Can we use contingent valuation to assess the demand for childhood immunisation in developing countries?: a systematic review of the literature.

Childhood immunisation is one of the most cost-effective public health interventions, yet its population coverage in low- and middle-income countries is severely limited by the fiscal constraints that health services face. A recent proposal suggested that commitments to purchase vaccines and make them available to developing countries for modest co-payments could solve the problem. However, this is dependent on communities being willing and able to share the cost in this way, which is difficult to assess. One possible method to assess this demand is contingent valuation (CV). This article evaluates the usefulness of using CV in this way, by reviewing applications of CV in developing countries against current 'standards' for CV of immunisation in the literature. A structured review was adopted with reference to the standard frameworks for methodological evaluation. A set of five criteria were developed for evaluating an 'acceptable' CV study: (i) response rate; (ii) association between willingness to pay (WTP) and socioeconomic status (SES); (iii) sensitivity of WTP to benefit scale/scope; (iv) predictive validity; and (v) reliability in elicitation formats. Two strands of literature search were conducted using electronic databases (MEDLINE, EMBASE, HEALTHSTAR and Econlit) from 1966 to 2003, one for CV studies of immunisation and one for CV studies in developing countries. Twelve CV studies of vaccination and 13 CV studies undertaken within developing countries were identified and reviewed. The quality of existing CV studies conducted in developing countries exceeded the benchmark standard set by studies of immunisation in the developed world in four of the five criteria. WTP estimates appeared both internally valid (i.e. associations with SES) and externally valid (i.e. predictive validity), reliability in developing countries was no less than that of the benchmark level in the existing literature, and the high response rates suggested that CV can be administered to a rural, and perhaps less literate, population. Only sensitivity to scale/scope was not well demonstrated. Our assessment indicated that the CV technique offers a promising tool to estimate the demand for childhood immunisation in low- and middle-income countries. International agencies are therefore encouraged to devote resources to such an application when designing their support to the immunisation programmes.