society will tolerate higher cost per life-years-saved for some health-care interventions.)

 

In cost-benefit (CB) analysis, both BENEFITS and COSTS are defined and measured in dollar amounts. 

 

            CB = BENEFITS-COSTS                                                                               (2)

 

BENEFITS are the cost-savings from prevention of a BW attack due to the protocol, which when in place, reduces direct and indirect costs of an attack and number of fatalities.  All benefits are expressed as NPV and weighted to account for the low probabilities of attack with no protocol or the protocol in place.

 

 

The model and results

 

In our simple model for pharmacoeconomic analysis, the probability of an attack in each year is independent of whether or not an attack occurred in the preceding year.  Also, the costs of the protocol and of an attack are assumed to be constant over the 25 years for which the analysis is carried out. The cost and probability data for the analysis are summarized in Table 5 and the flow or state diagram for modeling the risk of a BW attack on the US is presented in Figure 1. 

 

From Figure 1, the model begins its yearly progression in the upper box, which represents the state of healthy US citizens of average age 42 years.  In the first year, either the US remains attack free or it suffers an attack and moves to the state represented in the lower box.  In the next year, the US is either attacked again so remains in the lower box, or it is not attacked so moves to the upper box.  The probability of remaining in one state or moving to the other is given by the numbers next to the lines with arrows between the boxes or lines with arrows leaving and reentering the same box. 

 

The particular scenario represented by the probabilities in Figure 1 is for a relatively ineffective protocol.  Effectiveness of the protocol is represented solely by reduction of the probability of an attack--in this particular scenario by only a 10% reduction from 0.01 per year for a BW attack with no protocol in place (italics, normal type) to 0.009 (italics, condensed bold type). 

 

Listed in each box are the costs associated with that box, that is, costs for the protocol and for a BW attack.  Note that the size and cost of an attack remains the same.  The effect of the Protocol, therefore, is represented only in reduction of the probability of an attack. 

 

Cost-effectiveness analysis

 

Calculation of life-years-saved for Equation (1)

 

The calculation is summarized in Table 5 and expanded upon here.  The expected number of lives saved by the protocol is simply the difference in fatalities for no protocol vs. protocol in place:

 

The expected number of fatalities in any single year with no protocol is given by:

 

            [Probability of an attack (no protocol)] x [number of fatalities] =

0.01  x 9000 = 90

 

The expected number of fatalities in any single year with the protocol in place is given by:

 

            [Probability of an attack (protocol in place)] x [number of fatalities] =

0.009 x 9000 = 81

 

Thus, in any one year, the expected number of saved lives from the protocol is 90-81 = 9 and the expected total number of life-years-saved is therefore 9 x 35 = 315.  The NPV for life years saved over the 25 years of the analysis at a discount rate of 5% is 4440 lys (see below).

 

Discounting future costs

 

To determine cost effectiveness of the protocol using Equation (1), the total cost of the protocol to the US, COST=$8.7 million per year, was estimated in Table 1.  Using standard financial practice, the cost of the protocol in all future years must be discounted to net present value (NPV).  Then the total cost of the protocol over all years is obtained.  In equation form, this is

 

CE = {[C1/(1+d)1]+ [C2/(1+d)2]+ [C3/(1+d)3]…+ [C25/(1+d)25]} / lys                                   (3)

 

Where Ci is the protocol cost in year i from the present, and d is the discount rate used in this analysis, which is 5% or d=0.05.   This discount rate is reasonable for well-invested money at the present time.  The last term indicates that the analysis was carried out for 25 years. 

 

In the model, we assume that the costs of the protocol are constant, so Ci = C, so equation (3) becomes

 

  CE = C x {[1/(1+d)1]+ [1/(1+d)2]+ [1/(1+d)3]…+ [1/(1+d)25]} / lys                                   (4)

 

Equation (4) may now be rewritten

 

            CE = C x DF(d)/lys                                                                                          (5)

 

where

 

            DF(d) = 1/(1+d)1]+ [1/(1+d)2]+ [1/(1+d)3]…+ [1/(1+d)25]                               (6)

 

Using the formula for the sum of N terms in a geometric series, Equation (6) may be written more condensed in closed form.

