Geographic Distribution and Survival Outcomes for Rural Patients With Cancer Treated in Clinical Trials | Breast Cancer | JAMA Network Open | JAMA Network
[Skip to Content]
[Skip to Content Landing]
Figure 1.  Map Showing 36 995 SWOG Enrollments From 1986 to 2012 by Rural vs Urban County Origin
Map Showing 36 995 SWOG Enrollments From 1986 to 2012 by Rural vs Urban County Origin

The percentage of total SWOG and US cancer population cases by region are shown in the table, along with the estimated proportion in rural areas for each region.

Figure 2.  Forest Plot Showing the Association of Rural Residence and Survival Outcomes From Cox Regression Analyses
Forest Plot Showing the Association of Rural Residence and Survival Outcomes From Cox Regression Analyses

Results are grouped by adjuvant vs advanced disease and ordered in ascending order of the overall survival hazard ratio (HR). Each horizontal bar represents the 95% confidence interval for the associated HR (box). Hazard ratios to the left of the line of equal hazard indicate better survival for rural patients, and HRs to the right of line of equal hazard indicate worse survival for rural patients. ER indicates estrogen receptor; PR, progesterone receptor.

Figure 3.  Association of Rural Residency and Overall Survival for Different Cut Points of Rural-Urban Continuum Codes
Association of Rural Residency and Overall Survival for Different Cut Points of Rural-Urban Continuum Codes

The z scores (each represented by a circle) reflect the strength and direction of the association of residence and overall survival for each combination of Rural-Urban Continuum Code cut point (8 different cut points) and follow-up time (1-10 years) across the panel of 17 cancer cohorts. Negative z scores reflect better outcomes for rural patients; positive z scores reflect worse outcomes for rural patients. The dark gray line connects the mean values for each cut point. The blue horizontal lines show 2-tailed critical α levels for P = .05 and P = .01 (representing an informal adjustment for multiple comparisons), and the orange horizontal lines show 2-tailed critical α levels for P = .006 (representing a Bonferroni adjustment for multiple [n = 8] comparisons).

