Home | About | Journals | Submit | Contact Us | Français |

**|**HHS Author Manuscripts**|**PMC2941528

Formats

Article sections

Authors

Related links

Int J Infect Dis. Author manuscript; available in PMC 2011 September 1.

Published in final edited form as:

PMCID: PMC2941528

NIHMSID: NIHMS224780

Claudia Taylor, University of Pittsburgh ; Email: ude.ttip@4tcc;

The publisher's final edited version of this article is available at Int J Infect Dis

See other articles in PMC that cite the published article.

This research aims to determine if same Influenza vaccination strategies have the same level of effectiveness when applied to two different US metropolitan areas, Miami and Seattle, where the composition of the population differs significantly in age distribution and the household size distribution.

We use an individual-based network modeling approach in which every pair of individuals connected in the social network is represented. Factorial design experiments are performed to estimate the impact of age-targeted vaccination strategies to people to control the transmission of a “flu-like” virus.

The findings show that: (1) age composition of the city matters in determining the effectiveness of a vaccination strategy, (2) vaccinating the school children outperforms every other strategy.

The most significant policy implication of this research is that there may not be a universal vaccination srategy that works across all cities with the same level of effectiveness. Secondly, given the important role of school children in Influenza transmission, the US government should consider the vaccination of school children a top priority.

In a typical year, 10-20% of the US population is infected with the influenza virus [1]. Worldwide, influenza results in 250,000 to 500,000 deaths annually [2]. The primary method for influenza prevention is vaccination, which is usually 60% to 90% effective depending on the individual [2]. However, these vaccines are created based on predictions about what strains of influenza will be most prevalent in a given influenza season. Sometimes, as in the current H1N1 “swine-origin influenza”, a strain of influenza undergoes a sudden genetic shift, meaning that there is no vaccine readily available [3]. In other cases, such as the 2004-2005 factory contamination, the supply of vaccines may be less than expected [4]. In such situations, mass vaccination against influenza (as is attempted yearly) is not possible, and governments need to issue recommendations about how to most effectively use the limited number of vaccines in order to prevent or control a possible pandemic.

This presents an interesting policy dilemma: to whom do we distribute these vaccines? Current CDC recommendations prioritize, in the event of a pandemic: “critical occupations,” including deployed forces, healthcare workers, and emergency responders, and the “high risk population,” consisting of pregnant women, infants, and toddlers [5]. This first tier for vaccination comprises 24 million people. The reasoning for prioritizing these groups is that the critical infrastructure workers are vital to keep the nation running, and vaccinating pregnant women, infants, and toddlers will protect the highest-risk groups of the population [5].

There is a significant body of publications regarding influenza vaccine distribution [5-12]. Prioritization for vaccination, of course, depends partly upon the goal to be accomplished with the vaccine: various goals include protecting those most at risk, minimizing the number of infections, reducing influenza-related mortality, ensuring public order, saving the greatest number of life years, and reducing the economic costs of an influenza outbreak. In a pandemic, one vital priority is to slow transmission of the disease in order to prevent it from spreading out of control [9,10,11,16].

Those most important in sustaining transmission of influenza in the community are schoolchildren, and their vaccination may have a significant indirect effect on the rest of the community through increased herd immunity [7-11,13-15, 18]. Therefore, it is reasonable to consider other vaccination strategies, including placing a higher priority on vaccinating schoolchildren.

The focus of this research is to determine the effects of age-targeted vaccination on the transmission of influenza, not only among the general population but also among varying age groups, and household sizes. The goal is to study the results of applying the same vaccination strategies in two different metropolitan areas, Miami and Seattle, where the population differs significantly in age and household size distributions.

Please see the supplemental material for details.

This study uses factorial design experiments to estimate the impact of different vaccine distribution strategies on the populations of two geographic regions, the Miami and Seattle metropolitan areas. These regions were selected because they differ significantly in age and household size distributions (see Tables 1 and 2). For Miami and Seattle, respectively, schoolchildren (ages 5-18 years) are 15.03% and 20.33% of the areas’ populations, and seniors (aged 65+ years) compose 13.18% and 9.80% [17]. Preschool children (ages 0-4 years) and adults (ages 19-64 years) occur in about the same proportions in the two populations. The distribution of household sizes in the two regions is given in Table 2. There the proportion of Small (single person), Medium (2 or 3 person), and Large (more than 3 person) households is given. While there are more school-aged children in Seattle, the household sizes in Miami are generally larger, with 53.83% of the households having more than 3 persons.

