All statistical calculations were performed using Stata version 8.0 (College Station, Texas, Stata Corporation, 2003). Of the original sample of 1670 physicians, 120 were ineligible because they were retired or no longer in clinical practice. The final sample size included 1550 physicians, of which 1079 responded (overall response rate: 69.6%). Responders and non-responders were comparable in terms of demographic characteristics (location, gender, and age; p > 0.05). Most responding physicians were from Rome (73.8% of responders vs. 76.9% of non-responders) and male (56.2% of responders vs. 58.9% of non-responders), with a mean age of 50.7 (± 11.5) years (50.0 years INK1197 for non-responders).

The demographic characteristics of the sample were similar to those of all SP600125 Italian physicians, as 60.6% of the members of the National Board of Physicians are male and have a similar age distribution ( ENPAM, 2012). Other demographics,

professional and personal characteristics of the responding physicians are listed in Table 1. Italian physicians’ knowledge of predictive genetic testing for cancer appeared adequate in terms of BRCA1/BRCA2 testing, although knowledge of APC testing was lacking [ Table 2(A)]. Almost half of the sample (42.8%) answered all three questions about BRCA1/2 testing correctly. This knowledge was improved if physicians were exposed to cancer genetic testing during graduate or postgraduate training, and with the increase in the amount of time dedicated to continuing medical education. almost Female physicians were more likely to have adequate knowledge about BRCA1/2 testing, and this knowledge increased if genetic testing laboratories were located in the same geographical area as the physicians’ workplace (Model 1 in Table 3). Only 16.9% of physicians provided correct answers to all three questions about APC testing. This knowledge, as in the previous case, increased with exposure to cancer genetic testing during graduate and post-graduate training and with the amount of time dedicated to

continuing medical education (Model 2 in Table 3). Physicians’ knowledge was satisfactory on the penetrance of BRCA1/BRCA2 mutations, but not regarding the prevalence of hereditary breast cancer. Most physicians knew that the absolute risk of developing breast cancer in the presence of BRCA1/BRCA2 mutations is 40–80%, but less than one third recognized that the percentage of breast cancer cases associated with BRCA1/BRCA2 mutations is 1–10% [ Table 2(B)]. By contrast, knowledge concerning inherited forms of colorectal cancer was inadequate, as none of the surveyed physicians knew that the percentage of colorectal cancer cases associated with APC mutations is less than 5%, and only a small proportion of physicians recognized that the absolute risk of developing cancer in the presence of APC mutations is 100% [ Table 2(B)]. Attitudes toward predictive genetic testing for breast and colorectal cancer were quite heterogeneous (Table 4).