s / Annals of Epidemiology 24 (2014) 682e702 683 Purpose: No significant shift to high-volume centers for rectal cancer treatment has occurred despite complex staging, technical surgical challenges, and wide variations in outcomes. We evaluated the impact of facility characteristics on the performance of recommended staging and treatment, including pre-treatment transrectal ultrasound (TRUS), pelvic Magnetic Resonance Imaging (MRI), and carcinoembryonic antigen (CEA) testing, and pre-operative chemoradiation. We hypothesized that guideline-concordant care was provided more often in NCI-designated or academic facilities. Methods: Included were Medicare beneficiaries in SEER regions diagnosed with stage II/III rectal adenocarcinoma from 2005-2009. Patients were>1⁄4 66 years, had Parts A/B Medicare coverage for >1 year pre-diagnosis through 1 year post-diagnosis, and had a facility claim for cancer-directed surgery. Subjects were classified according to NCI designation or medical school affiliation of the treating facility. Logistic regression was used to evaluate association of facility type with staging/treatment received after controlling for patient demographics, stage and co-morbidities. Results: 2,380 subjects (average age1⁄475) met criteria. 61%, 69%, 65% and 72% of subjects treated at NCI-designated facilities received TRUS/MRI-pelvis, CEA, pre-operative chemotherapy and radiation therapy, respectively, compared to 29%, 52%, 40% and 45% at non-designated facilities (all p<.0001). On multivariate analysis, odds ratios (95% confidence intervals) for receiving TRUS/MRI-pelvis, CEA, pre-operative chemotherapy and radiation therapy in NCI-designated facilities were 3.66 (2.71,4.93), 1.93(1.41,2.63), 2.47(1.81,3.37), 2.94(2.12,4.07), respectively. Results by medical school affiliation were similar. Conclusions: Those receiving surgery at NCI-designated or medical schoolaffiliated hospitals received more guideline-recommended care. Initiatives that improve appropriate use of recommended pre-surgical staging and treatment may improve outcomes. P02. Quality of Breast Cancer Care in the U.S.: A National Analysis of Medicare Beneficiaries, 2003-2007 Jan Marie Eberth PhD, Ying Xu MD, MS, Grace L. Smith MD, PhD, MPH, Yu Shen PhD, Jing Jiang MS, Thomas A. Buchholz MD, Kelly K. Hunt MD, Dalliah M. Black MD, Sharon H. Giordano MD, MPH, Gary J. Whitman MD, Wei Yang MD, Chan Shen PhD, Linda Elting DrPH, Benjamin D. Smith MD. University of South Carolina Arnold School of Public Health Purpose: We examined the role of patient, physician, and structural factors in predicting needle biopsy use for breast cancer diagnosis and explored whether needle biopsy receipt was related to better patient outcomes. Methods: We analyzed a large, national cohort of female Medicare beneficiaries with incident breast cancer from 2003-2007. Binary outcomes included early surgical consultation prior to biopsy (needle or open), needle biopsy receipt, and single vs. multiple breast conserving surgeries. Results: The proportion of patients undergoing needle biopsy increased significantly over time (61% in 2003 to 77% in 2007; P<0.01). Variation across hospital referral regions was noted (24% receiving needle biopsy in Bismark, ND and 97% in Lynchburg, VA). Among women who were referred to a surgeon to examine their breast abnormality before biopsy (needle or open), only 54% went on to have a needle biopsy. Women who did not get a needle biopsy were twice as likely to have multiple breast surgeries (AdjRR1⁄42.08, 95% CI:2.07-2.11). Conclusions: Use of needle biopsy is heavily influenced by the patient’s surgeon. Unfortunately, for patients who did not get needle biopsy, 70% ended up having multiple surgeries to remove their breast cancer. We found greater variation by surgeon level factors, than geographic regions in our study. There are several avenues for improving use of needle biopsy: training surgeons to do needle biopsy themselves, educating and empowering patients/PCPs to ask for needle biopsy before surgical consultation, and increasing access to radiologic facilities performing needle biopsy. P03. Intrauterine Devices and Endometrial Cancer Risk: A Pooled Analysis of the Epidemiology of Endometrial Cancer Consortium Ashley Felix PhD, Mia M. Gaudet, Carlo La Vecchia, Christina M. Nagle, Xia Ou Shu, Elisabete Weiderpass, Louise