www.companyofscientists.com/index.php/chd e1 Cancer Health Disparities RESEARCH Cervical cancer outcome by type of health care facilities: National Cancer Database, 2004-2015 HyounKyoung G. Park1,4, Zhixin E. Wang1,4, Chenguang Wang2, Warner K. Huh3,4, Sejong Bae1,4 1Division of Preventive Medicine, Department of Medicine, School of Medicine, University of Alabama at Birmingham, Birmingham, AL 2Division of Biostatistics and Bioinformatics, Sidney Kimmel Comprehensive Cancer Center, The Johns Hopkins University, Baltimore, MD 3Division of Gynecology Oncology, Department of Obstetrics and Gynecology, School of Medicine, University of Alabama at Birmingham, Birmingham, AL 4O’Neal Comprehensive Cancer Center, University of Alabama at Birmingham, Birmingham, AL ABSTRACT The National Cancer Database from 2004 to 2015 was analyzed to identify cervical cancer outcomes associated with demographic and clinical characteristics measured by types of facility. Chi-Square tests were used to compare proportions and logistic regression to determine factors associated with cervical cancer outcomes. Women treated at Academic/Research Programs (ARPs) were younger at diagnosis, more likely black, less educated and more in Stage 2, lived further away from treatment facilities, had less comorbidities and better 5- year survival, and were more likely to be alive at 30 and 90 days after surgery compared to other programs. Women treated at Community Cancer Programs were more likely 75 and older at diagnosis, more likely to receive radiation treatment and more in Stage 4, more living in rural areas and less than 10 miles from the facility, and had more comorbidities, and lower 5- year survival compared to other programs. Women treated at Comprehensive Community Cancer Programs were more likely white and educated, had more private insurance, and underwent surgery. Women treated at Integrated Network Cancer Programs were more likely to live in urban, south region, and in Stage 1B2, had more surgery and one comorbidity, and died fewer than 30 days after surgery. The type of facility and treatment had varied effects on mortality and 5-year survival. Considering the different cervical cancer outcomes from different health care facilities, further research is needed to identify what factors influence women to choose a health care facility for their treatment and how this choice can affect different health outcomes. KEYWORDS: Cervical cancer, type of facility, mortality, survival Citation: Park HG et al (2019) Cervical cancer outcome by type of health care facilities: National Cancer Database, 2004-2015; Cancer Health Disparities. 4: e1-e-13. doi:10.9777/chd.2019.1011 www.companyofscientists.com/index.php/chd e2 Cancer Health Disparities RESEARCH Introduction Cervical cancer is the most common type of gynecological cancer but one of the most preventable and treatable cancers (Fedewa et al., 2012; Ferlay et al., 2015). In the United States, cervical cancer incidence and mortality rates continue to decline since regular cervical cancer screening tests and vaccination were introduced (Akinlotan et al., 2017; CDC, 2018; Scarinci et al., 2010.). Several studies have investigated risk factors associated with higher cervical cancer incidence and mortality. For example, being black (Arvizo and Mahdi, 2017; Beavis et al., 2017; Furlow, 2018; Powell et al., 2018 ; Scarinci et al., 2010. ; Singh and Jemal, 2017; Weragoda et al., 2016 ; Yoo et al., 2017), getting older (Fedewa et al., 2012; Furlow, 2018; Yosta and Hoekstra, 2018), living in The South (Yoo et al., 2017), less income (Singh and Jemal, 2017), less education (Singh and Jemal, 2017), having comorbidity (Diaz et al., 2018), stages of cancer (Acharya and Grigsby, 2016; Landy et al., 2016), insurance (Acharya and Grigsby, 2016; Churilla et al., 2016 ; Davis et al., 2018; Fedewa et al., 2012) and distance (Barrington et al., 2016) have been reported as risk factors responsible for the incidence and mortality of cervical cancer. Identifying risk factors for cervical cancer is critical for developing proper prevention and treatment approaches connecting to risk factors, thus reducing racial disparities on the outcomes of cervical cancer, such as higher incidence and mortality rates, as well as lower five-year survival and overall survival among black women (American Cancer Society, 2018). There were significant differences in the characteristics between white and black women with cervical cancer in the NCDB data. For example, black women were diagnosed with later- stage disease with more comorbidities than white women, which resulted in increased morbidity and mortality in black women. In addition, most black women had treatment in academic/research programs that were close to their residence and higher mortality rates among white women were identified in community cancer programs and Comprehensive Community Cancer Programs. Considering