OBJECTIVES: To derive a mapping algorithm to estimate scores (utility values) for the preference-based SF-6D measures from the non-preference-based disease-specific LupusQoL. METHODS: A total of 282 systemic lupus erythematosus (SLE) patients completed the LupusQoL and SF-6D at the same assessment. Models of the relationship between them were estimated using OLS regression. The SF-6D utility score was modelled using total scores on the 8 LupusQoL domains, employing a backward inclusion procedure. Model performance was judged using the root mean squared error (RMSE) and range of predicted values. RESULTS: The mean (SD) age of the sample was 45 (13.4) years and the mean (SD) SF-6D score was 0.61 (0.13). The mean scores for the LupusQoL domains ranged from 52.5 (Fatigue) to 73.5 (Body Image). Four of the eight LupusQoL domains were selected for inclusion in the final model (Physical Health, Pain, Emotional Health, Fatigue) because these domains were measured in both instruments. The root mean square error (RMSE) for the mapping function was 0.0701, lower than that reported for many published mapping functions. The overall model fit was good (R²=0.7155), although some under prediction at the upper end of the SF-6D was observed. CONCLUSIONS: There appears to be a strong relationship between the LupusQoL and SF-6D. Prediction errors are lower than for many published mapping functions, signifying that the mapping algorithm developed here provides a methodology for predicting SF-6D utility values from LupusQoL data. Potentially this could reduce patient burden if all of the necessary information can be obtain from administering the LupusQoL alone. However, the omission of disease-specific LupusQoL domains (intimate relationships, body image, burden to others, planning) from the final model, raises concerns that the specificity for SLE may be lost in this algorithm. Further out of sample testing will be useful to confirm the performance of this algorithm.