 

            DF(d) = (1/d) x [1-1/(1+d)25]                                                                           (7)

 

For the 5% discount rate used in this analysis DF(d) = 14.09

 

The question immediately arises, for a BW attack on the US in some future year, do we discount life-year-saved as well?  This is a philosophical and psychological question.  One could argue that our children’s and grandchildren’s lives in the future are as valuable as our lives now, in which case we would not wish to discount life-years-saved.  Here, life-years-saved are discounted at 5%, the same as money.  Most of us probably believe that future lives are more valuable than future money, so discounting future lives at the same rate as money is a conservative assumption, because it leads to higher cost/lys values making less of a case for the protocol. 

 

Since the expected lys for any year is constant in our model, so by the arguments leading to Equation (6), the total discounted expected life-years-saved is

 

            TOTAL discounted lys = lys x DF(d)                                                                (8)

 

Thus, discounting life-years-saved at the same rate as money also simplifies our calculations of cost effectiveness, since equation (5) becomes

 

            CE = {C x DF(d)}/{lys x DF(d)}= C/lys                                                          (9)                               

 

That is, the discounting factor cancels out and CE is simply given by the ratio of the undiscounted values of yearly cost of the protocol and undiscounted life years saved, namely

 

            CE = $8.7 million / 315 = $27,500 per lys,

 

where the value 315 lys was calculated above.

 

This cost-effective value is well within the limit for a good health-care buy ($34,000 per/lys); just as significantly, it was arrived at entirely by using only assumptions about the data that were unfavorable to effectiveness of the protocol.[19]  

 

Cost-benefit analysis

 

In equation (2) for cost-benefit, COST is solely the cost to the US for the protocol and benefits will be the cost-savings from prevention of a BW attack due to the protocol; where as before, all costs and benefits must be discounted over the 25 years.  The data before discounting is presented in Table 6.   Again, for our simple model discounted cost-benefit are given simply by

 

                CB = (BENEFITS-COSTS) x DF(d) = (BENEFITS-COSTS) x 14.09                        (10)

 

As in cost-effective analysis, the total cost is the cost of the protocol, which undiscounted is

 

COSTS = $8.7 million

 

The BENEFITS are cost-savings derived from the protocol.  There are two kinds of benefits.

 

(1) Costs saved from preventing an attack, which simply is the difference between the expected cost of an attack without the protocol and with the protocol in place.  From Table 2, this is

 

NPV cost of attack (no protocol) – NPV cost of attack (protocol) =

0.01 x $1,071million – 0.009 x $1,071million = $1.07 million

 

(2) Cost savings from the reduced number of fatalities is the probability weighted life-years-saved times the value of each saved life.  From Table 5                                             

           

            Value of life-years-saved = 315 x $33,885 = $10.7 million

 

Costs saved in (2) are considerably greater than costs saved in (1).

           

Substituting into Equation (10) yields

 

CB = BENEFITS – COSTS = ($1.07 +$10.7 – $8.7) million x 14.09 = $43.5 million

 

Thus, the protocol will actually save the US money; that is the NPV cost-savings are well above the cost of supporting the protocol for 25 years.  This conclusion is for a conservative case where data were chosen to make an unfavorable case for the protocol.

 

A less conservative case

 

For a BW attack of greater scope or with higher indirect cost, the protocol could be considerably less effective, and still be a good heath-care preventative measure.  The indirect costs of an attack could easily be ten times the direct costs, which may still be conservative.  Furthermore, the protocol could be more effective and thwart 20% of all potential attacks. 

 

But, how low can the probability of attack be to ensure that the protocol is still cost effective and has positive benefits over costs; that is, CE is less than $34,000 cost per lys and CB is greater than $0?

 

For this less conservative case, to be cost effective, the yearly probability of attack on the US can be as low as 0.004, which translates to a 50% chance of no BW attack on the US in the next 173 years.  Thus, the protocol would be cost effective even if the risk of attack on the US is very small.  Surely, an international treaty that eventually might be ratified eventually by over 100 States Parties must be at least that effective.  Benefits over costs remain positive as well, CB=$65 million NPV.

 

 

Conclusion

 

Pharmacoeconomics is one analytical tool to gauge the effectiveness of the protocol.  Based on this approach, it is in the US’ interest to support a BWC protocol.  The US’ arguments for rejection of the proposed protocol—on the grounds of effectiveness—are without merit. 