We believe that the difference in the age distribution of the populations will play a significant role in the performance of the age targeted vaccination strategy. To analyze this hypothesis we simulate distribution of influenza vaccines according to the following age groups: preschool; school-age; adults; and senior citizens. For both areas, we distribute the vaccine either at random across the population or to one of the age groups, in an amount equal to 10% of the total population of the area. We assume vaccine efficacy to be 67%, and vaccination begins when .01% of the population is infected. The outbreak originates in 5 randomly chosen infected individuals. Each experiment simulates the passage of 300 days and is repeated 25 times to overcome the effect of the stochastic nature of the simulations. For each area, we also run a base case in which no interventions take place.

The attack rates in the general population in the base cases were estimated by taking the mean proportion of infections from the 25 replicates. These are compared to the attack rate under each vaccination strategy. We also reviewed average attack rates by age group, the groups being: preschool (< 5 years); school-aged (5-18 years); adults (19-64 years); and seniors (65 years+). The average attack rates for the three household sizes, small (single-person), medium (2-3 people), and large (4+), for each region are computed.

- The larger proportion of school aged children in the Seattle region and larger household sizes in the Miami region tend to balance each other in terms of differences in the overall attack rates between the two regions.
- The vaccination strategy of inoculating school children has different effectiveness in the two US cities. The assumption of the study is that there is only enough vaccine to inoculate 10% of the total population. In this study then two-thirds the school age children in Miami and only half of them in Seattle are vaccinated. This differential in vaccination percentages leaves Seattle with twice the attack rate of Miami.
- The strategy of vaccinating the school children outperforms every other strategy. Not only was vaccinating schoolchildren the best strategy globally for the population as a whole, it was the best strategy locally for almost every age group.
- The outcome from vaccinating adults is almost as bad as the base case or no intervention at all. Vaccinating the seniors is marginally better than vaccinating the adults.
- The large households bear the biggest burden of the disease.

The bars on the left hand side of Figure 1 through Figure 5 show the results from the base case simulations where influenza was allowed to spread with no interventions. The average attack rate in Miami was estimated to be 28% [Figure 1]. Of those infected, 55.8% were adults (who make up over 65% of the area's population), and 30.8% were schoolchildren (who are 15% of the area's population) [Figures 2 & 3]. In Seattle, the estimated average attack rate was 28.6% [Figure 1]. Adults are 63% of that area's population, but they accounted for just under 50% of the cases of influenza, while schoolchildren accounted for nearly 40% of illnesses despite composing only about 20% of the population. [Figures 2 & 3]

In terms of household size we see a difference of nearly 24% between attack rates in small and large households in both Miami and Seattle. In Miami, the estimated baseline attack rates are 12.8% for single-person households, 19.3% for medium-sized households, and 36.4% for large households. In Seattle, estimated attack rates are 15.3% for small households, 22% for medium households, and 39.1% for large households [Figures 4 & 5]. There are highly significant differences between the attack rates in the three household groupings for both Miami and Seattle. A higher percentage of persons in larger households become infected. In larger households this is a consequence of more household members being in contact with infected persons in the household and the fact that, on the average, larger households have more children.

A fascinating result of the base case simulations is the relative consistency of the final attack rates in Seattle and Miami. From the figures, it is apparent that a higher percentage of school age children and larger household sizes increase the attack rate in the region. In the base case, the higher percentage of school age children in Seattle (20.33%) than in Miami (15.03%) is countered by the larger households in Miami, where 53.83% of the population lives in households with 4 or more persons compared to 43.34% in Seattle. These two demographics have offsetting effects, and the attack rates for the base cases in the two regions are nearly equal.

All intervention strategies were found to reduce the attack rate in both areas. With vaccination of 10% of the population of these metropolitan areas, attack rates in some cases dropped sharply. These are shown by the 5 sets of bar plots on the right hand side of the figures. Which age group received the vaccine was very important to the effectiveness of the vaccination strategy depends on which of the 5 age groups received the vaccine. Vaccinating adults dropped attack rates to 25.12% in Miami and 26.62% in Seattle; vaccinating seniors brought the attack rates to 23.97% in Miami and 23.68% in Seattle. Vaccinating preschoolers resulted in an attack rate of 23.65% in Miami and 24.45% in Seattle. For both cities, vaccinating 10% of the population at random performed better than any of the above strategies, with attack rates of 21.18% in Miami and 21.24% in Seattle. However, for both the cities, best strategy that resulted in the lowest attack rate was vaccinating school-age children: Miami's estimated attack rate under this strategy is 5.50%, while Seattle's is 10.86%. Not only was vaccinating schoolchildren the best strategy globally, it was the best strategy locally for every age group except preschoolers in Miami and preschoolers and seniors in Seattle. For these subgroups vaccinating schoolchildren was not the local optimum; the optimum was vaccination of their own age group [Figures 2 & 3].