Table 1.  Study Descriptions
Study Descriptions
Table 2.  Patient Characteristics
Patient Characteristics
1.
US Census Bureau. Measuring America: our changing landscape. https://www.census.gov/content/dam/Census/library/visualizations/2016/comm/acs-rural-urban.pdf. Published December 8, 2016. Accessed May 17, 2018.
2.
Henley  SJ, Anderson  RN, Thomas  CC, Massetti  GM, Peaker  B, Richardson  LC.  Invasive cancer incidence, 2004-2013, and deaths, 2006-2015, in nonmetropolitan and metropolitan counties—United States.  MMWR Surveill Summ. 2017;66(14):1-13. doi:10.15585/mmwr.ss6614a1PubMedGoogle ScholarCrossref
3.
Iglehart  JK.  The challenging quest to improve rural health care.  N Engl J Med. 2018;378(5):473-479. doi:10.1056/NEJMhpr1707176PubMedGoogle ScholarCrossref
4.
National Cancer Institute. National Cancer Advisory Board Cancer Moonshot Blue Ribbon Panel Report. https://www.cancer.gov/research/key-initiatives/moonshot-cancer-initiative/blue-ribbon-panel. Published October 17, 2016. Accessed January 26, 2018.
5.
Berger  ML, Mamdani  M, Atkins  D, Johnson  ML.  Good research practices for comparative effectiveness research: defining, reporting and interpreting nonrandomized studies of treatment effects using secondary data sources: the ISPOR Good Research Practices for Retrospective Database Analysis Task Force Report—Part I.  Value Health. 2009;12(8):1044-1052. doi:10.1111/j.1524-4733.2009.00600.xPubMedGoogle ScholarCrossref
6.
US Department of Agriculture. Rural-Urban Continuum Codes. https://www.ers.usda.gov/data-products/rural-urban-continuum-codes/.aspx. Updated October 12, 2016. Accessed May 17, 2018.
7.
Office of Policy Development and Research; US Department of Housing and Urban Development. HUD USPS Zip Code Crosswalk Files. https://www.huduser.gov/portal/datasets/usps_crosswalk.html. Accessed May 17, 2018.
8.
Aboagye  JK, Kaiser  HE, Hayanga  AJ.  Rural-urban differences in access to specialist providers of colorectal cancer care in the United States: a physician workforce issue.  JAMA Surg. 2014;149(6):537-543. doi:10.1001/jamasurg.2013.5062PubMedGoogle ScholarCrossref
9.
Singh  GK.  Rural-urban trends and patterns in cervical cancer mortality, incidence, stage, and survival in the United States, 1950-2008.  J Community Health. 2012;37(1):217-223. doi:10.1007/s10900-011-9439-6PubMedGoogle ScholarCrossref
10.
Weaver  KE, Palmer  N, Lu  L, Case  LD, Geiger  AM.  Rural-urban differences in health behaviors and implications for health status among US cancer survivors.  Cancer Causes Control. 2013;24(8):1481-1490. doi:10.1007/s10552-013-0225-xPubMedGoogle ScholarCrossref
11.
Kaplan  EL, Meier  P.  Nonparametric estimation from incomplete observations.  J Am Stat Assoc. 1958;53:457-481. doi:10.1080/01621459.1958.10501452Google ScholarCrossref
12.
Cox  DR.  Regression models and life tables.  J Royal Stat Soc. 1972;34(2):187-220. doi:10.1007/978-1-4612-4380-9_37Google Scholar
13.
Greenlee  H, Unger  JM, LeBlanc  M, Ramsey  S, Hershman  DL.  Association between body mass index and cancer survival in a pooled analysis of 22 clinical trials.  Cancer Epidemiol Biomarkers Prev. 2017;26(1):21-29. doi:10.1158/1055-9965.EPI-15-1336PubMedGoogle ScholarCrossref
14.
Rimsza  LM, Unger  JM, Tome  ME, Leblanc  ML.  A strategy for full interrogation of prognostic gene expression patterns: exploring the biology of diffuse large B cell lymphoma.  PLoS One. 2011;6(8):e22267. doi:10.1371/journal.pone.0022267PubMedGoogle ScholarCrossref
15.
Olson  RA, Nichol  A, Caron  NR,  et al.  Effect of community population size on breast cancer screening, stage distribution, treatment use and outcomes.  Can J Public Health. 2012;103(1):46-52.PubMedGoogle Scholar
16.
Onega  T, Duell  EJ, Shi  X, Wang  D, Demidenko  E, Goodman  D.  Geographic access to cancer care in the U.S.  Cancer. 2008;112(4):909-918. doi:10.1002/cncr.23229PubMedGoogle ScholarCrossref
17.
Williams  F, Thompson  E.  Disparity in breast cancer late stage at diagnosis in Missouri: does rural versus urban residence matter?  J Racial Ethn Health Disparities. 2016;3(2):233-239. doi:10.1007/s40615-015-0132-9PubMedGoogle ScholarCrossref
18.
Meilleur  A, Subramanian  SV, Plascak  JJ, Fisher  JL, Paskett  ED, Lamont  EB.  Rural residence and cancer outcomes in the United States: issues and challenges.  Cancer Epidemiol Biomarkers Prev. 2013;22(10):1657-1667. doi:10.1158/1055-9965.EPI-13-0404PubMedGoogle ScholarCrossref
19.
National Center for Health Statistics.  Health, United States, 2015: With Special Feature on Racial and Ethnic Health Disparities. Hyattsville, MD: National Center for Health Statistics; 2016.
20.
Kirkwood  MK, Bruinooge  SS, Goldstein  MA, Bajorin  DF, Kosty  MP.  Enhancing the American Society of Clinical Oncology workforce information system with geographic distribution of oncologists and comparison of data sources for the number of practicing oncologists.  J Oncol Pract. 2014;10(1):32-38. doi:10.1200/JOP.2013.001311PubMedGoogle ScholarCrossref
21.
Baldwin  LM, Cai  Y, Larson  EH,  et al.  Access to cancer services for rural colorectal cancer patients.  J Rural Health. 2008;24(4):390-399. doi:10.1111/j.1748-0361.2008.00186.xPubMedGoogle ScholarCrossref
22.
Meden  T, St John-Larkin  C, Hermes  D, Sommerschield  S.  Relationship between travel distance and utilization of breast cancer treatment in rural northern Michigan.  JAMA. 2002;287(1):111. doi:10.1001/jama.287.1.111-JMS0102-5-1PubMedGoogle ScholarCrossref
23.
Lin  CC, Bruinooge  SS, Kirkwood  MK,  et al.  Association between geographic access to cancer care, insurance, and receipt of chemotherapy: geographic distribution of oncologists and travel distance.  J Clin Oncol. 2015;33(28):3177-3185. doi:10.1200/JCO.2015.61.1558PubMedGoogle ScholarCrossref
24.
Mitchell  KJ, Fritschi  L, Reid  A,  et al.  Rural-urban differences in the presentation, management and survival of breast cancer in Western Australia.  Breast. 2006;15(6):769-776. doi:10.1016/j.breast.2006.04.001PubMedGoogle ScholarCrossref
25.
Sabesan  S, Burgher  B, Buettner  P,  et al.  Attitudes, knowledge and barriers to participation in cancer clinical trials among rural and remote patients.  Asia Pac J Clin Oncol. 2011;7(1):27-33. doi:10.1111/j.1743-7563.2010.01342.xPubMedGoogle ScholarCrossref
26.
Yu  XQ, O’Connell  DL, Gibberd  RW, Armstrong  BK.  Assessing the impact of socio-economic status on cancer survival in New South Wales, Australia 1996-2001.  Cancer Causes Control. 2008;19(10):1383-1390. doi:10.1007/s10552-008-9210-1PubMedGoogle ScholarCrossref
27.
Meit  M, Knudson  A, Gilbert  T,  et al. The 2014 Update of the Rural-Urban Chartbook. Bethesda, MD: Rural Health Reform Policy Research Center; 2014. https://ruralhealth.und.edu/projects/health-reform-policy-research-center/pdf/2014-rural-urban-chartbook-update.pdf. Accessed May 17, 2018.
28.
Ward  E, Halpern  M, Schrag  N,  et al.  Association of insurance with cancer care utilization and outcomes.  CA Cancer J Clin. 2008;58(1):9-31. doi:10.3322/CA.2007.0011PubMedGoogle ScholarCrossref
29.
Bettencourt  BA, Schlegel  RJ, Talley  AE, Molix  LA.  The breast cancer experience of rural women: a literature review.  Psychooncology. 2007;16(10):875-887. doi:10.1002/pon.1235PubMedGoogle ScholarCrossref
30.
Fiedler  M, Aaron  HJ, Adler  L, Ginsburg  PB.  Moving in the wrong direction—health care under the AHCA.  N Engl J Med. 2017;376(25):2405-2407. doi:10.1056/NEJMp1706848PubMedGoogle ScholarCrossref
31.
Underhill  CR, Goldstein  D, Grogan  PB.  Inequity in rural cancer survival in Australia is not an insurmountable problem.  Med J Aust. 2006;185(9):479-480.PubMedGoogle Scholar
32.
Campbell  NC, Ritchie  LD, Cassidy  J, Little  J.  Systematic review of cancer treatment programmes in remote and rural areas.  Br J Cancer. 1999;80(8):1275-1280. doi:10.1038/sj.bjc.6690498PubMedGoogle ScholarCrossref
33.
Sabesan  S, Larkins  S, Evans  R,  et al.  Telemedicine for rural cancer care in North Queensland: bringing cancer care home.  Aust J Rural Health. 2012;20(5):259-264. doi:10.1111/j.1440-1584.2012.01299.xPubMedGoogle ScholarCrossref
34.
National Cancer Institute. Community Oncology Research Program. https://ncorp.cancer.gov/about/. Accessed May 17, 2018.
35.
Unger  JM, Barlow  WE, Martin  DP,  et al.  Comparison of survival outcomes among cancer patients treated in and out of clinical trials.  J Natl Cancer Inst. 2014;106(3):dju002. doi:10.1093/jnci/dju002PubMedGoogle ScholarCrossref
36.
Lamont  EB, Hayreh  D, Pickett  KE,  et al.  Is patient travel distance associated with survival on phase II clinical trials in oncology?  J Natl Cancer Inst. 2003;95(18):1370-1375. doi:10.1093/jnci/djg035PubMedGoogle ScholarCrossref
37.
Chavez-MacGregor  M, Clarke  CA, Lichtensztajn  DY, Giordano  SH.  Delayed initiation of adjuvant chemotherapy among patients with breast cancer.  JAMA Oncol. 2016;2(3):322-329. doi:10.1001/jamaoncol.2015.3856PubMedGoogle ScholarCrossref
38.
Smedley  BD, Stith  AY, Nelson  AR, eds.  Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Washington, DC: The National Academies Press; 2003:118-143.
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    Health Policy
    August 17, 2018

