Geographic disparities in adherence to adjuvant endocrine therapy in Appalachian women with breast cancer

https://doi.org/10.1016/j.sapharm.2016.08.004Get rights and content

Abstract

Background

Appalachia is a largely rural, mountainous, poor and underserved region of the United States. Adherence to adjuvant endocrine therapy among Appalachian women with breast cancer is suboptimal.

Objectives

To explore small-area geographic variations and clustering patterns of breast cancer patient adherence to adjuvant endocrine therapy and associated factors in Appalachia.

Methods

In this retrospective study, we analyzed Medicare claims data linked with cancer registries from four Appalachian states (PA, OH, KY, and NC) in 2006–2008. We included adult women who were diagnosed with stage I–III, hormone-receptor positive, primary breast cancer and who newly started adjuvant endocrine therapy after the primary treatment for breast cancer. Hot spot analysis was conducted to explore geographic variations in adjuvant endocrine therapy adherence. Geographically weighted logistic regression (GWLR) was used to examine whether the impacts of factors associated with adherence varied across the region.

Results

Breast cancer patients living in PA and OH showed higher adherence to adjuvant endocrine therapy than those living in KY and NC. We identified clusters of high adherence in most of PA but poor adherence in Erie County, PA and in Buncombe, Transylvania, Henderson, and Polk Counties, NC. Adherence to adjuvant endocrine therapy was significantly associated with the Health Professional Shortage Area designation, catastrophic coverage, dual-eligibility status of Medicaid and Medicare, adjuvant endocrine therapy drug class, and side effects. And among these factors, the impacts of dual-eligibility status and the use of pain medications to treat side effects on adherence were more pronounced in KY and NC than in PA.

Conclusions

There were significant geographic disparities in adherence to adjuvant endocrine therapy in the Appalachian counties in PA, OH, KY, and NC. This study explored these geographic areas with poor adherence as well as geographically varying effects of predictors on adherence; our results may provide more localized information that may be used to improve adjuvant therapy use and breast cancer care in these high-risk and underserved areas.

Introduction

Adjuvant endocrine therapy (AET) is an important treatment modality for hormone-receptor (HR) positive breast cancer due to its significant benefits in reducing recurrence and mortality.1, 2, 3, 4 To achieve the optimal clinical benefits, adherence to AET is crucial.5, 6, 7 The current literature has identified many individual characteristics that may be inversely related to AET adherence such as extreme age (younger [under 40–45 years old] or older [over 75–85 years old]), higher out-of-pocket drug costs, switching drugs, drug class (aromatase inhibitors vs. tamoxifen), suboptimal patient-centered communication, lack of perceived self-efficacy in patient–physician interaction, and adverse drug reactions,5, 8, 9, 10, 11 but the literature has paid limited attention to geography or associated factors. In fact, geography can serve as a proxy or composite measure for various observed and unobserved variables that may be related to medication adherence, such as access to care, available health resources, socioeconomic status, disease burden, race/ethnicity, and culture.12 Examining geography and associated factors may help advance AET adherence research by further explaining individual variations in AET adherence that cannot be fully explained by individual characteristics. Small-area geographic variations in medication adherence may be attributable to the neighborhood effect, which describes the social interactions impacting an individual's behavior or outcomes.13 Theoretically, people residing in the same neighborhood are more likely to share common social norms, cultural background, socioeconomic status, and systemic and lifestyle characteristics compared to people living in different neighborhoods, which may further shape health behaviors, including medication-taking behaviors, above the individual-level.14 There may also be provider neighborhood effects, such as possible ineffective or inadequate patient-provider communication in the Health Professional Shortage Area (HPSA) that leads to the failure to underscore the importance of AET adherence, or similar prescribing or practice behaviors under the influence of similar policies, regulations or interventions in one area.