significant racial disparities in healthcare outcomes, it is essential to understand the impact of the type of health care facility on health outcomes among cervical cancer patients. Several studies have been conducted on the relationship between outcomes in cervical cancer patients based on distance from hospitals (Barrington et al., 2016; Gunderson et al., 2013; Powell et al., 2018; Spees et al., 2018) and the impact of facility volume on quality treatment and survival (Ross et al., 2010; Showalter et al., 2016). In their study on racial disparities in outcomes after surgical procedures, Haider and associates (2013) reported systemic factors such as access to care, hospital volume, and hospital patient population have been shown to contribute to disparities (Haider et al., 2013). Study results report that high volume hospitals have more favorable outcomes than low volume hospitals. Connecting to the choice of hospitals for treatment, it is also important to understand possible factors affecting the patient’s choice of treatment options resulting in the subsequent oncologic outcome and causing significant disparities in survival among minority patients (Luo et al., 2015). For example, black patients and living in rural areas were associated with the choice of surgery to treat other cancers, resulting in less www.companyofscientists.com/index.php/chd e3 Cancer Health Disparities RESEARCH surgery compared to white patients (Cykert et al., 2010; Steenland et al., 2011). The purpose of this study was to investigate the impact of different health care facilities on types of treatments and cervical cancer outcomes using the National Cancer Database (NCDB) data. The research question was “How would different types of health care facilities influence the outcomes of patients with cervical cancer and their types of treatment?” Materials and Methods To assess cervical cancer outcomes by different facility types, cervical cancer data were drawn from the National Cancer Database (NCDB) for the years 2004 to 2015. NCDB is jointly sponsored by the American College of Surgeons and the American Cancer Society, and it is the largest clinical cancer registry in the world (Steenland et al., 2011). It covers more than 70% of newly diagnosed cancer cases in the United States (Lin et al., 2014). Each facility reporting cases to the NCDB is assigned to a category classification, i.e., the facility type, by the Commission on Cancer Accreditation program. The NCDB classified all facilities into four types: Academic Cancer Programs, Community Cancer Programs, Comprehensive Community Cancer Programs, and Integrated Network Cancer Programs (Source: http://ncdbpuf.facs.org/content /participant-use-file-facility-type). Academic Comprehensive Cancer Program or Academic/Research programs (ARPs) participate in postgraduate medical education in at least four program areas, including internal medicine and general surgery. The facility accessions more than 500 newly diagnosed cancer cases each year, and offers the full range of diagnostic and treatment services either on-site or by referral. Community Cancer Programs (CCPs) accession more than 100 but fewer than 500 newly diagnosed cancer cases each year and provides a full range of diagnostic and treatment services, but referral for a portion of diagnosis or treatment may occur. Comprehensive Community Cancer Programs (CCCPs) accession 500 or more newly diagnosed cancer cases each year and provide a full range of diagnostic and treatment services either on-site or by referral. Integrated Network Cancer Programs (INCPs) own, operate, lease, or are part of a joint venture with multiple facilities providing integrated cancer care and offer comprehensive services. We used this classification to analyze demographic characteristics, disease status, treatment and cervical cancer outcomes by the type of facility. For our data analysis, women younger than 40 years old were excluded from this study in addition to women other than white and black (Hispanic origin was not distinguished). Age was categorized into 5 groups: 40-49, 50-59, 60-69, 70-79, and 80+. Regional information was deduced from the location of the reporting facility (suppressed for cases aged 0-39). There are 6 categories for insurance in NCDB: uninsured, private insurance/managed care, Medicaid, Medicare, other government, and unknown. We further separated the Medicare group into younger Medicare (< 65 years old) and older Medicare (≥ 65 years old) because of the difference in eligibility for these two groups. Cervical cancer stages were identified based on the T, N, and M elements as defined by the American Joint Committee on Cancer (AJCC). To analyze the association between disease status and facility types, “TNM Clin Stage Group” that identifies the anatomic extent of disease based on http://ncdbpuf.facs.org/content/participant-use-file-facility-type