Statistical and Economic Aspect of Herd Immunity in Vaccination: An Illustration with Influenza in Taiwan

• Main Authors: Sen-Te Wang, 王森德
• Other Authors: Hsiu-Hsi Chen
• Format: Others
• Language: en_US
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/35317374127265962901

博士 === 國立臺灣大學 === 預防醫學研究所 === 97 === Abstract Introduction Vaccination can not only directly protect persons from infection but it can confer partial protection by raising the coverage rate up to a threshold value, which is so-called herd immunity. Most of the previous studies that evaluated the efficacy of influenza vaccination program lay emphasis on the comparison of the difference of outcome between the vaccinated and unvaccinated group, i.e. in the perspective of individual. The indirect benefit that unvaccinated persons accrue from vaccination policy was largely neglected. The consideration of the benefit as a result of herd immunity is worthy of being investigated. Objectives 1. Develop a generalized computer-aided software for predicting epidemic curve. 2. Apply the extended dynamic models to analyze the effect of herd immunity that the policy of massive vaccination against influenza brings to the population. 3. Apply the extended dynamic models to analyze how the policy of massive vaccination against influenza influences the basic reproductive rate. 4. Explore how the efficacy and cost-effectiveness of vaccination against influenza is underestimated when herd immunity is not considered in the analysis. Materials and Methods We begin with the development of new extended epidemic model with influenza-related complications in it. We then developed a computer-aided software for covering a series of model for producing different epidemic curves. We select influenza A as our illustration for the demonstration of computer-aided software for producing these epidemic curves by using different dynamic epidemic models. Empirical data sources on influenza in Taipei County and nationwide are applied to the extended epidemic model to estimate the transmission parameters that then were translated to the basic reproductive rate and threshold coverage rate of mass vaccination to probably eradicate the viral circulation in the studied epidemic season. The empirical data and the transmission parameters of flu estimated by simulation were applied in cost-effectiveness analysis of mass vaccination. The incremental cost-effectiveness ratio expressed by trading extra cost for per death averted, per flu-related complication (outpatient visiting or/and hospitalization) averted for vaccination against non-vaccination was calculated. We compared the result of cost-effectiveness analysis evaluated by dynamic model to that by fixed model to explore the impact of neglect of herd immunity on efficacy of vaccination. Results According to the surveillance data of CDC, there were 5 outbreaks of influenza in Taipei County in the studied season. The estimated transmission probability of five outbreaks obtained from the simulation by the extended dynamic model was 0.55, 0.6, 0.75, 0.6 and 0.75. Then the basic reproductive rate (R0) can be estimated as 3.3, 3.6, 6.75, 5.4 and 4.5. So the estimated threshold coverage rate of vaccination to eradicate influenza circulation at these 5 outbreaks was 70%, 72%, 85%, 81%, and 78%. To validate the SHIRO model, we applied the model and parameters available from the process of simulation for influenza epidemic in Taipei County to the national situation in Taiwan in the same manner. We found that the simulated epidemic curve of infection fit well to the national empirical situation only except in outbreak 3 (from 2001 Dec to 2002 Feb). In the economic evaluation of vaccination, vaccination with consideration of herd immunity, as opposed to no vaccination, could save 158 of OPD & hospitalization due to flu and 146 of deaths with extra cost of 249,000 US dollars, based on the. The incremental cost-effective ratio was 1,619 US dollars per each complication averted due to flu and 1,705 US dollars per death saved. Vaccination without consideration of herd immunity, as opposed to no vaccination, only could save 18 of OPD & hospitalization due to flu and 11 of deaths with extra cost of 322,000 US dollars. The incremental cost-effective ratio was 17,443 US dollars per each complication averted due to flu and 29,273 US dollars per death saved. Obviously, the benefit of influenza without considering herd immunity is largely underestimated. Conclusion The application of this model to immunization program of influenza in Taiwan found the basic reproductive number is reduced from 12 to 6.75 for model considering her immunity and 8.65 for the model without considering herd immunity. The benefit of vaccination without considering herd immunity has been substantially underestimated. Key words: Influenza, Influenza vaccine, Herd immunity, Dynamic model, Cost-effectiveness analysis