Japanese encephalitis (JE), a mosquito-borne viral infection, remains a major public health problem in Asia, reportedly causing 16 000 to 50 000 acute encephalitic episodes and 5000 to 10 000 deaths annually. The syndromes caused by the virus range from encephalomyelitis to mild febrile illness. Where JE is endemic it is principally a disease of children, the highest rates occurring in children aged under 10 years living in rural areas, especially where rice is grown and pigs are reared (1). Furthermore, long-term neurological disability occurs in an appreciable fraction of JE survivors (2).
In Japan and the Republic of Korea, national programmes of routine immunization with inactivated JE vaccine derived from mouse brain have nearly eliminated the disease, even in the areas of highest risk (1). Since 1968, JE vaccines have been provided for certain populations at high risk in China. Early success in controlling the disease in China was achieved by means of an inactivated vaccine that was developed and produced locally. It was prepared from the P3 strain of the JE virus (3) and approximately 70 million doses were administered annually (4). In 1988 a live attenuated vaccine (SA 14-14-2), also developed and produced locally, was licensed for use in China after studies had demonstrated its safety and a high degree of efficacy (1). This vaccine has been routinely used in provinces of south-west China, more than 20 million doses being given annually (4).
Because these Chinese vaccines can be produced inexpensively, they are potentially attractive tools for other developing countries in Asia where JE is endemic. However, no studies on their cost-effectiveness have been reported. In order to obtain information on this matter we calculated the cost-effectiveness of a JE immunization programme in Shanghai, China, where the use of JE vaccines over many years provides an empirical basis for evaluating the economic consequences of vaccination.
We compared the costs and outcomes for three hypothetical cohorts of 100 000 neonates followed up to the age of 30 years which received the following: no JE vaccine; inactivated vaccine (P3); or live attenuated vaccine (SA 14-14-2). The setting for the analysis was Shanghai, where JE is endemic and JE immunization has been implemented since 1968 (Fig. 1). The purpose of following up to 30 years of age is to capture the disability associated with JE. Although most cases occur in young children in settings where the disease is endemic, the neurological sequelae can be frequent and permanent. Because we express costs in 1997 US$, the analysis models a birth cohort assembled in 1997 and followed up for 30 years after that date. The risk for JE in the cohort that does not receive JE vaccine is taken as the cumulative JE incidence rate from birth until 30 years of age, estimated from data reported during the prevaccination era in Shanghai between 1952 and 1967.
[FIGURE 1 OMITTED]
Construction of decision tree
In order to compare the three strategies, a decision-analytical model was constructed (DATA, Version 3.5, TreeAge Software, Inc., Boston, MA, USA). The model considered all costs and outcomes related to JE in the hypothetical cohorts from birth until 30 years of age. The routine JE vaccination schedule in Shanghai was used in the model. The P3 vaccine was given in a five-dose schedule administered as two doses one week apart at 12 months of age and one dose at 2, 6, and 10 years of age. The SA 14-14-2 vaccine was administered in two doses, one at 12 months and the second at 2 years of age. A simplified decision tree for the analysis is presented in Fig. 2.
[FIGURE 2 OMITTED]
As in routine public health practice, children in the two cohorts targeted for vaccination could either receive or not receive vaccine, and, among those vaccinated, either complete or incomplete regimens could be administered. …