OBJECTIVES: In Health Economics, the estimation of disease specific attributable cost is of major importance. For this estimation, cost data of cases (patients with the disease) and comparable controls (patients without the disease) are often utilized. When individual level data are available, regression and GLM models, addressing issues such as skeweness and heteroscedasticity, can be applied. When only aggregate level data (e.g. sample means and standard deviations per strata) are available, these models may not be appropriate. METHODS: Here, motivated by real pressure ulcer cost data, we propose and study a Bayesian Gamma regression mixed model that utilizes as stochastic nodes both sample means and inverse coefficients of variation. We investigate its performance and goodness of fit (using deviance) using various simulated data and compare it with two linear models, assuming known and unknown cost variances per stratum. We also use the method for estimating pressure ulcer attributable costs. RESULTS: In most cases, the linear models give more accurate estimates of the attributable cost, with significantly shorter computational time. The random effects adapt to the multiplicative nature of the data, posterior means between intercept and slope are positively correlated. CONCLUSIONS: When only aggregate data are available, the simplest linear model seems to estimate the attributable costs sufficiently well. The proposed Gamma model, despite being more theoretically justifiable, is of questionable benefit. Further investigation is needed for refining the Gamma model and selecting appropriate measures of model assessment and comparison.