 

 

Figure 1. State diagram for the pharmacoeconomic model

 

 


Cost for protocol organization

 

$30,000,000 

per yr

Percentage of organization cost paid by US

 

22%

 

Cost to US

 

 

$6,600,000

per yr

 

 

 

 

 

 

Facility site costs to host visits in US

 

 

 

(10 people, 2 days @ $75/hr, 5 visits per yr)

$60,000

per yr

 

 

 

 

 

 

 

Cost of US "implementing organization"

 

 

 

(20 people at $100,000 per year per person, fully loaded)

 

 

$2,000,000

 

per yr

 

 

 

 

 

 

 

 

 

 

 

 

Total cost of protocol:

 

 

$8,660,000

per yr

 


Table 1.  The cost of the protocol to the US.

 

 

 

Size of population infected in a BW attack:

 

30,000

Average age of person infected

 

45 yrs

Average life span in US

 

 

80 yrs

Average life-years-saved for prevented fatality

35 yrs

Percent fatalities:

 

 

30%

Number fatalities

 

 

9,000


Table 2.  The statistics for the BW attack of small to moderate size of this analysis.

 

 

 

Average length of hospital stay (survivors):

 

 

14 days

 

 

 

 

Average length of hospital stay (fatalities):

 

 

7 days

  

 

 

 

 

 

Average direct daily cost of treatment in hospital

$1,500

 

 

 

 

 

 

Total direct hospitalization cost of attack

 

 

$535,500,000

 

 

 

 

 

 

Indirect costs of attack (as % of direct costs)

 

100%

 

 

 

 

 

Total indirect cost

 

 

 

$535,500,000

 

 

 

 

 

 

 

TOTAL cost of attack to US:

 

 

$1,071,000,000


Table 3.  Estimates of the cost of a BW attack on the US where 30,000 people are infected with 9,000 fatalities

­­­­­­­­­­­­­­­­­­­­­­

 

Yearly probability of BW attack (no protocol)

0.01

per yr

Protocol effectiveness (% reduction in prob. of BW attack)

10%

 

Yearly probability of BW attack (protocol in place)

0.0090

per yr


Table 4.  Summary of assumptions used for conservative analysis case for a weakly effective protocol.

 

 

 

 

No Protocol

Protocol in Place

Yearly Probabilities:

 

 

Attack

0.010

0.009

No Attack

0.990

0.991

 

 

 

Yearly Cost:

 

 

Protocol Implementation

$0

$8.7 million

 

 

 

Attack Related Costs:

 

 

Direct Cost of Attack

$535 million

$535 million

Indirect Cost of Attack

$535 million

$535 million

Cost of Fatalities

$10,700 million

$10,700 million


 

Table 5. Data used in pharmacoeconomic analysis.  Note that attack related cost are only realized in years when an attack on the US occurs.

 

FOOTNOTES


[1] See for example: B. Rosenberg, “Allergic Reaction: Washington’s response to the BWC Protocol,” Arms Control Today, July/August 2001. Link: http://www.armscontrol.org/act/2001_07-08/rosenbergjul_aug01.asp

 

[2] The detailed description of the data and analysis can be found in the FAS Working Group Papers section of the FAS Chemical and Biological Arms Control Project web site (http://www.fas.org/bwc/papers/protocol_ce.htm) in the paper entitled "The Biological Weapons Protocol as a Health Care Intervention: Cost-effectiveness and Cost-benefit." The spreadsheet developed for this analysis allows testing of many scenarios and is available from Lynn Klotz (lynnklotz@compuserve.com)

 

[3] Details and a spreadsheet for the operating budget may be found on the FAS web site  (http://www.fas.org/bwc/papers/structure/start.htm).  The US 22% contribution is from the UN scale of assessment.

 

[4] Chairman’s text of the “Protocol to the Convention on the Prohibition of the Development, Production and Stockpiling of Bacteriological (Biological) and Toxin Weapons and on their Destruction,” Article 6, B, 17.  The text may be found on the FAS website. Link: http://www.fas.org/bwc/protocol.htm

 

[5] Chairman’s text, op. cit., Article 6, B, 31b,c.

