OBJECTIVES: Currently in Brazil, there is a lack of epidemiology data of pregnant women with suspected pre-eclampsia (PE). As long as the ICD can only identify women diagnosed with PE, the aim of this study was to identify patients suspected to have PE through an algorithm based on a retrospective analysis of the medical procedures on an administrative database from private healthcare insurance companies.

METHODS: A database of patients with ICD of PE or hypertension with significantly proteinuria was retrospectively analyzed. Only medical procedures related to suspected PE were previously selected to be evaluated. The study was divided in five steps: 1) analysis of the other ICD before the diagnosed of PE; 2) analysis of the medical procedure related to suspected PE that occurred before the ICD PE; 3) development of an algorithm with a group of procedures that can identify patients suspected to have PE; 4) evaluation of the sensitivity of the algorithm; 5) evaluation of the specificity of the algorithm in the population with ICD not related to PE or hypertension.

RESULTS: 98.6% of the patients had procedures with ICD not identified. Groups of procedures were developed based on its association. The algorithm was developed based on a decision tree including 5 groups of procedures, with 4 or 3 procedures in each. The sensitivity and specificity of the algorithm were 70% and 71%, respectively.

CONCLUSIONS: The algorithm can be used to identify women with suspected PE and gather a more accurately information of its epidemiology and the cost to manage these patients.