    临床试验中农村癌症患者治疗的地理分布和生存结果

    Author Affiliations
    • 1SWOG Statistics and Data Management Center, Seattle, Washington
    • 2Public Health Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, Washington
    • 3Sweetwater Regional Cancer Center, Memorial Hospital of Sweetwater County, Rock Springs, Wyoming
    • 4Department of Health Services Research, The University of Texas MD Anderson Cancer Center, Houston
    • 5Columbia University Medical Center, New York, New York
    JAMA Netw Open. 2018;1(4):e181235. doi:10.1001/jamanetworkopen.2018.1235
    关键点 español English

    问题  在临床试验中接受类似治疗的农村和城市癌症患者是否具有类似的结果?

    结果  在这项疗效比较研究中,来自所有50个州的36995例患者,自1986年至2012年接受了44次SWOG治疗试验,分成了17组不同的癌症特异性分析组。不论农村居民是如何确定的,17组分析组(处于辅助阶段、雌激素受体阴性和孕激素受体阴性的乳腺癌患者)中仅有1组农村患者的生存率在统计学上显著下降。

    意义  改善统一治疗策略(如临床试验中发现的策略)的可获取程度,可能有助于解决农村和城市患者之间癌症结果的差异。

    Abstract

    Importance  Studies showing that patients with cancer from rural areas have worse outcomes than their urban counterparts have relied on cancer population data and did not account for differences in access to care. Clinical trial patients receive protocol-directed care by design, so large clinical trial databases are ideal for examining the impact of rural vs urban residency on outcomes.

    Objective  To compare the geographic distribution and survival outcomes for rural vs urban patients with cancer treated in clinical trials.

    Design, Setting, and Participants  In this comparative effectiveness retrospective cohort analysis, 36 995 patients from all 50 states enrolled in 44 phase 3 and phase 2/3 SWOG (formerly the Southwest Oncology Group) treatment trials from January 1, 1986, to December 31, 2012, were examined. Seventeen different cancer-specific analysis cohorts were constructed. Data through January 30, 2018, were analyzed.

    Main Outcomes and Measures  Rural vs urban residency was defined using the Rural-Urban Continuum Codes developed by the US Department of Agriculture. Multivariate Cox regression was used to estimate the association of residency with overall survival, progression-free survival, and cancer-specific survival, controlling for major disease-specific prognostic factors and demographic variables and stratifying by study. Different definitions of rurality were examined. The distribution of rural vs urban patients by geographic region was described.

    Results  Overall, 27.7% of patients were 65 years or older (range across 17 cohort analyses, 7.8%-74.5%), 40.3% were female in the non-sex-specific analyses (range across 17 cohort analyses, 28.1%-45.9%), and 10.8% were black (range across 17 cohort analyses, 1.9%-22.4%). Overall, 19.4% of patients (7184 of 36 995) were from rural locations. Rural patients were more likely to be aged 65 years or older (rural, 30.7% aged ≥65 years vs urban, 27.0% aged ≥65 years; difference, 3.7%; 95% CI, 2.5%-4.9%; P < .001), were less likely to be black (rural, 5.4% vs urban, 12.1%; difference, 6.7%; 95% CI, 6.1%-7.3%; P < .001), were similar with respect to sex (rural, 40.4% female vs urban, 39.7% female; difference, 0.6%; 95% CI, −1.4% to 2.6%; P = .53), and were well represented within major US geographic regions (West, Midwest, South, and Northeast). Clinical prognostic factors were similar. In multivariable regression, rural patients with adjuvant-stage estrogen receptor–negative and progesterone receptor–negative breast cancer had worse overall survival (hazard ratio, 1.27; 95% CI, 1.06-1.51; P = .008) and cancer-specific survival (hazard ratio, 1.26; 95% CI, 1.04-1.52; P = .02). No other statistically significant differences for overall, progression-free, or cancer-specific survival were found. Results were consistent regardless of the definition of rurality.

    Conclusions and Relevance  Rural and urban patients with uniform access to cancer care through participation in a SWOG clinical trial had similar outcomes. This finding suggests that improving access to uniform treatment strategies for patients with cancer may help resolve the disparity in cancer outcomes between rural and urban patients.