http://ncdbpuf.facs.org/content/participant-use-file-facility-type www.companyofscientists.com/index.php/chd e4 Cancer Health Disparities RESEARCH the T, N, and M elements known prior to the start of any therapy was re-categorized into 5 groups:1A-1B1 (TNM Clin Stage Group=1,1A,1A1, 1A2,1B,1B1), 1B2 (TNM Clin Stage Group=1B2), 2 (TNM Clin Stage Group=2,2A,2A1,2A2,2B), 3 (TNM Clin Stage Group=3,3A,3B), and 4 (TNM Clin Stage Group=4,4A,4B). Comorbidity was measured by Charlson score, which is a weighted score derived from the sum of the scores for each of the comorbid conditions listed in the Charlson Comorbidity Score Mapping Table. Household income was categorized as quartiles based on equally proportioned income ranges among all US zip codes. Education was estimated by matching the zip code of the patient recorded at the time of diagnosis against files derived from the 2012 American Community Survey data, spanning years 2008-2012. This item provides a measure of the number of adults in the patient's zip code who did not graduate from high school, and is categorized as equally proportioned quartiles among all US zip codes. Residence, which represents the area-based measure of rurality and urban influence, was estimated by matching the state and county FIPS code of the patient recorded at the time of diagnosis against 2013 files published by the United States Department of Agriculture Economic Research Service. Metro counties and urban counties (as defined in the 2014 NCDB Participant Use Data File (PUF) Data Dictionary) were combined together as “Urban”, and compared with rural counties. SAS version 9.4, a software package for statistical analysis, was used to compute descriptive statistics. Chi-Square tests were used to compare the differences in demographic characteristics, disease status (cervical cancer stage at diagnosis), treatment, and cervical cancer outcomes among programs. Multivariable analysis was performed using stepwise modeling for variables associated with the outcomes of cervical cancer including vital status and 5 year survival. Odds ratio with 95% confidence intervals were calculated. Results Table 1 presents the association between demographic information and facility types. About forty-four percent of women used ARPs, followed by CCCPs (37.4%), INCPs (11.4%), and CCPs (6.8%). ARPs had more black women and younger women aged 40-59 years at diagnosis compared to CCPs and CCCPs, had less income and education (more women from the area who did not graduate from high school) compared to CCCPs and INCPs, more income compared to CCPs, more living far from hospitals, more uninsured, more lived in Midwest and Northeast compared to all other programs. CCPs had more women aged 75 and older when diagnosed compared to ARPs and INCPs, income with $38,000-47,999, education with 13-20.9% who did not graduate from high school, more living in rural areas and less than 10 miles away from the facility compared to all other programs. CCCPs had more white women compared to ARPs and INCPs, more women aged 65-74 years when diagnosed compared to ARPs, more education with less than 7% and with 7-12.9% who did not graduate from high school compared to ARPs and CCPs, more living in 21-50 miles away from the facility compared to all other programs, more private and other government insurance program www.companyofscientists.com/index.php/chd e5 Cancer Health Disparities RESEARCH compared to ARPs and CCPs, and more living in West region compared to all other programs. INCPs had more income with $48,000-$62,999 compared to ARPs and CCPs, and more living in urban, 11-20 miles from the facility and in South region compared to all other programs. Table 1. The demographic information by facility types (NCDB, 2004-2015). Type of facility Demo Facility Type Total N (%) ARPs CCPs CCCPs INCPs Facility 35719(44.43) 5447(6.78) 30050(37.38) 9169(11.41) 80385 Race* White 28096(78.66) 4670(85.74) 25911(86.23) 7328(79.92) 66005 Black 7623(21.34) 777(14.26) 4139(13.77) 1841(20.08) 14380 Age*† 40-44 6489(18.17) 914(16.78) 5088(16.93) 1636(17.84) 14127 45-49 6466(18.10) 883(16.21) 4937(16.43) 1567(17.09) 13853 50-54 5690(15.93) 818(15.02) 4416(14.70) 1406(15.33) 12330 55-59 4970(13.91) 699(12.83) 4016(13.36) 1262(13.76) 10947 60-64 3956(11.08) 609(11.18) 3257(10.84) 1026(11.19) 8848 65-69 2912(8.15) 472(8.67) 2779(9.25) 782(8.53) 6945 70-74 2026(5.67) 354(6.50) 1987(6.61) 551(6.01) 4918 75+ 3210(8.99) 698(12.81) 3570(11.88) 939(10.24) 8417 Income* <$38,000 9807(27.46) 1452(26.66) 6756(22.48) 2172(23.69) 20187 $38,000-47,999 8356(23.39) 1669(30.64) 7961(26.49) 2245(24.48) 20231 $48,000-62,999 8450(23.66) 1322(24.27) 7841(26.09) 2444(26.66) 20057 >=$63,000 8733(24.45) 942(17.29) 7095(23.61) 2183(23.81) 18953 Unknown 373(1.04) 62(1.14) 397(1.32) 125(1.36) 957 Education*†† <7% 5646(15.81) 606(11.13) 5236(17.42) 1561(17.02) 13049 7-12.9% 9094(25.46) 1530(28.09) 9099(30.28) 2752(30.01) 22475 13-20.9% 9928(27.79) 