METHODS: We used a 20% sample of national Medicare claims to identify urologists and gynecologists who performed SNM procedures from 2005 to 2010. We then determined physician-level rates of device testing [number of tests/number of patients with di- agnoses of overactive bladder (OAB) or non-obstructive urinary reten- tion] and of device implantation (number of implants/number of tests). We then fit a Poisson model to examine factors associated with de- vice testing. RESULTS: The number of physicians performing SNM test procedures tripled from 2005 to 2010. During the same time period, average rates of SNM testing per physician increased from 2.8 tests to 4.9 tests (p<0.01), while rates of device implantation remained stable (p¼0.80), see figure. Based on our model, we found that physicians who have a history of lower rates of device implantation were associ- ated with higher rates of device testing (p<0.01). Physicians who had high rates of testing were also less likely to be urologists compared to gynecologists, to perform the test procedure in an ambulatory surgery center or in a hospital-based inpatient or outpatient facility compared to the office setting, and were more likely to test on patients who were older and white, all p values <0.01. CONCLUSIONS: Over time, physicians are testing more pa- tients but are not necessarily implanting more devices. In addition, we found that physicians who historically have lower rates of device im- plantation are associated with higher rates of device testing. Better strategies need to be developed to help reduce rates of unsuccessful testing in the future. Source of Funding: AHRQ R01 HS018726-01A1 MP11-12 TRENDS IN TESTOSTERONE REPLACEMENT THERAPY AMONG OLDER MEN IN A REGIONAL VETERANS AFFAIRS HEALTH CARE SYSTEM Thomas Walsh*, Alvin Matsumoto, Alexandra Fox, Kathryn Moore, Christopher Forsberg, Susan Heckbert, Seattle, WA; Steve Zeliadt, Seattle, United States Minor Outlying Islands; MaryLou Thompson, Nicholas Smith, Molly Shores, Seattle, WA INTRODUCTION AND OBJECTIVES: Low serum testosterone (T) is common, is increasingly prevalent with age, and is the only indication for T treatment. Recent studies have reported an “epidemic” of T prescribing over the last decade and there is concern about un- necessary initiation of treatment. The Veterans Affairs (VA) Health Care System provides care to a large population of older men. We investigated trends in serum T testing, the prevalence of low serum T levels among those tested, and the initiation of T treatment among those with low T. METHODS: We identified all men aged 40 and older who used any VA health care in a single Veterans Integrated Service Network (VISN) in the Pacific Northwest from 2002 to 2011. For inclusion, men needed to present for care at a VA clinic on at least one occasion during the year prior to initial T testing. We excluded men with a history of prostate or breast cancer or a history of past T treatment. Low T was defined as an abnormally low value by the testing laboratory. T treat- ment initiation was defined as the first prescription for T. RESULTS: The total number of men in the VISN increased from 129,247 in 2002 to 163,572 in 2011. During this decade, the number of men undergoing serum T testing more than doubled from 4,078 in 2002 to 9,452 in 2011; the proportion of men being tested increased by 83%, from 3.2% to 5.8% of all men. In men ages 40-49, the increase T testing was more substantial over this time period (increasing by 150% from 2.6% to 6.6% of all men) than in other decade-age groups. The observed increase in T testing occurred primarily between 2007 to 2011 (increasing by 71% in all men, and 123% in men ages 40-49). Among those tested, there was an increase over time in the proportion of men found to have low T levels: 35.0% of men tested in 2002 to 47.3% in 2011. The number of men treated with T tripled from 443 in 2002 to 1251 in 2011; however, there was no increase in the proportion of men with low T who were treated with T (31% of men with low T in 2002; 28% in 2011). Throughout the period of study, most men were treated with injectable T (81.4%), as compared to T patches (13.2%) and T gels (5.4%). CONCLUSIONS: Despite secular increases in the absolute number and proportion of T testing and men with low T levels, and absolute number of men treated with T, the proportion of men with low T levels who were treated with T remained unchanged in the Pacific Northwest VA Healthcare System. The decision to treat men with T does not appear to be governed by a low serum T level alone. Source of Funding: Research reported in this publication was supported by the National Institute On Aging of the National Institutes of Health under Award Number R01AG042934. MP11-13 PREDICTORS FOR THE USE OF INTRAVESICAL MITOMYCIN-C FOLLOWING TRANSURETHRAL RESECTION OF BLADDER TUMORS IN THE UNITED STATES: A POPULATION BASED ANALYSIS Nedim Ruhotina*, Jeffrey Leow, Wei Jiang, Boston, MA; Benjamin L. Chung, Stanford, CA; Jonathan Rosenberg, New York, NY; Adam Kibel, Quoc Dien Trinh, Steven L. Chang, Boston, MA INTRODUCTION AND OBJECTIVES: Multiple phase III trials have demonstrated that a single intravesical instillation of mitomycin C (MMC) within 24 hours after transurethral resection of non-muscle- invasive bladder tumor reduces the risk of locally recurrent disease. We analyzed a contemporary population-based cohort to determine the prevalence and predictors of immediate post operative administration of MMC (IPOMC) in the United States. METHODS: We analyzed patient-level data from a proprietary national inpatient discharge database, which collects data from over 400 non-federal hospitals throughout the United States. We captured all men who underwent a transurethral resection of bladder tumor (TURBT) or cystoscopy with bladder biopsy between January 1, 2003, and December 31, 2010, based on the International Classification of Disease, 9th edition, codes 57.49 and 57.33 respectively. We identi- fied the use of intravesical MMC within 24 hours of endoscopic therapy through a detailed review of the hospital charge data. The data were analyzed with descriptive statistics and logistic regres- sion models. RESULTS: Overall, the rate of IPOMC utilization increased from 4.0% in 2003 to 10.0% in 2010. While in providers who administered IPOMC at least once, the rate of IPOMC use increased from 14.6% in 2003 to 25.4% in 2010. There was a higher likelihood of IPOMC use in patients who were older than age 75 compared to patients less than age 55 (OR 1.73, p¼0.03). Patients who had a Charlson co-morbidity index (CCI) of 1 were less likely to receive IPOMC compared to patients with CCI of 0 (OR 0.81, p¼0.01). In subgroup analysis of providers who administered at least one dose of IPOMC, the highest annual volume providers were less likely to administer IPOMC (OR 0.54, p¼0.01) than lower volume providers. Compared to lowest volume hospitals, in the e100 THE JOURNAL OF UROLOGY â Vol. 191, No. 4S, Supplement, Saturday, May 17, 2014