    Introduction

    Nineteen percent of the US population overall, and of the US population with cancer in particular, are from rural areas.1,2 Rural patients with cancer have been shown to have worse outcomes than their urban counterparts. A major recent report indicated that the age-adjusted rate of cancer deaths in rural areas from 2011 to 2015 was 180.4 per 100 000 individuals, compared with just 157.8 per 100 000 individuals in large metropolitan areas.2 Thought leaders have expressed concerns that these differences might be attributed to rural individuals’ reduced access to medical, technological, and financial resources, as well as adverse health status and shortened life spans.3 The Cancer Moonshot initiative emphasized that rural patients with cancer experience disproportionate morbidity and mortality and indicated the importance of increased research into disparities between rural and urban patients.4 Indeed, disparities in cancer outcomes for rural patients in the United States may actually be increasing rather than decreasing.2 Whether this disparity is due to inadequate access to quality cancer care or other characteristics of patients residing in rural areas, such as different clinical, demographic, or disease profiles, is unclear.

    In this context, an important question is whether rural and urban patients with cancer who receive similar care have similar outcomes. To address this, we compared survival outcomes between patients with cancer from rural vs urban locales who participated in therapeutic clinical trials. Patients receiving care in this setting are uniformly staged, treated, and followed up under protocol-specific guidelines, reducing the potential influences of inconsistent pretreatment evaluation, care, and posttreatment surveillance. If outcomes between rural and urban patients with cancer in clinical trials are similar, then access to uniform treatment strategies of the type represented by clinical trial care could help alleviate disparities in outcomes.

    Methods
    Patients

    Data were derived from patient medical records for trial participants enrolled between January 1, 1986, and December 31, 2012, to clinical treatment trials conducted by SWOG (formerly the Southwest Oncology Group), a National Clinical Trials Network and Community Oncology Research Program group sponsored by the National Cancer Institute (NCI). Only data from phase 3 or large phase 2/3 trials for which the primary analysis was previously published were included. The following cancer types were included: acute myeloid leukemia, brain, breast, colorectal, lung, lymphoma, myeloma, ovarian, prostate, and sarcoma (gastrointestinal stromal tumors). We combined trials with similar histology and stage to increase our power to identify potential differences in survival outcomes between rural and urban patients. In the case of advanced prostate cancer, we analyzed SWOG trials S8894 and S9346 separately, because each of these trials enrolled more than 1000 patients. Each trial included in this analysis was previously approved by an institutional review board; informed consent was previously obtained from all patients for each study included. Institutional review board approval and informed consent of study participants for this comparative effectiveness study was not required because secondary data that were not identifiable were used. The research question and analysis plan were defined prospectively in accordance with the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) reporting guideline5 for the conduct of comparative effectiveness studies.

    Covariates and Outcome Variables

    We defined rural residence using 2003 Rural-Urban Continuum Codes (RUCCs) developed by the Economic Research Service of the US Department of Agriculture (eTable in the Supplement).6 We matched RUCCs with patient zip codes using 2010 United States Postal Service Zip Code Crosswalk Files from the US Department of Housing and Urban Development.7 Guided by recent studies in the literature, our primary analysis used a prespecified cut point, dividing the 9 RUCCs into 2 categories: urban (RUCCs 1-3) vs rural (RUCCs 4-9).8-10

    Each analysis adjusted for the demographic variables age (<65 years vs ≥65 years), self-reported race (black vs other), and sex (where appropriate), which could potentially influence the relationship between residency status and survival outcomes. In addition, each analysis adjusted for important disease-specific clinical adjustment variables. Only clinical adjustment variables with a known effect on survival that were measured in all studies within an analysis were included (see Table 1 for clinical adjustment variables). These variables often represented those factors used to balance randomization assignment (ie, stratification factors).

    Because men represented a very small fraction of the patients with breast cancer, male patients were excluded from the 3 breast cancer groups, and sex was not included as a covariate in these analyses. Analyses of non-Hodgkin lymphoma did not include age as a separate covariate because age is included in the International Prognostic Index.

    The primary outcome was overall survival (OS), measured as days from study registration to death by any cause or, for those patients still alive, as days to last contact (censored). Patients lost to follow-up were also censored at their date of last contact. We also examined progression-free survival (PFS) and cancer-specific survival (CSS) as secondary outcomes. We defined PFS as days from study registration to the date of death by any cause, evidence of protocol-defined relapse or progressive disease, or date of last contact for those alive and progression free (censored). We measured CSS as days from study registration to date of cancer-specific death, death by another cause (censored), or last contact for those alive at last contact (censored). Detailed cause-of-death information was available for 28.0% of patient deaths. For the remaining 72.0%, we considered any death preceded by documented relapse or progression a cancer-specific death. We limited analyses to the first 5 years after registration in our main analysis in order to focus on cancer-related and treatment-related survival. Patients with last contact date or death date greater than 5 years postregistration were censored at 5 years of follow-up. Survival outcomes were grouped by adjuvant (ie, nonmetastatic) vs advanced (ie, metastatic) to identify whether global patterns of outcomes for rural vs urban patients differed by these disease settings.

    Statistical Analysis

    To examine observed survival differences by residency status, we generated Kaplan-Meier OS curves for each of the 17 cancer cohorts, separately for rural and urban patients.11 We used multivariate Cox regression to estimate the effect of patient residence on survival outcomes while controlling for the major disease-specific prognostic factors and the demographic variables.12 Each analysis was stratified by study.

    We further explored whether patterns of survival between rural and urban patients might substantively differ for different definitions of rural residency. We used variable cut point analysis to examine the association of residency and OS for each of the 8 possible definitions of rural residency based on the RUCCs.13,14 Although the primary analysis was based on 5 years of follow-up, for the variable cut point analysis, we allowed follow-up to vary from 1 to 10 years to allow for different prognosis patterns across the panel of 17 cancer types, resulting in 80 separate analyses per cancer type, or 1360 analyses overall. For each analysis, we derived the signed (according to the direction of the effect) χ2 statistic for the model-estimated association of rural or urban residence and OS in multivariable Cox regression.12 We treated these 17 χ2 statistics as independent random variables and used a 1-sample t test to assess whether they collectively differed from 0. We then plotted the corresponding signed z scores.

    Tests for statistical significance were 2-sided, α = .05. Informal (P = .01) and formal (per Bonferroni) multiple comparison tests for the variable cut point analysis were provided. Data through January 30, 2018, were examined.