A. Brinton. National Cancer Institute Purpose: Intrauterine devices (IUDs), a long-acting and reversible contraceptive method, are known to induce a number of immunological and biochemical changes in the intrauterine environment. As such, we examined relationships between endometrial cancer (EC) risk and IUD use in a pooled analysis of data collected in the Epidemiology of Endometrial Cancer Consortium. Methods: We pooled data from twelve case-control and five cohort studies resulting in 8,801 EC cases and 15,357 controls. Multivariable-adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were estimated from logistic regression models stratified by study. Results: A total of 1,433 (16.3%) cases and 2,288 controls (14.9%) reported ever use of IUDs, which was inversely related to EC risk (OR1⁄40.82, 95% CI1⁄40.67-0.99). Compared with never use, reduced EC risk was associated with inert IUDs (OR1⁄40.72, 95% CI1⁄40.60-0.85) but not with copper or hormone-releasing IUDs. In models adjusted for IUD type, older age at last use ( 45 years OR1⁄40.52, 95% CI1⁄40.41-0.68, ptrend1⁄40.03) and shorter time since last use (use within 1 year of study entry OR1⁄40.36, 95% CI1⁄40.26-0.49 ptrend <0.001) were inversely related to EC risk. Conclusion: Our data suggest that IUD use, specifically more recent use, is inversely related to EC risk. Future studies addressing reasons for this relationship, notably the extent to which detection biases might be involved as opposed to biologic mechanisms, such as foreign body responses in the endometrium, heavier bleeding (and increased clearance of carcinogenic cells), and localized hormonal changes are needed. P04. Crowdsourcing the Collateral Damage from Breast Cancer Treatment Christine Fischetti MPH, Amaka Obidegwu MPH, CHES, Susan M. Love MD. Dr. Susan Love Research Foundation Purpose: Survivorship research focuses on quality of life after cancer treatment, but is usually initiated by providers and lacks the voice of the patient. In an effort to explore the patient perspective, Dr. Susan Love Research Foundation convened a collaboration of advocacy groups to “crowdsource” women’s questions regarding collateral damage from breast cancer treatment. These concerns will be incorporated into the Health of Women [HOW] Study, an ongoing, online cohort study open to anyone aged 18 or older. Methods: Emails to current HOW participants were sent out July 2013 to solicit questions. In October 2013, a new webpage was designed to collect responses; we then partnered with other advocacy organizations and began collecting responses, which were then categorized to guide questionnaire development. Results: Sixteen advocacy organizations came together to support this project. Emails to current HOW participants resulted in 1191 responses. The website landing page resulted in 5099 responses. Overall, 16.8% of respondents complained of fatigue, 16.3% of memory problems, 15.7% of anxiety and/or depression, and 14.0% of numbness/neuropathy. Other less frequently reported problems included problems with nail growth, vision, hearing, urinary tract infections, and allergy-like symptoms. Conclusion: We received an overwhelming number of responses about collateral damage. Many of the issues are known side effects, while others are less commonly reported. We will next compare responses to previously validated questionnaires, develop a comprehensive collateral damage questionnaire for inclusion in the HOW Study, and generate a report for publication to be distributed to participants and the medical community. P05. Is Rural the Same? How ‘Rural’ Cancer Risk Varies Across the US Amanda Fogleman B.S, Wiley Jenkins PhD, MPH, Georgia Mueller MS. Southern Illinois University School of Medicine Purpose: Multiple studies have examined disparities between rural and urban areas, with some studies finding significantly increased risk for some cancers. However, the data are inconsistent, perhaps due in part to the use of ‘rural’ as a generic term implying homogeneity of risk/protective factors across all rural US residents. This is an intuitively false assumption as the cancer experience would be expected to significantly differ between, say, rural Georgia versus rural Utah. We sought to determine if cancer incidence rates (combined breast, colorectal, lung and prostate) in similar rural areas significantly varied.