1709(31.38) 8763(29.16) 2761(30.11) 23161 >=21% 10694(29.94) 1542(28.31) 6574(21.88) 1971(21.50) 20781 Missing 357(1.00) 60(1.10) 378(1.26) 124(1.35) 919 Residence* Urban 34135(95.57) 5125(94.09) 28463(94.72) 8819(96.18) 76542 Rural 489(1.37) 195(3.58) 656(2.18) 61(0.67) 1401 Missing 1095(3.07) 127(2.33) 931(3.10) 289(3.15) 2442 Distance* <=10 15743(45.65) 3301(62.80) 14379(49.80) 4757(53.80) 38180 11-20 5578(16.18) 981(18.66) 5250(18.18) 1702(19.25) 13511 21-50 6742(19.55) 742(14.12) 5758(19.94) 1585(17.93) 14827 >50 6421(18.62) 232(4.41) 3486(12.07) 798(9.03) 10937 Insurance* Uninsured 3829(10.78) 424(7.83) 2113(7.06) 800(8.77) 7166 Medicaid 7373(20.76) 1273(23.50) 4320(14.44) 1601(17.55) 14567 Younger Medicare 1579(4.45) 312(5.76) 1343(4.49) 368(4.03) 3602 Older Medicare 6039(17.00) 1265(23.35) 6815(22.79) 1863(20.42) 15982 Private 14866(41.85) 1979(36.53) 14443(48.29) 4262(46.71) 35550 Other 301(0.85) 41(0.76) 371(1.24) 109(1.19) 822 www.companyofscientists.com/index.php/chd e6 Cancer Health Disparities RESEARCH Government Missing 1536(4.32) 123(2.27) 504(1.69) 121(1.33) 2284 Region* Midwest 9076(25.41) 1777(32.62) 6209(20.66) 1867(20.36) 18929 Northeast 9083(25.43) 900(16.52) 4534(15.09) 873(9.52) 15390 South 12833(35.93) 1928(35.40) 13429(44.69) 5421(59.12) 33611 West 4727(13.23) 842(15.46) 5878(19.56) 1008(10.99) 12455 * P<.0001 † All women whose age were less than 40 years were missing in this facility analysis (26,927, 25.1%). †† Education was measured using the number of adults in the patient's zip code who did not graduate from high school and is categorized as equally proportioned quartiles among all US zip codes. TNM Clin Stage Group presented that ARPs had more women in Stage 2, CCPs in Stage 4, and INCPs in Stage 1B2 compared to all other programs. ARPs had more women with no comorbidity while INCPs had more women with 1 comorbidity compared to all other programs (Table 2). Table 2. Disease status by facility types (NCDB, 2004-2015). Type of facility Disease status Facility Type Total N (%) ARPs CCPs CCCPs INCPs TNM Clin Stage Group* 1A-1B1 9357(32.92) 1293(30.71) 7702(33.67) 2324(33.71) 20676 1B2 1516(5.33) 176(4.18) 1186(5.18) 405(5.87) 3283 2 5806(20.42) 784(18.62) 4270(18.67) 1313(19.04) 12173 3 6885(24.22) 989(23.49) 5456(23.85) 1692(24.54) 15022 4 4862(17.10) 968(22.99) 4262(18.63) 1161(16.84) 11253 Comorbidity* 0 29874(84.61) 4507(83.66) 24861(83.59) 7434(82.19) 66676 1 4521(12.81) 722(13.40) 4031(13.55) 1361(15.05) 10635 2 911(2.58) 158(2.93) 849(2.85) 250(2.76) 2168 * P<.0001 The association between treatment types and race by the types of facility was analyzed. Black women had more surgery, radiation, and chemotherapy in ARPs compared to all other programs, while white women had more surgery and chemo in CCCPs compared to ARPs and INCPs and more radiation in CCPs compared to ARPs and INCPs. Women who did not have surgery were about five times as likely to die of cervical cancer (Odds ratio 4.90, 95%CI: 4.74-5.06), while women who did not have radiation or chemotherapy were less likely to die of cervical cancer (Odds ratio 0.84, 95%CI: 0.81-0.87; Odd ratio 0.86, 95%CI:0.82-0.89) (data not shown). INCPs had more women with surgery compared to all other programs. Table 3. Treatment status by race and facility types (NCDB, 2004-2015). www.companyofscientists.com/index.php/chd e7 Cancer Health Disparities RESEARCH Type of facility Treatment Facility Type Total N (%) ARPs CCPs CCCPs INCPs Surgery* Yes 16432(46.10) 2293(42.40) 14818(49.51) 4682(51.22) White 13597(82.75) 2013(87.79) 13126(88.58) 3910(83.51) 32646 Black 2835(17.25) 280(12.21) 1692(11.42) 772(16.49) 5579 No 19210(53.90) 3115(57.60) 15113(50.49) 4459(48.78) 41897 Chemo* Yes 22110(62.43) 3328(62.03) 18034(60.73) 5641(62.13) 49113 White 17165(77.63) 2836(85.22) 15487(85.88) 4479(79.40) 39967 Black 4945(22.37) 492(14.78) 2547(14.12) 1162(20.60) 9146 No 13308(37.57) 2037(37.97) 11663(39.27) 3439(37.87) 30447 Radiation* Yes 22470(63.31) 3425(63.50) 18547(62.26) 5612(61.51) 50054 White 17409(77.48) 2940(85.84) 15832(85.36) 4453(79.35) 40634 Black 5061(22.52) 485(14.16) 2715(14.64) 1159(20.65) 9420 No 13024(36.69) 1969(36.50) 11242(37.74) 3511(38.49) 29746 * P<.0001 Five outcomes were analyzed to identify their association with facility types. ARPs had more women alive at 30 and 90 days after surgery compared to CCPs, while INCPs had more cases died less than 30 days after surgery compared to ARPs and CCCPs. CCPS had more cases that died fewer than 90 days after surgery, more cases with no surgery or alive cases that had fewer than 30 and 90 days of follow up and more deceased cases compared to all other programs. CCCPs had more cases with no surgery or alive cases that had fewer than 30 days of follow up compared to all other programs. Regarding the outcome of readmission within 30 days of surgical discharge, overall 94.6% was either no surgical procedure of the primary site