OBJECTIVES:  Despite increasing use of economic evaluation to inform decision-making in Asia, decisions were made in the face of uncertainty due to poor quality evidence, if not absent. As wrong decision made comes with costly opportunity loss, value of information (VOI) analysis can provide a rational approach by estimating the monetary value of reducing uncertainty through future additional research. This study aims to investigate the use of VOI in a case of selection of strategies in the prevention of carbamazepine (CBZ)-induced Stevens-Johnson syndrome and toxic epidermal necrolysis among newly-diagnosed epilepsy patients in Malaysia. METHODS:  A hybrid model of decision tree and Markov model was developed to evaluate three strategies - (i) CBZ initiation without HLA-B*15:02 genetic screening (reflecting current practice); (ii) universal HLA-B*15:02 genetic screening prior CBZ initiation; and (iii) alternative treatment (valproic acid) without HLA-B*15:02 genetic screening. This analysis undertook the societal perspective over a lifetime time horizon. The population expected value of perfect information (PEVPI) and partial EVPI (EVPPI) were calculated for the three model parameter subsets: probabilities, direct medical costs, and utilities. A 10-year expected lifetime of the technology was assumed. RESULTS: Based on a 10-year stream of 72,422 newly-diagnosed epilepsy population, the PEVPI was MYR739 million (USD185 million) at a willingness-to-pay threshold of MYR36,000 (USD9,000) per quality-adjusted life-year (QALY). The EVPPI results suggested that direct medical costs subset represented the largest amount of uncertainty, at MYR736 million (USD184 million). The EVPPIs of probabilities and utilities subsets were relatively modest at MYR18 million (USD5 million) and MYR11 million (USD3 million), respectively. CONCLUSIONS:  This study demonstrated that VOI analysis is useful in quantifying and decreasing uncertainty surrounding the existing evidence by prioritizing evidence needs to support a decision-making process. As a result, the risk of making wrong decision can be reduced, leading to a better informed decision made.