An epidemiology-based model to estimate the rate of inappropriateness of tumor marker requests

Background: Appropriateness of tumor markers (TMs) has been retrospectively studied in limited patients' series, matching the requests to clinical records. Methods to monitor appropriateness suitable for use on a large scale are required. This study aims to establish and validate an innovative model to estimate appropriateness based on the comparison between the number of TMs requested and the expected requests inferred from epidemiological data. Methods: The number of CA15.3, CA19.9 and CA125 requests theoretically expected according to the epidemiology of malignancies in a known geographic area (2 Italian regions) was compared with the number of TMs actually requested - the surveyed requests projected on a regional scale - during a given time span (1 year). The expected number of requests was calculated comparing TMs recommended by guidelines in different clinical scenarios with the prevalence or incidence figures of the examined diseases (carcinomas of breast, pancreas and biliary tract, ovary and endometrium). Results: Suitability of the model was demonstrated with the analysis of 1,891,070 TM requests surveyed in 66 laboratories from Veneto and Tuscany regions. The percentage difference over the total of expected TMs (delta%) ranged from -6.9% for CA15.3 to +1022.6% for CA19.9 in Veneto and from +35.7% for CA15.3 to +1842.6% for CA19.9 in Tuscany. Conclusions: The presented model was effective in demonstrating higher than expected TM request rates, possibly associated with inappropriate use. Moreover, it can be applied on a large scale survey setting since it circumvents the unavailability of clinical information on test orders.