    Results

    Overall, 36 995 patients enrolled between January 1, 1986, and December 31, 2012, in 44 SWOG phase 2/3 or phase 3 trials were analyzed (Table 1). Of the total study population, 19.4% resided in rural areas, the same as the rural proportion of the US population with cancer. Figure 1 shows a map of the trial registrations by rural vs urban status. Although all states are represented, the states with the fewest enrollments were Maine (82), Wyoming (93), and Rhode Island (102). Registrations in SWOG trials generally reflect the geographic population patterns of the United States, although with more patients enrolled from the Midwest (39% vs 21%) and fewer from the south (24% vs 37%). Rural patients were well represented within major geographic regions in the United States (West, Midwest, South, and Northeast) (Figure 1).

    Patient Characteristics

    Overall, 27.7% (range across the 17 cohort analyses, 7.8%-74.5%) of patients (10 247 of 36 995) were aged 65 years or older, 40.3% (range across 17 cohort analyses, 28.1%-45.9%; 5381 of 13 360 patients) were female in the non-sex-specific analyses, and 10.8% (range across 17 cohort analyses, 1.9%-22.4%; 4003 of 36 995 patients) were black. Table 2 shows these statistics as well as descriptions of the clinical prognostic factors used in the analyses, both overall and separately by rural and urban residence. Patients residing in rural areas were older on average (rural, 30.7% aged ≥65 years vs urban, 27.0% aged ≥65 years; difference, 3.7%; 95% CI, 2.5%-4.9%; P < .001), were similar with respect to sex (rural, 40.4% female vs urban, 39.7% female; difference, 0.6%; 95% CI, −1.4% to 2.6%; P = .53), and were less likely to report being black (rural, 5.4% vs urban, 12.1%; difference, 6.7%; 95% CI, 6.1%-7.3%; P < .001) than patients residing in urban areas. There were very few statistically significant differences between rural and urban patients with respect to clinical prognostic factors.

    Survival Outcomes

    The median follow-up time among patients still alive at the date of last contact for the primary analysis was the same for rural and urban patients (5.0 years each). There were few clear observed differences by residency status in OS using Kaplan-Meier curves (eFigure in the Supplement). In patients with adjuvant-stage estrogen receptor–negative, progesterone receptor–negative breast cancer, multivariate Cox regression results showed worse OS (hazard ratio, 1.27; 95% CI, 1.06-1.51; P = .008) and worse CSS (hazard ratio, 1.26; 95% CI, 1.04-1.52; P = .02) for patients residing in rural areas. No other statistically significant differences for any survival outcome for any of the remaining 16 disease cohorts were observed (Figure 2).

    Variable Cut Point Analyses

    Figure 3 shows the results of the variable cut point analyses. Negative z scores reflect better outcomes for rural patients; positive z scores reflect worse outcomes for rural patients. The division of the RUCCs into urban (1) vs rural (2-9) maximized the association between residence and OS across the panel of 17 cancer cohorts. This categorization is problematic, however, because defining codes 2 and 3 as rural is not representative and results in a rural-to-urban ratio that is much different from that of the US population. The second highest average association between residence and OS occurred for rural defined as RUCCs 1 to 3 vs urban defined as RUCCs 4 to 9, reflecting the cut point specified for our primary analysis. Importantly, there was no evidence that any combination of RUCC cut point and follow-up time (1 to 10 years) resulted in a general pattern of better or worse survival for rural patients with cancer across the 17 cancer cohorts.

    Discussion

    Numerous prior studies have documented inferior survival for rural patients with cancer.10,15-17 However, these studies did not adequately account for differences in access to cancer care. To our knowledge, no prior study has systematically compared cancer outcomes in clinical trial participants who reside in urban vs rural settings. This comprehensive analysis examined nearly 37 000 patients with a wide variety of cancer types and cancer stages. All participants were uniformly staged, treated, and followed up under protocol-directed care in a trial setting. Although rural trial participants were older on average and less likely to identify as black than their urban counterparts, after adjustment for these factors and important clinical prognostic factors, we found no systematic pattern of differences in survival outcomes by residency status, even under varied definitions of rural residency. Thus, this study contributes to the overall evidence base about differences in survival outcomes between rural and urban patients with cancer by showing that under similar treatment conditions within a clinical trial setting, rural and urban patients having the included cancer types do in fact have similar outcomes. This finding suggests that previously observed differences in outcomes for rural and urban patients with cancer may be due, in part, to inadequate receipt of guideline-concordant cancer care (as provided in clinical trials) rather than other factors intrinsic to residing in rural areas, such as unmeasurable differences in cancer prognosis or socioeconomic status.18 This conclusion is reinforced by the observation of very few differences in clinical prognostic factors between rural vs urban patients in our study.

    It is more difficult to access adequate medical care for rural individuals.3,19 In the United States, rural residency has emerged as an important predictor of care access in the general population of individuals with cancer.2 At the same time, rural oncology resources are sparse; although 20% of the population is rural, only 3% of oncologists work in rural areas.20 Rural patients with cancer are required to travel much greater distances to receive care, adding time and financial burdens to treatment. One study found that fewer than 50% of rural patients with colorectal cancer in small or isolated rural areas had access to an oncologist within 30 miles.21 This travel burden can result in lower rates of standard care; rural patients with cancer were found to be about half as likely to receive breast-conserving therapy for early-stage breast cancer compared with the national average, and rates declined further as travel distance increased.22 A large study of nearly 35 000 patients found that patients with stage III colon cancer who had to travel very long distances (≥250 miles) to visit an oncologist were only one-third as likely to receive adjuvant chemotherapy.23 Research from Australia proposes numerous access issues that affect receipt of quality care, including delayed diagnosis due to limited access to screening or prevention tools, limited access to treating specialists, financial barriers to travel for treatment, and the physical burden of travel.24-26 Furthermore, residents of rural areas are least likely to have private health care coverage, accentuating disparities in access to health care and prevention services.27,28 Unfortunately, our study is unable to shed light on these issues given that all patients we studied were enrolled in trials.

