Background: Tumor markers (TM) are a routine test that are not always used well, and can lead to unnecessary additional tests, which are not without risks for the patients. So, to implement appropriate strategies to improve the adequate use of TM and, therefore, improve patient safety, is required to analyze the use of TM, identifying risks and establishing if there are differences in their use as a function of their utility. Methods: The study was a descriptive, longitudinal, retrospective and systematic study in the area covered by the University Hospital of Padua. In the follow-up 2-year study, 23,059 analytical requests of TM, corresponding to 14,728 patients, were analyzed. For the level of statistical significance it applies an approximation of the normal law (Z statistic) and χ²-test. Results: Only 9196 requests (39.88%) out of a total of 23,059 on 5080 patients with neoplastic diseases have been classified as adecuate according to current guidelines. The number of requests per patient was variable (1.57±1.35). In patients with neoplastic diseases this increased to 1.80±1.68. The mean of number of TM per request was 2.4±1.73. The analysis showed an association between the number of requests and the type of marker used. Conclusions: The use of TM is variable, mainly of the follow-up markers, when they are used as screening. This inappropriate use, minimizes their utility favoring erroneous interpretations and increases the risk of damage to the patient. So it is essential to implement safe practices in the use of TM.