was performed, or patient was not readmitted. INCPs had more cases with unplanned readmission within 30 days of discharge compared to all other programs. ARPs presented more five-year survival compared to CCPs and less comorbidity compared to INCPs, while CCPs had less five year survival compared to all other programs. Regarding the outcome of vital status (mortality), ARPs had more cases who were alive compared to CCPs and CCCPs, while CCPs had more cases who were deceased compared to all other programs (Table 4). Table 4. Survival outcomes by facility types (NCDB, 2004-2015). Type of facility Outcomes Facility Type Total N (%) ARPs CCPs CCCPs INCPs 30 Day Mortality* Alive at 30 days after surgery Died < 30 days after surgery No surgery, or alive cases have < 30 days of follow up 14681(98.38) 1988(96.18) 13102(97.81) 4127(98.10) 33898 56(0.38) 11(0.53) 56(0.42) 25(0.59) 148 186(1.25) 68(3.29) 237(1.77) 55(1.31) 546 90 Day Mortality* Alive at 90 days after surgery Died < 90 days after surgery No surgery, or alive cases have < 90 days of follow up 14461(96.90) 1949(94.29) 12902(96.32) 4052(96.32) 33364 168(1.13) 34(1.64) 173(1.29) 62(1.47) 437 294(1.97) 84(4.06) 320(2.39) 93(2.21) 791 www.companyofscientists.com/index.php/chd e8 Cancer Health Disparities RESEARCH Readmission Within 30 Days of Surgical Discharge* No surgical procedure Unplanned readmission Planned readmission Planned and unplanned readmission Unknown 33713(94.38) 5111(93.83) 28519(94.91) 8675(94.61) 76018 690(1.93) 87(1.60) 454(1.51) 190(2.07) 1421 417(1.17) 66(1.21) 278(0.93) 102 (1.11) 863 33(0.09) 2(0.04) 23 (0.08) 5(0.05) 63 866(2.42) 181(3.32) 776(2.58) 197(2.15) 2020 5 Year survival* Yes No 10895(33.27) 1471(29.53) 8918(32.49) 2684(32.28) 23968 21855(66.73) 3511(70.47) 18531(67.51) 5630(67.72) 49527 Vital status (Mortality) * Dead Alive 12456(38.03) 2225(44.63) 11480(41.82) 3219(38.72) 29380 20299(61.97) 2760(55.37) 15971(58.18) 5095(61.28) 44125 * P<.0001 Table 5 presents the result of multivariate logistic regression on vital status (mortality) and 5-year survival in cervical cancer. There were slight differences in predicting vital status and 5-year survival. Black women, insurances other than private insurance, less income, more education, more comorbidity, more advanced stages, older women, treatment in CCPS and CCCPs, and not having surgery and chemo were more likely to die of cervical cancer compared to the reference groups. Black women, insurances other than private insurance, more education, more comorbidity, more advanced stages, living far from facilities, older women, treatment in CCCPs and INCPs, and not having surgery and radiation were less likely to survive for five years compared to the reference groups. Although distance in predicting vital status and income in predicting 5-year mortality were not significant from stepwise, we added them to the model to adjust confounding effects. Table 5. Predictors of the outcomes of cervical cancer (NCDB, 2004-2015). Vital status (Mortality) 5 year survival Adjusted Odds Ratio 95% confidence interval Adjusted Odds Ratio 95% confidence interval Race White Black 1.00 1.10 - 1.04-1.16 1.00 0.94 - 0.89-0.99 Region South Midwest 1.00 1.03 - 0.98-1.09 1.00 1.03 - 0.97-1.08 Northeast 0.94 0.88-1.00 1.01 0.95-1.07 West 0.94 0.89-1.01 1.10 1.04-1.17 Insurance Private Uninsured 1.00 1.03 - 0.96-1.11 1.00 0.80 - 0.74-0.86 Medicaid 1.24 1.17-1.31 0.75 0.71-0.79 Younger Medicare 1.49 1.35-1.64 0.70 0.63-0.78 Older Medicare 1.54 1.40-1.70 0.84 0.75-0.93 Other Government 0.85 0.69-1.05 0.85 0.70-1.03 Missing 1.05 0.94-1.19 0.92 0.82-1.03 www.companyofscientists.com/index.php/chd e9 Cancer Health Disparities RESEARCH Income <$38,000 $38,000-47,999 1.00 0.90 - 0.84-0.95 1.00 1.03 - 0.97-1.09 $48,000-62,999 0.86 0.80-0.92 1.03 0.96-1.10 >=$63,000 0.78 0.72-0.84 1.05 0.96-1.13 Unknown 1.02 0.40-2.58 1.03 0.41-2.62 Education† >=21% <7% 1.00 1.12 - 1.03-1.23 1.00 0.94 - 0.86-1.02 7-12.9% 1.16 1.08-1.24 0.93 0.87-0.99 13-20.9% 1.13 1.07-1.20 0.98 0.92-1.03 Missing 2.60 1.01-6.68 0.30 0.11-0.78 Comorbidity 0 1 1.00 1.37 - 1.29-1.45 1.00 0.79 - 0.74-0.84 2 1.98 1.75-2.24 0.60 0.52-0.69 Stage 1A-1B1 1B2 1.00 1.34 - 1.21-1.49 1.00 0.64 - 0.59-0.71 2 1.84 1.72-1.97 0.76 0.71-0.81 3 3.62 3.38-3.87 0.42 0.39-0.45 4 11.52 10.68-12.43 0.12 0.10-0.13 Distance(miles) <=10 11-20 1.00 0.99 - 0.94-1.05 1.00 0.94 - 0.89-0.99 21-50 1.00 0.95-1.06 0.91 0.86-0.96 >50 1.05 0.98-1.11 0.82 0.77-0.87 Age 40-44 45-49 1.00 1.20 - 1.12-1.29 1.00 0.99 - 0.93-1.06 50-54 1.27 1.18-1.36 0.95 0.89-1.02 55-59 1.38 1.28-1.49 0.88 0.82-0.95 60-64 1.52 1.41-1.65 0.85 0.79-0.92 65-69 1.51 1.36-1.69 0.84 0.75-0.94 70-74 1.91 1.69-2.15 0.71 0.63-0.81 75+ 3.53 3.15-3.96 0.46 0.40-0.51 Facility Academic/Research Program Community Cancer Program 1.00 1.16 - 1.07-1.27 1.00 0.93 - 0.86-1.02 Comprehensive Community Cancer Program 1.19 1.14-1.25 0.94 0.90-0.99 Integrated Network Cancer Program 