All vaccination strategies reduced attack rates in all income groups, with the greatest reduction resulting from the vaccination of schoolchildren. Reduction of attack rate was consistent with reduction in the general population.

For all household size subgroups, vaccination of school-aged children was the best strategy. It is the only strategy that came close to evening out the attack rate difference between household sizes. In Miami, under this strategy, the attack rates were 4.5% for small households, 4.7% for medium households, and 6.2% for large households. In Seattle, attack rates when vaccinating schoolchildren were 8.5% for small households, 9.8% for medium households, and 12.4% for large households. Interestingly, the groups that faced the highest attack rates gained the most from this strategy i.e. the large families.

Vaccinating school children with a fixed supply of vaccine that totals 10% of the complete populations in Miami and Seattle leads to some very interesting comparisons. In Seattle 20.33% of the population is composed of school children. Therefore, in this study where the vaccine is available to only 10% of the population, one half of the school children in Seattle are vaccinated. In Miami where the percentage of school children is 15.03%, about 2/3 of the school children are vaccinated. This differential in the percentage of school children vaccinated has a great effect on the attack rates in the two regions. From Figure 1, this scenario leads to an attack rate in Seattle of 10.86% but it is only half that in Miami where the attack rate is just 5.5%.

Our results corroborate others which suggest that age is an important factor in disease transmission. The disproportionate attack rate among schoolchildren, and the vast reduction in the overall attack rate among all subpopulations when schoolchildren are vaccinated, shows them to be crucial disease vectors. Vaccinating school children reduced the overall attack rate by 18-22% and even more notably, by 27-30% in large households. Since in a pandemic one of the main concerns is quelling disease transmission, our findings suggest that it may be prudent to give school children a higher priority in the pandemic influenza preparedness plan.

A 10% level of vaccination is a reasonable figure; the CDC's goal is to have a stockpile of vaccines able to cover 6.7% of the population, and obtain more as soon as there are clear signs of a pandemic outbreak [5]. However, many of these vaccines will go to healthcare and emergency workers crucial to keep the health system running, so actual vaccination levels among non-critical personnel may be lower in the early stages of a pandemic.

The strength of this study is that it is an exceptionally high-resolution simulation of an influenza outbreak in a social network. The individuals in the simulation behave much like the real individuals on whom they are modeled. The simulation is therefore able to return accurate results about transmission of the disease. The fact that vaccinating schoolchildren has a greater effect in Miami than in Seattle makes two important points: (1) age composition of the city matters in determining the effectiveness of a vaccination strategy, (2) school aged children are an important vector in the spread of the disease.

The dramatic results in favor of vaccinating school children show that vaccination of children can be expected to significantly reduce the transmission of the influenza virus. The US pandemic influenza guidelines published in June 2009 recommends vaccination of the school-aged children only after 39 million others, or after 10% of the U.S. population, have been vaccinated [5]. The results of this research have important implications for the policy makers. The most important is that there may not be a universal vaccination strategy that works across all cities with the same level of effectiveness. It is important to be cognizant of the differences in the demographics of the cities to accurately estimate the performance of different intervention strategies. Secondly, in light of this research and other recent studies on the important role of school children in influenza transmission, the United States government should consider the vaccination of school children a top priority.

This work has been partially supported NSF Nets Grant CNS- 0626964, NSF HSD Grant SES-0729441, CDC Center of Excellence in Public Health Informatics Grant 2506055-01, NIHNIGMS MIDAS project 5 U01 GM070694-05, NIH MIDAS project 2U01GM070694-7, NSF PetaApps Grant OCI-0904844, DTRA R&D Grant HDTRA1-0901-0017, DTRA CNIMS Grant HDTRA1-07-C-0113 and NSF NETS CNS-0831633.

**Publisher's Disclaimer: **This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

**Conflict of Interest Statement:** The authors have no conflict of interest.

Claudia Taylor, University of Pittsburgh ; Email: ude.ttip@4tcc.

Achla Marathe, 1880 Pratt Drive, Bldg XV, Virginia Bioinformatics Institute, Virginia Tech, Blacksburg, VA 24061. Email: ude.tv.ibv@ehtarama.

Richard Beckman, Virginia Tech, Email: ude.tv.ibv@namkcebr.

