OBJECTIVES: The objective of the study is to derive a method to predict a cohort EQ-5D preference-based index score using published statistics of the eight dimension scores describing the SF-36 health profile. METHODS: Ordinary least square regressions are used to obtain models from patient level data covering a wide range of health conditions. The eight dimension scores, the squares age and gender are used to derive a relationship with the EQ-5D index. Models obtained are compared for goodness of fit using standard techniques such as descriptive statistics, variance explained, the residuals and the proportion of values within the minimal important difference. Predictive abilities are also compared when using summary statistics from both within-sample subgroups and datasets published studies. RESULTS: The models obtained explain more than 56% of the variance in the EQ-5D scores. For the individual predicted values, the mean predicted EQ-5D score is correct to two decimal places and the mean absolute error is approximately 0.13. Using summary statistics to predict within-sample subgroup mean EQ-5D scores, the mean errors (mean absolute errors) range from 0.021 to 0.077 (0.045 to 0.083). When predicting baseline cohort EQ-5D scores using published mean dimension scores the models produce mean errors ranging from 0.048 to 0.099 with 76% of values correct to within the minimal important difference. When predicting out-of sample incremental differences between study arms and incremental changes over time, over 71% of values are within the minimal important difference. CONCLUSIONS: The models provide researchers with a mechanism to estimate EQ-5D utility data from published mean dimension scores. This research is unique in that it uses mean statistics from published studies to validate the results. While further research is required to validate the results in additional health conditions, the algorithms can be used to derive additional preference-based measures for use in economic analyses.