OBJECTIVES: To compare three methods of measuring multimorbidity according to the use of health resources (cost of care) in primary health care (PHC). METHODS: Design: retrospective study using computerized medical records. Setting: thirteen PHC teams in Catalonia (Spain). Participants: assigned patients requiring care in 2009. Main measurements: the socio-demographic variables, co-morbidity and costs. Methods of comparison were: a) Combined Comorbidity Index (CCI): an index itself was developed from the scores of acute and chronic episodes; b) Charlson Index (ChI); and c) Adjusted Clinical Groups case-mix: resource use bands (RUB). The cost model was constructed by differentiating between fixed (operational) and variable costs. Statistical analysis: was developed 3 multiple lineal regression models to assess the explanatory power of each measure of co-morbidity were compared from the of determination coefficient (R2), p<0.05. RESULTS: A total fo 227,235 patients were included. Woman: 55.6%, average age was 44.1 years, mean episodes/year: 4.5; average visits/patient/year: 8.1, the mean unit of cost was €654.2. The CCI explained a R2=50.4%, the ChI a R2=29.2% and RUB a R2=39.7% of the variability of the cost. The ICC is acceptable behaviour, albeit with low scores (1 to 3 points), showed no conclusive results. CONCLUSIONS: The CCI may be a simple method of predicting PHC costs in routine clinical practice. If confirmed, these results will allow improvements in the comparison of the case-mix.