Methods Exit surveys were undertaken at inpatient and outpatient health facilities. In order to handle uncertainty and facilitate comparisons, parametric and non-parametric bootstrap methods were used to estimate direct and indirect costs at the individual data level. The inpatient and outpatient incremental costs from Madang and Maprik health facilities were compared and significant differences between provinces were identified. Results Differences were noted between provinces for both inpatient and outpatient household costs. Total arithmetic mean costs for an outpatient malaria episode were US$7.54 in Madang and US$9.20 in Maprik. Total mean inpatient malaria episode costs were US$25.20 in Madang and US$14.08 in Maprik. As cost distributions were not normal, non-parametric bootstrap techniques were used for cost comparisons. Total household costs per outpatient episode of malaria were lower, although not significantly, in Maprik than in Madang (incremental cost of US$ -1.67; 95% CI -4.16, 0.31), while total household costs per inpatient episode were significantly higher in Madang than in Maprik (difference of US$11.16; 95% CI 5.47, 25.33). A difference was noted between provinces in the proportion of indirect costs in total household costs for an outpatient visit: 76% in Madang vs 94% in Maprik. The proportion for indirect costs associated with inpatient visits varied less: 63% in Madang vs 68% in Maprik. Conclusions Intra-country differences need to be considered in estimating household costs for both outpatient and inpatient malaria treatment. Our findings suggest that it is important to recognize the impact of both direct and indirect costs on individuals' capacity to afford treatment. Certain indirect costs are difficult to measure accurately, particularly respondents' interpretations of their productive versus non-productive time. Despite this, exploring intra-country cost variation can provide important information to health policy makers.