 

[6] Chairman’s text, op. cit., Article A, 7 a.  While a total of seven clarification plus transparency visits maximum are allowed per year, it is assumed that the maximum will not be reached in most years, so for the analysis here five visits was assumed.

 

[7]G. Christopher, et al., “Biological Warfare: a Historical Perspective,” Journal of the American Medical Association (JAMA), August 1997.   U.S. Congress, Office of “Technology Assessment, Proliferation of Weapons  of Mass Destruction: Assessing the Risk, OTA-ISC-559 (Washington, DC: U.S.  Government Printing Office, August 1993).  Link: http://www.wws.princeton.edu/~ota/disk1/1993/9341_n.html pg 54

 

[8] Disease by Design: De-mystifying the Biological Weapons Debate.  Crowley, Michael BASIC Research Report 2001.2, November 2001.  Link: http://www.basicint.org/BWreport.htm Part II, Section 7.3

 

[9] Thirty percent fatalities is commonly accepted for smallpox.  See, for example, the links http://www.hopkins-biodefense.org/pages/agents/agentsmallpox.html and http://www.who.int/emc/diseases/smallpox/faqsmallpox.html

 

[10] See, for example, http://www.fda.gov/cber/vaccine/anthrax.htm or Meselson M, Guillemin J, Hugh-Jones M, et al. “The Sverdlovsk anthrax outbreak of 1979.” Science 1994;266:1202-1208.

 

[11] Reference: Mean age of Americans: 36.6 (Nov. 1, 2000).  Population Estimates. Program, Population Division, U.S. Census Bureau, Washington, D.C. 20233.  Internet Release Date:  January 2, 2001.  Link: http://eire.census.gov/popest/archives/national/nation2/intfile2-1.txt

 

[12] Reference: Life Expectancy of Americans: 76.7 yrs (1998).  Anderson RN. National Vital Statistics Reports; vol 48 no 18. Hyattsville,  Maryland: National Center for Health Statistics. 2001.  Link: http://www.cdc.gov/nchs/fastats/lifexpec.htm

 

[13] Hospital cost and stay estimates were derived from several different severe illnesses:  (1) Respiratory syncytial virus (RSV) among the elderly costs $11,000 per RSV pneumonia hospitalization (Han LL, Alexander JP, Anderson LJ. J Infect Dis  1999 Jan;179(1):25‑30).  (2) Costs for treating viral meningitis are over $2,000 per day, but hospital stays are short, 4 to 5 days. (Parasuraman TV, Deverka PA, Toscani MR. Managed Care  2000 Jan;9(1):41-6). (3) For, severe sepsis, the average costs per case were $22,100 (Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Crit Care Med  2001 Jul;29(7):1303-10).  (4) Estimate three weeks in hospital for smallpox survivors (Forman J. Boston Globe, Tuesday November 20, 2001, page C1)

 

[14] Massachusetts has 17,400 community hospital beds and perhaps 20% more total beds (Web version  (www.census.gov/prod/www/statistical-abstract-us.html) of US Census Bureau, Statistical Abstract of the United States 2000, health and Nutrition, Tables No. 204 and 205, Page 136-137.

 

[15] Some opponents to the protocol would argue that the protocol could increase the probability of attack, not reduce it, because it would give us a false sense of security or reveal our defensive strategies.  Almost all protocol supporters understand that the protocol is only one of a number of measures needed to reduce the risk of biological weapons, so they would not have a false sense of security nor reduce efforts in other areas to prevent BW proliferation.  Regarding revealing defensive strategies, the managed access procedures built into the proposed Protocol are more than adequate to protect confidential national-security information, as the US has stated to be the case in the Chemical Weapons convention.

  

[16] From the internet version of the Stanford Encyclopedia of Philosophy (http://plato.stanford.edu/entries/pascal-wager/)

 

[17] See for example: BR Luce and A Elixhauser, “Socioeconomic Evaluation and the Health Care Industry,” Journal of Research in Pharmaceutical Economics, Vol. 2(4) 1990.

 

[18] Throughout the analysis, the interest or discount rate used to calculate NPV is 5%.

 

[19] The spreadsheet developed for this analysis allows testing of any data and is available from Lynn Klotz (lynnklotz@compuserve.com).