1.04 0.97-1.11 0.87 0.81-0.93 Surgery Yes No 1.00 1.90 - 1.80-2.01 1.00 0.77 - 0.73-0.82 Radiation Yes 1.00 - 1.00 - www.companyofscientists.com/index.php/chd e10 Cancer Health Disparities RESEARCH No 0.90 0.85-0.96 0.70 0.66-0.75 Chemo Yes No 1.00 1.16 - 1.09-1.22 1.00 1.09 - 1.03-1.15 † Education was measured using the number of adults in the patient's zip code who did not graduate from high school and is categorized as equally proportioned quartiles among all US zip codes. Discussion To prevent and treat cervical cancer effectively, it is critical to understand risk factors and how they impact the disease. While assessing the differences in risk factors and outcomes of cervical cancer between races in our previous study, we found there were differences in using health care facilities between black and white women. Therefore, we further analyzed the NCDB data to identify differences in demographics, disease status, treatment status and health outcomes by facility type. ARPs had more black women, women younger at diagnosis with less education, more living far from treatment facilities, more in Stage 2, less comorbidity and more 5-year survival, and more cases alive at 30 and 90 days after surgery compared to other programs. CCPs had more women aged 75 and older at diagnosis, living in rural areas and fewer than 10 miles from hospitals, more Stage 4 and comorbidity, less 5 year survival, and more radiation compared to other programs. CCCPs had more white women, more education and more private insurance, and more surgery. INCPs had more women living in urban and south region, more Stage 1B2 and more surgery, more 1 comorbidity, and more women who died less than 30 days after surgery. Previous studies reported having treatment at high-volume hospitals was associated with lower mortality rates and better survival (Haider et al., 2013; Lin et al., 2014). However, when assessing the impact of the type and volume of facility on health care outcome, it is critical to consider patients’ disease status. ARPs that would be classified as high volume centers presented better outcomes in being alive at 30 and 90 days after surgery, 5-year survival and, vital status, compared to CCPs that would be classified as low volume centers. However, it was noted that CCPs had more women diagnosed at Stage 4 while ARPs had more women diagnosed at Stage 2, so the result should be interpreted carefully. It is also interesting that younger age groups had more treatment in ARPs while older age groups had more treatment in CCPs and CCCPs. That may explain the possible association with a better survival outcome in cervical cancer in ARPs, which is consistent with prior works (Fedewa et al., 2012; Furlow, 2018; Yosta and Hoekstra, 2018). From multivariate analysis results, under-insured women were more likely to die of cervical cancer and, uninsured and under-insured women were less likely to survive for five years compared to privately insured women. It is interesting that the uninsured women were not more likely to die of cervical cancer compared to private patients. Although women without health insurance are less likely to receive cervical cancer treatment and survive (Acharya and Grigsby, 2016; Fedewa et al., 2012), our analysis shows over 53% of uninsured patients were treated in ARPs and had better health outcomes in 30 day and 90 day mortality www.companyofscientists.com/index.php/chd e11 Cancer Health Disparities RESEARCH and vital status, which proves promising in developing cervical cancer centers that may improve healthcare access and provide a high quality of care to those with under- or no insurance. However, when the population was stratified by race within ARPs, black women had more deaths less than 30 days and 90 days after surgery (respectively p<.05) and had more deaths (Vital status: p<.0001) compared to white women. Further analyses showed black women in ARPs were older, less income, more living in urban area and close to treatment facilities, more uninsured and underinsured, more advanced stages (Stage 3 & 4), more comorbidities, less surgery, and more chemo and radiation therapy compared to white women (data not shown). Considering the proportion of black women (21%) and still overall better outcomes for women with cervical cancer in ARPs, it addresses clear cervical cancer outcome disparities between black and white women (Arvizo and Mahdi, 2017; Dalton & Farley, 2017; Fleming and et al., 2014). Therefore, it is needed to assess the effectiveness of cervical cancer treatment/programs according to the facility type, especially, to explore how to manage patients without insurance and with less income and how to reduce outcome disparities by targeting barriers identified. As another interesting result, we found there was an interaction between education and income (Data not shown) that women from higher education areas had less survival, which is contradicting