A breast cancer clinical registry in an Italian comprehensive cancer center: an instrument for descriptive, clinical, and experimental research.

In clinical research, many potentially useful variables are available via the routine activity of cancer center-based clinical registries (CCCR). We present the experience of the breast cancer clinical registry at Fondazione IRCCS "Istituto Nazionale dei Tumori" to give an example of how a CCCR can be planned, implemented, and used. Five criteria were taken into consideration while planning our CCCR: (a) available clinical and administrative databases ought to be exploited to the maximum extent; (b) open source software should be used; (c) a Web-based interface must be designed; (d) CCCR data must be compatible with population-based cancer registry data; (e) CCCR must be an open system, able to be connected with other data repositories. The amount of work needed for the implementation of a CCCR is inversely linked with the amount of available coded data: the fewer data are available in the input databases as coded variables, the more work will be necessary, for information technology staff, text mining analysis, and registrars (for collecting data from clinical records). A cancer registry in a comprehensive cancer center can be used for several research aspects, such as estimate of the number of cases needed for clinical studies, assessment of biobank specimens with specific characteristics, evaluation of clinical practice and adhesion to clinical guidelines, comparative studies between clinical and population sets of patients, studies on cancer prognosis, and studies on cancer survivorship.

Risk of death for hematological malignancies for residents close to an Italian petrochemical refinery: a population-based case-control study.

PURPOSE: We investigated the risk of death for hematological malignancies (HMs) in the area surrounding an Italian petrochemical refinery, where atmospheric concentrations of benzene (known carcinogen) had not been adequately monitored in the past. METHODS: We performed a population-based case-control study, using conditional logistic regression to estimate odds ratios (ORs) of HM death, with 95 % confidence intervals (CIs), and p trends, in relation to tertiles of time-weighted average residential proximity to the refinery. We identified 177 HM deaths and 349 sex- and age-matched controls from municipal files. Confounding factors were investigated from interviews with consenting relatives for 109 HM deaths and 178 matched controls. RESULTS: For males and females combined, risk of HM death was unrelated to residential proximity. For females, ORs of HM death by increasing tertiles of proximity were 1, 2.74 (95 % CI 1.48-5.09, significant) and 1.49 (95 % CI 0.76-2.92) (p trend 0.184). For the subgroup of persons who plausibly spent most of their time at home (long-term retired, homemakers or unemployed, 53 cases, 79 controls), the ORs of leukemia plus non-Hodgkin lymphoma death (38 cases, 56 controls) by increasing tertiles of proximity were 1, 3.44 (95 % CI 1.04-11.37, significant) and 3.25 (95 % CI 0.82-12.87) (p trend 0.083). CONCLUSIONS: No increased risk of HM death for males and females combined living close to the refinery was found. However, the findings for females and a subgroup plausibly spending most of their time at home suggest a relation between increased risk of HM death and residential proximity to the refinery.