Japanese encephalitis vaccination in the Philippines: A cost-effectiveness analysis comparing alternative delivery strategies

Abstract

Introduction: Japanese encephalitis (JE) is a mosquito-borne viral infection of the brain that can cause permanent brain damage and death. In the Philippines, efforts are underway to deliver a live attenuated JE vaccine (CD-JEV) to children under five years of age (YOA), who are disproportionately infected. Multiple vaccination strategies are being considered.

Methods: We conducted a cost-effectiveness analysis comparing three vaccination strategies to the current state of no vaccination from the societal and government perspectives: (1) national routine vaccination only, (2) sub-national campaign followed by national routine, and (3) national campaign followed by national routine. We developed a Markov model to estimate impact of vaccination or no vaccination over the child's lifetime horizon, assuming an annual incidence of 10.6 cases per 100,000. Costs of illness ($859/case), vaccine ($0.50/dose), routine vaccination ($0.95/dose), and campaign vaccination ($0.98/dose) were based on hospital financial records, expert opinion and literature. The societal perspective included transportation and opportunity costs of caregiver time, in addition to costs incurred by the health system.

Results: JE vaccination via national campaign followed by national routine delivery was the most cost-effective strategy modeled with a cost per disability adjusted life year (DALY) averted of $233 and $29 from the government and societal perspectives, respectively. Results were similar for other delivery strategies with cost/DALY ranging from $233 to $265 from the government perspective and $29-$57 from the societal perspective. JE vaccination was projected to prevent 27,856-37,277 cases, 5571-7455 deaths, and 173,233-230,704 DALYs among children under five over 20 consecutive birth cohorts. Total incremental costs of vaccination versus no vaccination over 20 birth cohorts were $6.6-$9.8 million from the societal perspective ($230 K-$440 K annually) and $45.9-$53.9 million ($2.2 M-$2.7 M annually) from the governmental perspective.

Conclusion: Vaccination with CD-JEV in the Philippines is projected to be cost-effective, reducing long-term costs associated with JE illness and improving health outcomes compared to no vaccination.

Keywords: Cost analysis; Cost-effectiveness; Japanese encephalitis; Japanese encephalitis vaccine; Philippines.

Fig. 1

Markov model. All individuals enter the model with no JE. Acute JE implies symptomatic JE and is a tunnel state, meaning that any individual in that health state stays there for exactly one cycle. Those who had acute JE do have a higher mortality rate but must accrue the costs and DALYs of the acute event before transitioning to death. Asymptomatic JE is not associated with higher mortality, costs or DALYs; rather it eliminates any transition to acute JE. Costs and DALYs for acute and post-acute JE are distributed by sequelae presence and severity. Vaccination changes the probability of transitioning from no JE to acute or asymptomatic JE. No other probabilities are changed by presence or absence of vaccination. Each state is associated with an annual cost and disability weight where applicable.

Fig. 2

Typical patient flow for suspected acute JE in the Philippines^ⱡ. ^ⱡThis diagram represents comprehensive treatment for a typical patient entering a health care facility with suspected JE. The diagram was developed based on local expert opinion and iterated upon based on discussions with pediatric neurologists and clinicians in local health care facilities. Percentages included in the “Patient screening + diagnosis; inpatient admission” and “Daily inpatient supportive care” boxes represent the proportion of patients expected to receive the identified health care services. This patient flow map informed the secondary, unit-based costing exercise. Each component of the patient flow was assigned a cost based on unit data provided by facilities in Region III and Regional XI. A population-weighted cost including medicines, supplies, and staff time was applied to determine treatment costs relative to average length of patient stay. This cost represents the typical set of treatments we would expect for a patient with suspected JE without censoring that occurred in financial records when a patient transfers health care facilities. It was used as the high value for costs of treatment in the one-way sensitivity analysis.

Fig. 3

One-way sensitivity analysis of key cost drivers* for cost per DALY averted from the government perspective with Subnational Campaign + National Routine over 20 birth cohorts. *Key drivers were defined as parameters whose impact on model uncertainty was ≥15% of the total cost per DALY averted.