1. Sullivan KM, Monto AS, Longini IM., Jr Estimates of the US health impact of influenza. Am J Public Health. 1993;83(12):1712–6. [PubMed]

2. WHO World Health Organization: Influenza (Seasonal) 2009. [August 5, 2009]. http://www.who.int/med iacentre/factsheets/fs211/en/index.html.

3. Kilbourne ED. The Influenza Viruses and Influenza. Academic Press; New York: 1975.

4. Enserik M. Influenza: Crisis Underscores Fragility of Vaccine Production System. Science. 2004;306(5695):385. [PubMed]

5. Shimabukuro T, Adirim T. Review of Existing Pandemic Influenza Vaccine Priority: Group Guidance. ACIP; Jun 252009. [August 4, 2009]. http://www.cdc.gov/vaccines/recs/acip/downloads/mtg-slides-jun09/15-5-inf.pdf.

6. Holmberg SD, Layton CM, Ghneim GS, Wagener DK. State Plans for Containment of Pandemic Influenza. Emerging Infectious Diseases. 2006;12(9):1414–7. [PMC free article] [PubMed]

7. Patel R, Longini IM, Jr, Halloran ME. Finding optimal vaccination strategies for pandemic influenza using genetic algorithms. J Theor Biol. 2005;234:201–12. [PubMed]

8. Longini IM, Jr, Halloran ME. Strategy for Distribution of Influenza Vaccine to High-Risk Groups and Children. Am J Epidemiol. 2005;161(4):303–6. [PubMed]

9. Piedra PA, et al. Herd immunity in adults against influenza-related illnesses with use of the trivalent-live attenuated influenza vaccine (CAIV-T) in children. Vaccine. 2005;23:1540–8. [PubMed]

10. Weycker D, et al. Population-wide benefits of routine vaccination of children against influenza. Vaccine. 2005;23:1284–93. [PubMed]

11. Basta NE, et al. Strategies for Pandemic and Seasonal Vaccination of Schoolchildren in the United States. Am J Epidemiology Advance Access. August 13;:2009. [PMC free article] [PubMed]

12. Emanuel EJ, Wertheimer A. Who Should Get Influenza Vaccine When Not All Can? Science. 2006;312(5775):854–5. [PubMed]

13. Longini IM, Jr, Koopman JS, Monto AS, Fox JP. Estimating household and community transmission parameters for influenza. Am J Epidemiology. 1982;115(5):736–51. [PubMed]

14. Fox JP, Hall CE, Cooney MK, Foy HM. Influenza infections in Seattle families, 1975-1979. Am J Epidemiology. 1982;116(2):228–42. [PubMed]

15. Reichert TA, et al. The Japanese Experience with Vaccinating Schoolchildren against Influenza. N Engl J Med. 2001;344(12):889–96. [PubMed]

16. Halloran ME, Longini IM., Jr Community Studies for Vaccinating Schoolchildren Against Influenza. Science. 2006;311(5761):615–6. [PubMed]

17. U.S. Census Bureau Census 2000. http://www.census.gov/main/www/cen2000.html.

18. Vibout C, et al. Risk factors of influenza transmission in households. Br J Gen Pr. 2004;54:684–9. [PMC free article] [PubMed]

19. Barrett CL, Smith JP, Eubank S. Modern Epidemiology Modeling. Scientific American. 2005

20. Eubank S, Guclu H, Anil Kumar VS, Marathe MV, Srinivasan A, Toroczkai Z, Wang N. Modeling Disease Outbreaks in Realistic Urban Social Networks. Nature. 2004;429:180–184. [PubMed]

21. Barrett CL, Eubank S, Feng X, Marathe MV. EpiSimdemics: an efficient algorithm for simulating the spread of infectious disease over large realistic social networks.. Proc. of the ACM/IEEE conference on Supercomputing (SC); Austin, Texas. 2008; 1-12.

22. Halloran M, Ferguson N, Eubank S, Longini I, Jr., Cummings D, Lewis B, Xu S, Fraser C, Vullikanti A, Germann TC, Wagener D, Beckman R, Kadau K, Barrett C, Macken C, Burke D, Cooley P. Modeling targeted layered containment of an influenza pandemic in the United States. Proc. of National Academy of Sciences (PNAS) 2008;105(12) [PubMed]

PubMed Central Canada is a service of the Canadian Institutes of Health Research (CIHR) working in partnership with the National Research Council's national science library in cooperation with the National Center for Biotechnology Information at the U.S. National Library of Medicine(NCBI/NLM). It includes content provided to the PubMed Central International archive by participating publishers. |