    These findings have implications for the current policy debates regarding the Affordable Care Act (ACA). Nineteen percent of patients with cancer reside in rural areas, the same as the rate of clinical trial patients from rural areas observed in this study.1,2 Therefore, rural patients with cancer make up a sizeable minority of patients with cancer who may have special needs.29 It is estimated that at least 22 million people will lose health care coverage if the ACA is dismantled, 14 million through loss of expanded Medicaid coverage.30 In many states, rural patients benefit disproportionately from Medicaid expansion. Adequate insurance is key to ensuring access to quality care. Therefore, reducing insurance access through policy prescriptions that limit or dismantle the ACA may disproportionately impact health care access for rural patients, again further widening disparities in outcomes between rural and urban patients with cancer.

    Models for improving access to quality cancer care in rural settings do exist. In Australia, the establishment of Regional Cancer Centers of Excellence are linked to major urban cancer centers to provide multidisciplinary care, improve support services, and improve clinical trial participation.31 Successful centers have substantially increased the number of patients treated locally, removing a significant burden for rural patients. A shared approach between local practitioners and specialists could be a particularly useful model.32 Communications facilitated through teleoncology models have been found to improve access to specialist consultations and receipt of chemotherapy closer to home.32,33 In the United States, the NCI’s National Community Oncology Research Program, of which SWOG is a member, is designed to bring clinical trials to community investigators and patients.34 The success of this program is reflected by the results of our study, showing representative enrollment of rural patients with cancer from different regions throughout the United States (Figure 1). The Cancer Moonshot Blue Ribbon Report recommends the establishment of a large-scale patient participation network for comprehensive tumor profiling.4 Patients would enroll directly in the network, an approach that could reduce barriers to access to comprehensive cancer testing and novel treatments for rural and other medically disadvantaged groups.

    Limitations

    The results of this analysis are limited by the fact that they may not represent patterns of outcomes for rural vs urban patients receiving quality care outside of a trial setting. Clinical trial participants have been shown to have better outcomes than those treated in a nontrial setting in the short term (ie, the first year after diagnosis), likely due to differences in baseline comorbid conditions.35 Although this could impact the absolute estimates of survival outcomes for patients examined in this analysis, it is less likely to affect the relative rates by rural vs urban status. Moreover, a randomized comparison of rural vs urban patients and cancer survival outcomes is clearly not feasible in this setting; thus, although our regression results accounted for important demographic and clinical prognosis factors, the potential remains that unknown confounders could influence the results. For instance, rural patients who participate in clinical trials may also be more likely to have improved health behaviors,2 although we found very little evidence of measurable differences between rural and urban patients in clinical risk factors (Table 2). Also, travel distance may lead to trial participation barriers that differ between rural and urban patients with cancer. However, the influence of this factor for patients choosing to participate in trials (as opposed to receiving care outside of a trial) is uncertain,36 especially because patients in urban areas may also choose to travel long distances for trial participation. The analysis of CSS was limited by incomplete cause-of-death information. Furthermore, we cannot presently explain why rural patients in the estrogen receptor–negative, progesterone receptor–negative, adjuvant breast cancer cohort had worse survival. One possibility is that residency is more relevant in adjuvant cancer settings, where prognosis is better overall, although we observed survival differences in only 1 of the 4 adjuvant cohorts we examined. Another possibility is that rural patients with cancer may be more likely to have delays in chemotherapy administration, which can adversely affect survival in patients with estrogen receptor–negative, progesterone receptor–negative cancers in particular.37

    Conclusions

    Substantial efforts to identify and mitigate disparities in access to care and outcomes based on race, ethnicity, and socioeconomic factors have received great attention.38 Yet few efforts or research have focused on disparities related to geographic residency, despite the fact that the rural population in the United States represents 1 in 5 individuals. This is the largest examination of survival outcomes for rural vs urban patients with cancer receiving care in a clinical trial setting. The finding of almost no differences in outcomes by residency status across 17 different patient cohorts has potential policy implications. If rural and urban patients with cancer receiving similar care also have similar outcomes, then a reasonable inference is that the best means by which to improve outcomes for rural patients with cancer may be to improve their access to quality care. Better access to affordable health care insurance, better access to screening and prevention tools, better access to treating specialists, improved resources for traveling to receive care, and innovative new networks to give rural patients better access to new, novel treatments and clinical trials are all likely to improve outcomes for patients with cancer in rural areas. Thus multiple parallel efforts may begin to ameliorate the persistent disparity in cancer outcomes between rural and urban patients.

    Back to top
    Article Information

    Accepted for Publication: May 22, 2018.

    Published: August 17, 2018. doi:10.1001/jamanetworkopen.2018.1235

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2018 Unger JM et al. JAMA Network Open.

    Corresponding Author: Joseph M. Unger, PhD, MS, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave N, Ste M3-C102, PO Box 19024, Seattle, WA 98109-1024 (junger@fredhutch.org).

    Author Contributions: Dr Unger and Ms Moseley had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Unger, Moseley, Symington, Ramsey, Hershman.

    Acquisition, analysis, or interpretation of data: Unger, Moseley, Chavez-MacGregor, Hershman.

    Drafting of the manuscript: Unger, Moseley, Symington, Hershman.

    Critical revision of the manuscript for important intellectual content: Unger, Moseley, Chavez-MacGregor, Ramsey, Hershman.

    Statistical analysis: Unger, Moseley.

    Obtained funding: Unger, Symington.

    Administrative, technical, or material support: Symington.

    Supervision: Unger, Ramsey, Hershman.