to previous study results (Franceschi et al., 2009; Singh and Jemal, 2017). This might be associated with the fact that education was measured using the number of adults in the patient's zip code who did not graduate from high school, not based on an individual’s educational attainment. Therefore, further study is necessary to investigate how the level of education and income are related to cervical cancer risk and mortality. Our study has some limitations especially in assessing the outcomes of cervical cancer in each facility. We compared the association between cervical cancer outcomes and facility type based on responses from a large national sample of cancer patients, so our results should not be interpreted as actual evaluation of cervical cancer treatment in each facility classified by NCDB. Also, considering a disproportionate number of women treated at each facility, for example, 44.4% patients at ARPs compared with about 6.8% at CCPs, the generalizability of this study may be limited. More importantly, a significant number of patients had missing data on facility at each stage of their diagnosis (e.g., 34.99% did not have facility information in Stage 1A-1B1). As another limitation, our analysis was limited by the variables collected in the NCDB. For example, there was no data to assess factors contributing to choosing a facility and treatment, so the reason for racial differences in using different facilities could not be explored. Considering a complex situation that patients with cervical cancer may confront, it is imperative to understand what factors contribute to choosing their treatment option including health care facility. For example, possible reasons resulting from limited access to high-volume/quality hospitals, such as geographic location and under- or no insurance among minorities leading to outcome disparities, should be considered and reflected when developing cervical cancer centers and www.companyofscientists.com/index.php/chd e12 Cancer Health Disparities RESEARCH programs (Beavis et al., 2017; Haider et al., 2013; Singh and Jemal, 2017). Our study results include some contradicting outcomes, which may encourage further research on understanding treatment choices and outcomes better. This will help identify critical factors that make cancer care programs successful and effective. Acknowledgements This article was partially supported by the National Cancer Institute (P30CA013148, P50CA098252, 2U54CA118948). The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the funding agency, NIH. Conflict of interest The authors declare that there are no conflicts of interest or financial disclosures. Authors’ contributions Conceptualization: HP SB CW Formal analysis: HP ZW SB Methodology: HP SB Supervision: SB WH Writing ± original draft: HP SB Writing ± review & editing: HP ZW CW WH SB REFERENCES Acharya, S and Grigsby, P.W., Access to Health Care and Disparities in Cervical Cancer Diagnosis, Treatment, and Survival. International Journal of Radiation Oncology, 2016. 96(2): p. e290. Akinlotan M, B.J., Helduser J, Ojinnaka C, Lichorad A, and McClellan D, Cervical Cancer Screening Barriers and Risk Factor Knowledge Among Uninsured Women. J Community Health, 2017. 42(4): p. 770–778. American Cancer Society, Cancer Facts & Figures for African Americans 2016-2018. Atlanta: American Cancer Society, 2018. Arvizo, C., and Mahdi, H., Disparities in cervical cancer in African American women: What primary care physicians can do. Cleve Clin J Med, 2017. 84(10): p. 788-794. Barrington, D.A., Dilley, S.E., Landers, E.E., Thomas, E.D., Boone, J.D., Straughn, J.M. Jr, McGwin, G. Jr, and Leath, C.A., Distance from a Comprehensive Cancer Center: A proxy for poor cervical cancer outcomes? Gynecol Oncol., 2016. 143(3): p. 617-621. Beavis, A.L., Gravitt, P.E. and Rositch, A.F., Hysterectomy- corrected cervical cancer mortality rates reveal a larger racial disparity in the United States. . Cancer, 2017. 123(6): p. 1044-1050. CDC, HPV Vaccine. Accessed on Dec 20, 2018 https://www.cdc.gov/cancer/knowledge/provider- education/hpv/hpv-vaccine.htm, 2018. Churilla, T., Egleston, B., Dong, Y., Shaikh, T., Murphy, C., Mantia-Smaldone, G., Chu, C., Rubin, S., and Anderson, P., Disparities in the management and outcome of cervical cancer in the United States according to health insurance status. Gynecol Oncol, 2016 141(3): p. 516-523. Cykert, S., Dilworth-Anderson, P., Monroe, M.H., Walker, P., McGuire, F.R., Corbie-Smith, G., Edwards, L.J., and Bunton, A.J., Factors associated with decisions to undergo surgery among patients with newly diagnosed early-stage lung cancer. JAMA, 2010 303(23): p. 2368-76. Dalton, H.J., and Farley, J.H., Racial disparities in cervical cancer: Worse than we thought. Cancer, 2017. 123(6): p. 915-916. Davis, M., Strickland, K., Easter, S.R., Worley, M. Jr, Feltmate, C., Muto, M., Horowitz, N., Berkowitz, R., and Feldman, S., The impact of health insurance status on the stage of cervical cancer diagnosis at a tertiary care center in Massachusetts. Gynecol Oncol., 2018. 