| Contents |
| Cost-effectiveness analysis (CEA) is used routinely to inform vaccination
policy decisions. An incremental cost-effectiveness ratio (ICER) of a
vaccination program aimed at controlling the spread of infections in a population
represents a measure of how efficient that program may be in improving
the health of a population. When estimating these ICERs, however, the traditional
static approach used in CEA methods does not account for the potential
herd-immunity/protection effects of vaccination. To account for these indirect
effects of vaccination when estimating ICERs, we borrow from the field
of mathematical infectious disease modeling to extend the traditional cost-effectiveness
methods using a dynamic approach. We characterize the difference
between the estimates of ICER of vaccination programs under the static
and the dynamic approaches. We use a general SIRS (susceptible-infected-removed-
susceptible) model featuring a vaccine with several properties. The
special case of an SIR model with an all-or-nothing vaccine is studied analytically.
We also numerically simulate the general model, trace the transient
dynamics, and conduct a probabilistic sensitivity analysis. Measures of vaccine
effects differ across the static and dynamic models. We find that the static
model generates ICERs showing vaccination programs are less valuable (i.e.,
less efficient) in improving health than the dynamic model. This gap in efficiency
is biggest for diseases with low basic reproduction numbers and grows
with increases in vaccine cost, waning immunity, and decrement in quality of
life from disease. Analytic results from this study suggest it may be possible
to reduce this bias by adjusting ICERs generated by a static model to better
approximate what the ICER may be for the dynamic model counterpart. We
recommend, however, use of a dynamic model within a CEA when the vaccine
is likely to have important effects on transmission. |
|