    Conflict of Interest Disclosures: Dr Unger reported grants from the HOPE Foundation and the National Cancer Institute during the conduct of the study. Ms Moseley reported other support from the National Cancer Institute during the conduct of the study. Dr Chavez-MacGregor reported employment in MD Anderson Physician’s Network, consulting or advisory roles for Pfizer and Roche/Genentech, research funding from Novartis, and travel funding from Pfizer. Dr Ramsey reported consulting or advisory roles for Bayer, Bristol-Myers Squibb, Genentech, Kite Pharma, and Seattle Genetics; research funding from Bayer and Bristol-Myers Squibb; and travel funding from Bayer Schering Pharma, Bristol-Myers Squibb, and Flatiron Health. No other disclosures were reported.

    Funding/Support: Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health (grant 5UG1CA189974) and by the HOPE Foundation.

    Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

    References
    1.
    US Census Bureau. Measuring America: our changing landscape. https://www.census.gov/content/dam/Census/library/visualizations/2016/comm/acs-rural-urban.pdf. Published December 8, 2016. Accessed May 17, 2018.
    2.
    Henley  SJ, Anderson  RN, Thomas  CC, Massetti  GM, Peaker  B, Richardson  LC.  Invasive cancer incidence, 2004-2013, and deaths, 2006-2015, in nonmetropolitan and metropolitan counties—United States.  MMWR Surveill Summ. 2017;66(14):1-13. doi:10.15585/mmwr.ss6614a1PubMedGoogle ScholarCrossref
    3.
    Iglehart  JK.  The challenging quest to improve rural health care.  N Engl J Med. 2018;378(5):473-479. doi:10.1056/NEJMhpr1707176PubMedGoogle ScholarCrossref
    4.
    National Cancer Institute. National Cancer Advisory Board Cancer Moonshot Blue Ribbon Panel Report. https://www.cancer.gov/research/key-initiatives/moonshot-cancer-initiative/blue-ribbon-panel. Published October 17, 2016. Accessed January 26, 2018.
    5.
    Berger  ML, Mamdani  M, Atkins  D, Johnson  ML.  Good research practices for comparative effectiveness research: defining, reporting and interpreting nonrandomized studies of treatment effects using secondary data sources: the ISPOR Good Research Practices for Retrospective Database Analysis Task Force Report—Part I.  Value Health. 2009;12(8):1044-1052. doi:10.1111/j.1524-4733.2009.00600.xPubMedGoogle ScholarCrossref
    6.
    US Department of Agriculture. Rural-Urban Continuum Codes. https://www.ers.usda.gov/data-products/rural-urban-continuum-codes/.aspx. Updated October 12, 2016. Accessed May 17, 2018.
    7.
    Office of Policy Development and Research; US Department of Housing and Urban Development. HUD USPS Zip Code Crosswalk Files. https://www.huduser.gov/portal/datasets/usps_crosswalk.html. Accessed May 17, 2018.
    8.
    Aboagye  JK, Kaiser  HE, Hayanga  AJ.  Rural-urban differences in access to specialist providers of colorectal cancer care in the United States: a physician workforce issue.  JAMA Surg. 2014;149(6):537-543. doi:10.1001/jamasurg.2013.5062PubMedGoogle ScholarCrossref
    9.
    Singh  GK.  Rural-urban trends and patterns in cervical cancer mortality, incidence, stage, and survival in the United States, 1950-2008.  J Community Health. 2012;37(1):217-223. doi:10.1007/s10900-011-9439-6PubMedGoogle ScholarCrossref
    10.
    Weaver  KE, Palmer  N, Lu  L, Case  LD, Geiger  AM.  Rural-urban differences in health behaviors and implications for health status among US cancer survivors.  Cancer Causes Control. 2013;24(8):1481-1490. doi:10.1007/s10552-013-0225-xPubMedGoogle ScholarCrossref
    11.
    Kaplan  EL, Meier  P.  Nonparametric estimation from incomplete observations.  J Am Stat Assoc. 1958;53:457-481. doi:10.1080/01621459.1958.10501452Google ScholarCrossref
    12.
    Cox  DR.  Regression models and life tables.  J Royal Stat Soc. 1972;34(2):187-220. doi:10.1007/978-1-4612-4380-9_37Google Scholar
    13.
    Greenlee  H, Unger  JM, LeBlanc  M, Ramsey  S, Hershman  DL.  Association between body mass index and cancer survival in a pooled analysis of 22 clinical trials.  Cancer Epidemiol Biomarkers Prev. 2017;26(1):21-29. doi:10.1158/1055-9965.EPI-15-1336PubMedGoogle ScholarCrossref
    14.
    Rimsza  LM, Unger  JM, Tome  ME, Leblanc  ML.  A strategy for full interrogation of prognostic gene expression patterns: exploring the biology of diffuse large B cell lymphoma.  PLoS One. 2011;6(8):e22267. doi:10.1371/journal.pone.0022267PubMedGoogle ScholarCrossref
    15.
    Olson  RA, Nichol  A, Caron  NR,  et al.  Effect of community population size on breast cancer screening, stage distribution, treatment use and outcomes.  Can J Public Health. 2012;103(1):46-52.PubMedGoogle Scholar
    16.
    Onega  T, Duell  EJ, Shi  X, Wang  D, Demidenko  E, Goodman  D.  Geographic access to cancer care in the U.S.  Cancer. 2008;112(4):909-918. doi:10.1002/cncr.23229PubMedGoogle ScholarCrossref
    17.
    Williams  F, Thompson  E.  Disparity in breast cancer late stage at diagnosis in Missouri: does rural versus urban residence matter?  J Racial Ethn Health Disparities. 2016;3(2):233-239. doi:10.1007/s40615-015-0132-9PubMedGoogle ScholarCrossref