150(1): p. 67-72. Diaz, A., Baade, P.D., Valery, P.C., Whop, L.J., Moore, S.P., Cunningham, J., Garvey, G., Brotherton, J.M.L., O'Connell, D.L., Canfell, K., Sarfati, D., Roder, D., Buckley, E., and Condon, J.R., Comorbidity and cervical cancer survival of Indigenous and non-Indigenous Australian women: A semi-national registry-based cohort study (2003-2012). PLoS One, 2018. 13(5). Fedewa, S.A., Cokkinides, V., Virgo, K.S., Bandi, P., Saslow, D., and Ward, E.M., Association of Insurance Status and Age With Cervical Cancer Stage at Diagnosis. AJPH, 2012. Ferlay, J., Soerjomataram, I., Dikshit, R., Eser, S., Mathers, C., Rebelo, M., and Bray, F., Cancer incidence and mortality worldwide: Sources, methods and major patterns in GLOBOCAN 2012. International Journal of Cancer, 2015. 136. Fleming, S., Schluterman, N.H., Tracy, J.K., and Temkin, S.M., Black and white women in Maryland receive different treatment for cervical cancer. PLoS One. 9(8): p. e104344. www.companyofscientists.com/index.php/chd e13 Cancer Health Disparities RESEARCH Franceschi, S., Plummer, M., Clifford, G., De Sanjose, S., Bosch, X., Herrero, R., Muñoz, N. and Vaccarella, S., Differences in the risk of cervical cancer and human papillomavirus infection by education level. British Journal of Cancer, 2009. 101(5): p. 865. Furlow, B., Nearly One-in-Five Women With Cervical Cancer are Diagnosed After Age 65. 2018. Gunderson, C.C., N.E., McMeekin DS, and Moore KN., Distance traveled for treatment of cervical cancer: who travels the farthest, and does it impact outcome? Int J Gynecol Cancer, 2013 Haider, A.H., Scott, V.K., Rehman, K.A., Velopulos, C., Bentley, J.M., Cornwell, E.E. and Al-Refaie, W., Racial disparities in surgical care and outcomes in the United States: a comprehensive review of patient, provider, and systemic factors. J Am Coll Surg, 2013. 216(3): p. 482-92. Landy, R., Pesola, F., Castañón, A., and Sasieni P., Impact of cervical screening on cervical cancer mortality: estimation using stage-specific results from a nested case–control study. Br J Cancer, 2016. 115(9): p. 1140-1146. Lin. J.F., B., J.L., Krivak, T.C., Beriwal, S., Chan, J.K., Sukumvanich, P., Monk, B.J., and Richard, S.D., Impact of facility volume on therapy and survival for locally advanced cervical cancer. Gynecol Oncol, 2014 132(2): p. 416-22. Luo, T., Spolverato, G., Johnston, F., Haider, A.H., and Pawlik, T.M., Factors That Determine Cancer Treatment Choice Among Minority Groups. J Oncol Pract, 2015 11(3): p. 259-61. Powell, T.C., Dilley, S.E., Bae, S., Straughn, J.M.Jr,, Kim, K.H., and Leath, C.A., The Impact of Racial, Geographic, and Socioeconomic Risk Factors on the Development of Advanced-Stage Cervical Cancer. J Low Genit Tract Dis, 2018 22(4): p. 269-273. Ross, J.S., Normand, S.L., Wang, Y., Ko, D.T., Chen, J., Drye, E.E., Keenan, P.S., Lichtman, J.H., Bueno, H., Schreiner, G.C., and Krumholz, H.M., Hospital Volume and 30-Day Mortality for Three Common Medical Conditions. N Engl J Med, 2010. 362(12): p. 1110-8. Scarinci IC, G.F., Kobetz E, Partridge EE, Brandt HM, Bell MC, Dignan M, Ma GX, Daye JL, and Castle PE., Cervical cancer prevention: new tools and old barriers. Cancer, 2010. . 116(11): p. 2531-42. Showalter, T.N., Camacho, F., Cantrell, L.A., and Anderson, R.T., Determinants of Quality Care and Mortality for Patients With Locally Advanced Cervical Cancer in Virginia. Medicine, 2016. 95(8): p. e2913. Singh, G.K., and Jemal, A., Socioeconomic and Racial/Ethnic Disparities in Cancer Mortality, Incidence, and Survival in the United States, 1950-2014: Over Six Decades of Changing Patterns and Widening Inequalities. J Environ Public Health, 2017. 2017. Spees, L.P., Wheeler, S.B., Varia, M., Weinberger, M., Baggett, C.D., Zhou, X., Petermann, V.M., and Brewster, W.R., Evaluating the urban-rural paradox: The complicated relationship between distance and the receipt of guideline-concordant care among cervical cancer patients. Gynecol Oncol, 2018 12(18): p. 31407-0. Steenland, K., Goodman, M., Liff, J., Diiorio, C., Butler, S., Roberts, P., Smith, J.L., Ekwueme, D. and Hall, I.J., The effect of race and rural residence on prostate cancer treatment choice among men in Georgia. Urology, 2011 77(3): p. 581-7. Weragoda, J., Azuero, A., Badiga, S., Bell, W.C., Matthews, R., and Piyathilake, C., An examination of racial differences in 5-year survival of cervical cancer among African American and white American women in the southeastern US from 1985 to 2010. Cancer Med, 2016 5(8): p. 2126-35. Yoo, W., et al., Recent trends in racial and regional disparities in cervical cancer incidence and mortality in United States. PLoS One, 2017. 12(2): p. e0172548. Yosta and Hoekstra, A., Cervical cancer in women over 65: An analysis of screening. Gynecol Oncol Rep, 2018. 25: p. 48–51.