br Study variables br by zip code and distance

2.3. Study variables
>21%) by zip code, and distance traveled to hospital. Hospital variables included facility location (Northeast, South, Midwest, and West), facility type defined as non-academic or academic (minimum of 4 areas of postgrad-uate medical education and more than 500 newly diagnosed cancer cases each year), location type defined as metro, urban, or rural. Hospital volume was catego-rized into quartiles based on annual RC volume (low: 1 −3 cases, low-intermediate: 4−7 cases, high-intermedi-ate: 8−16 cases, and high: 17−98 cases). Pathologic stage was categorized by the American Joint Committee on Cancer TNM system (extent of the tumor (T), extent of spread to the Malonyl Coenzyme A nodes (N), and the presence of metastasis (M)). The NCDB began capturing surgical approach in 2010 and was categorized as open MIS. We calculated the percentage of MIS RCs performed at each facility since 2010 and dichotomized the percent-age of MIS RC performed into the top 90th percentile of facilities and <90th percentile. The top 90th percen-tile equated to ≥75% of RCs performed by MIS.
The primary outcome was receipt of CD after RC, with a focus on identification of patient and facility factors associ-ated with receiving CD. Trends in type of UD, patient, and facility characteristics were also examined comparing
2.5. Statistical analysis
Categorical variables and trends were analyzed using Pearson’s chi-squared test. ANOVA test was used to ana-lyze mean age across the study period. Variables associated with CD on univariate analysis were included in a multivar-iable logistic regression model. All P values were 2-sided with P < 0.05 considered statistically significant. Statistical analysis was performed using Stata 13 (StataCorp., College Station, TX).
3. Results
3.1. Patient-related demographics
Table 1
Patient demographics and hospital features. Statistically significant if bold
Incontinent diversion Continent diversion P
Age
Sex
Male
Female
Race
White
Non-white
Charlson comorbidity index
% of no high school graduates in patient’s zip code of residence
Insurance status
Not insured
Private/managed care
Medicaid
Medicare
Other
Median income
Facility location
Northeast
South/southeast
Midwest
West
Facility type
Nonacademic facility
Academic program
Hospital volume
County description
Metropolitan
Urban
Rural
Miles from patient’s residence to hospital
Surgical approach
Open
Minimally invasive
Pathologic stage
Surgical margin
Negative
Positive
(continued)
Table 1 (Continued)
Incontinent diversion Continent diversion P
Surgical approach ratio
Surgical approach ratio (low volume)
Surgical approach ratio (low-intermediate volume)
Surgical approach ratio (high-intermediate volume)
Surgical approach ratio (high volume)
3.2. Socioeconomic-related demographics
3.3. Hospital-related demographics
3.4. Tumor characteristics and surgical approach
Overall, MIS surgical approach was associated with a higher rate of CD compared to open RC (13.3% vs. 11.4%, P = 0.01). When individual facilities were analyzed, there was no difference in CD rates between those facilities with <75% MIS cases and those ≥75% MIS, P = 0.37. However, after stratification by facility volume, high-volume centers performing ≥75% MIS RCs were associated with fewer CD (10.2%) compared to high-volume facilities with <75% MIS cases (19.7%), P < 0.01. There was no association on subgroup analysis in low, low-intermediate, and high-inter-mediate-volume facilities. Worse pathologic tumor stage and positive surgical margins were associated with decreased rate of CD, P < 0.01. 
3.5. Variables associated with type of urinary diversion
Variables associated with CD on univariate analysis are shown in Table 2. On multivariable analysis (Table 2), age >80, CCI ≥ 1, Medicare insurance, urban county, patho-logic T3/T4 stage, and positive surgical margins were asso-ciated with decreased rates of undergoing CD. On multivariable analysis, higher income, facility located in the West, academic program, high hospital volume, and >60 miles traveled to the hospital were associated with higher rates of CD. Prior history of neoadjuvant chemother-apy was not associated with type of UD.
3.6. Trends in urinary diversion and population
4. Discussion
In our analysis of the NCDB, we report a low and declin-ing rate of continent diversion after RC for bladder cancer in the United States from 2004 to 2013. Multiple indepen-dent variables were identified as being positively and nega-tively associated with undergoing CD. The overall CD rate