    18.
    Meilleur  A, Subramanian  SV, Plascak  JJ, Fisher  JL, Paskett  ED, Lamont  EB.  Rural residence and cancer outcomes in the United States: issues and challenges.  Cancer Epidemiol Biomarkers Prev. 2013;22(10):1657-1667. doi:10.1158/1055-9965.EPI-13-0404PubMedGoogle ScholarCrossref
    19.
    National Center for Health Statistics.  Health, United States, 2015: With Special Feature on Racial and Ethnic Health Disparities. Hyattsville, MD: National Center for Health Statistics; 2016.
    20.
    Kirkwood  MK, Bruinooge  SS, Goldstein  MA, Bajorin  DF, Kosty  MP.  Enhancing the American Society of Clinical Oncology workforce information system with geographic distribution of oncologists and comparison of data sources for the number of practicing oncologists.  J Oncol Pract. 2014;10(1):32-38. doi:10.1200/JOP.2013.001311PubMedGoogle ScholarCrossref
    21.
    Baldwin  LM, Cai  Y, Larson  EH,  et al.  Access to cancer services for rural colorectal cancer patients.  J Rural Health. 2008;24(4):390-399. doi:10.1111/j.1748-0361.2008.00186.xPubMedGoogle ScholarCrossref
    22.
    Meden  T, St John-Larkin  C, Hermes  D, Sommerschield  S.  Relationship between travel distance and utilization of breast cancer treatment in rural northern Michigan.  JAMA. 2002;287(1):111. doi:10.1001/jama.287.1.111-JMS0102-5-1PubMedGoogle ScholarCrossref
    23.
    Lin  CC, Bruinooge  SS, Kirkwood  MK,  et al.  Association between geographic access to cancer care, insurance, and receipt of chemotherapy: geographic distribution of oncologists and travel distance.  J Clin Oncol. 2015;33(28):3177-3185. doi:10.1200/JCO.2015.61.1558PubMedGoogle ScholarCrossref
    24.
    Mitchell  KJ, Fritschi  L, Reid  A,  et al.  Rural-urban differences in the presentation, management and survival of breast cancer in Western Australia.  Breast. 2006;15(6):769-776. doi:10.1016/j.breast.2006.04.001PubMedGoogle ScholarCrossref
    25.
    Sabesan  S, Burgher  B, Buettner  P,  et al.  Attitudes, knowledge and barriers to participation in cancer clinical trials among rural and remote patients.  Asia Pac J Clin Oncol. 2011;7(1):27-33. doi:10.1111/j.1743-7563.2010.01342.xPubMedGoogle ScholarCrossref
    26.
    Yu  XQ, O’Connell  DL, Gibberd  RW, Armstrong  BK.  Assessing the impact of socio-economic status on cancer survival in New South Wales, Australia 1996-2001.  Cancer Causes Control. 2008;19(10):1383-1390. doi:10.1007/s10552-008-9210-1PubMedGoogle ScholarCrossref
    27.
    Meit  M, Knudson  A, Gilbert  T,  et al. The 2014 Update of the Rural-Urban Chartbook. Bethesda, MD: Rural Health Reform Policy Research Center; 2014. https://ruralhealth.und.edu/projects/health-reform-policy-research-center/pdf/2014-rural-urban-chartbook-update.pdf. Accessed May 17, 2018.
    28.
    Ward  E, Halpern  M, Schrag  N,  et al.  Association of insurance with cancer care utilization and outcomes.  CA Cancer J Clin. 2008;58(1):9-31. doi:10.3322/CA.2007.0011PubMedGoogle ScholarCrossref
    29.
    Bettencourt  BA, Schlegel  RJ, Talley  AE, Molix  LA.  The breast cancer experience of rural women: a literature review.  Psychooncology. 2007;16(10):875-887. doi:10.1002/pon.1235PubMedGoogle ScholarCrossref
    30.
    Fiedler  M, Aaron  HJ, Adler  L, Ginsburg  PB.  Moving in the wrong direction—health care under the AHCA.  N Engl J Med. 2017;376(25):2405-2407. doi:10.1056/NEJMp1706848PubMedGoogle ScholarCrossref
    31.
    Underhill  CR, Goldstein  D, Grogan  PB.  Inequity in rural cancer survival in Australia is not an insurmountable problem.  Med J Aust. 2006;185(9):479-480.PubMedGoogle Scholar
    32.
    Campbell  NC, Ritchie  LD, Cassidy  J, Little  J.  Systematic review of cancer treatment programmes in remote and rural areas.  Br J Cancer. 1999;80(8):1275-1280. doi:10.1038/sj.bjc.6690498PubMedGoogle ScholarCrossref
    33.
    Sabesan  S, Larkins  S, Evans  R,  et al.  Telemedicine for rural cancer care in North Queensland: bringing cancer care home.  Aust J Rural Health. 2012;20(5):259-264. doi:10.1111/j.1440-1584.2012.01299.xPubMedGoogle ScholarCrossref
    34.
    National Cancer Institute. Community Oncology Research Program. https://ncorp.cancer.gov/about/. Accessed May 17, 2018.
    35.
    Unger  JM, Barlow  WE, Martin  DP,  et al.  Comparison of survival outcomes among cancer patients treated in and out of clinical trials.  J Natl Cancer Inst. 2014;106(3):dju002. doi:10.1093/jnci/dju002PubMedGoogle ScholarCrossref
    36.
    Lamont  EB, Hayreh  D, Pickett  KE,  et al.  Is patient travel distance associated with survival on phase II clinical trials in oncology?  J Natl Cancer Inst. 2003;95(18):1370-1375. doi:10.1093/jnci/djg035PubMedGoogle ScholarCrossref
    37.
    Chavez-MacGregor  M, Clarke  CA, Lichtensztajn  DY, Giordano  SH.  Delayed initiation of adjuvant chemotherapy among patients with breast cancer.  JAMA Oncol. 2016;2(3):322-329. doi:10.1001/jamaoncol.2015.3856PubMedGoogle ScholarCrossref
    38.
    Smedley  BD, Stith  AY, Nelson  AR, eds.  Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Washington, DC: The National Academies Press; 2003:118-143.
    ×