Assessment of the total cost of FNAC as a diagnostic tool in patients with thyroid nodules
6628 Background: The prevalence of thyroid nodules in the general population ranges from 4 to 7% and fine-needle aspiration cytology (FNAC) is a widely used diagnostic procedure. As the final diagnosis is based on post-surgical pathology of the lesion, a number of patients with nodules classified as benign or indeterminate by cytology may have a postponed final diagnosis, impacting the procedure cost. This study aims to determine the total cost of FNAC, by taking into account of diagnostic mistakes (false-positive and false-negative), failures of FNAC (indeterminate results) and the follow-up of non operated patients. Methods: A Markov model was built to describe the management of patients through time, from the first FNA-t. We derived estimates for diagnostic accuracy values and follow-up from a retrospective study, including all patients who had their first FNAC from 2003 to 2005 at Institut Gustave Roussy. Costs were computed from the viewpoint of the hospital. A microcosting study was performed to determine the unit cost of FNAC. Sensitivity analyses were performed. Results: Out of the 624 patients who had FNAC, 15 were cytologically classified as malignant, 57 as suspicious and 171 as indeterminate. Surgical excision was carried on in 155 cases. The unit cost of FNAC was 121 €. Markov modelling showed that the mean total cost of a true diagnosis including unnecessary surgeries, repeat FNAC and follow-up was 1,145 Euros. The number of false negative and false positive cases had little impact on total cost, whereas it was significantly affected by the percentage (27%) of unsatisfactory specimens. The absence of such specimens would reduce the total cost to 738 Euros (35%). Conclusions: The cost of FNAC exceeds its unit production cost. It depends on the performance of the cytologist and on the rate of indeterminate results that might be reduced by routine ultrasound-guidance. No significant financial relationships to disclose.