OBJECTIVES: In the absence of evidence from randomised controlled trials on the relative effectiveness of treatments, cost-effectiveness analyses (CEAs) may use observational evidence instead. Treatment assignment is not, however, randomised, and naïve estimates of treatment effect may therefore be biased. To appropriately deal with this form of bias, one may need to adjust for observed and unobserved confounders. In this work we explore these adjustment strategies within a case study of negative pressure wound therapy (NPWT) in the healing of surgical wounds healing by secondary intention (SWHSI).  METHODS: Time to healing of SWHSI patients, the main effectiveness outcome, was estimated using i) OLS models, ii) OLS model adjusting for potential confounders and iii) two-stage instrumental variable (IV) models. All econometric models were Bayesian and used MCMC simulation. CEA estimates were obtained for selected models. RESULTS: The case study was a longitudinal cohort study that included 393 participants followed up by on average 500 days. Unadjusted estimates of the additional days NPWT patients take to heal was 69.1 (mean, SE=10.2), compared to other treatments.  When adjusting for observables, 77.2 (mean, SE=19.9) and when using the IV approach, 61.0 (mean, SE=64.9) days were estimated.  NPWT was not cost-effective across all approaches implemented. CONCLUSIONS: This study demonstrates the feasibility of analysing observational evidence for CEA by adjusting for both observable and non-observable confounders. Within the case study, we could not demonstrate that the existing endogeneity affects the effectiveness of NPWT, and thus cost-effectiveness results were consistently negative.