The Appalachian region of the United States (US) covers 204,452 square miles in 420 counties along the spine of the Appalachian Mountains.15 This region contains all of West Virginia, and portions of 12 other states: New York, Pennsylvania, Ohio, Maryland, Kentucky, Virginia, Tennessee, North Carolina, South Carolina, Georgia, Alabama, and Mississippi. The Appalachian population in the US is a special population of interest in cancer research because it consistently suffers from a significant cancer burden, with higher cancer incidence and mortality than the non-Appalachian population.16, 17 In terms of breast cancer, compared to other regions, Appalachia experienced a slower decline in breast cancer mortality,18 and its patients receive guideline-recommended breast cancer screening and primary treatment at lower rates than those in other regions.19, 20, 21 The factors leading to poor access to and utilization of care in this region may include rural residence, geographic isolation, lack of public transportation, underdeveloped telecommunication infrastructure, high poverty and unemployment rates, inadequate medical resources, a shortage of healthcare professionals, lower levels of educational attainment, and attitudinal and cultural factors.20, 22, 23 Given the largely rural, mountainous environment and unsatisfactory patient adherence to AET in Appalachia as a whole,5 we need to measure geographic variations in AET adherence beyond the general urban and rural classification. The identification of “hot spots” that require monitoring and intervention can help local communities to develop strategies to improve cancer treatment use and outcomes. However, there have been very few studies examining geographic disparities in adjuvant cancer treatment use in this region, primarily due to the lack of data and of a representative study sample. Therefore, we pursued the following study aims: 1) to explore small-area geographic variations and clustering patterns of AET adherence; and 2) to examine spatial non-stationarity of the relationships between potential predictors and AET adherence.

Section snippets

Study design and study population

In this retrospective study, we analyzed Medicare claims data linked with cancer registries from four Appalachian states (PA, OH, KY, and NC) between January 1, 2006 and December 31, 2008. We only assessed the Appalachian counties in these four states, not including the non-Appalachian counties. The study design included a baseline period that began one year before the diagnosis date, and patients were followed from the date of the first AET prescription until death or until the end of the

Results

A total of 428 eligible women with an average age of 74.8 years old were included in the study. Table 1shows the basic characteristics of the study sample. The MPR values ranged from 0.06 to 1.20, with a mean of 0.83 and a standard deviation (SD) of 0.24. Only about 69.4% of the population were adherent to AET. Significant bivariate predictors of MPR >0.80 at α = 0.05 included having dual status (65.8% yes vs. 78.5% no), having catastrophic coverage (82.3% yes vs. 64.8 no), positive lymph nodes

Discussion

Patient adherence to AET is essential to maximize its significant benefits in cancer outcomes for HR-positive breast cancer survivors; therefore disparities in AET adherence may partly contribute to the disparities in breast cancer outcomes including mortality. Appalachia has experienced substantial cancer disparities over the years. In this study, we used innovative geographic analytic tools and a unique dataset linking Medicare claims with cancer registries from four Appalachian states (PA,

Conclusions

This study is among the first to demonstrate the utility and feasibility of using geographic techniques as a tool to account for geographic variations and neighborhood effects on medication adherence and its predictors in a specified region. It explored specific geographic areas in Appalachia with poor AET adherence as well as geographically varying effects of predictors on AET adherence, which may help direct future research, policy, and interventions to focus on these high-risk areas and

Acknowledgment

This study was funded by the National Cancer Institute (NCI) (Grant R21 CA168479) and the National Institutes of Health (NIH) office of Women's Health (Grant: 1 R21 CA168479; PI: Balkrishnan).

References (46)

  • L.R. Ayres et al.

    Adherence and discontinuation of oral hormonal therapy in patients with hormone receptor positive breast cancer

    Int J Clin Pharmcol

    (2014)
  • C.C. Murphy et al.

    Adherence to adjuvant hormonal therapy among breast cancer survivors in clinical practice: a systematic review

    Breast Cancer Res Treat

    (2012)
  • J.H. Lin et al.

    Predicting adherence to tamoxifen for breast cancer adjuvant therapy and prevention

    Cancer Prev Res

    (2011)
  • Y. Liu et al.

    Adherence to adjuvant hormone therapy in low-income women with breast cancer: the role of provider-patient communication

    Breast Cancer Res Treat

    (2013)
  • J.E. Couto et al.

    Geographic variation in medication adherence in commercial and Medicare Part D populations

    J Manag Care Spec Pharm

    (2014)
  • H.F. de Vries McClintock et al.

    Neighborhood social environment and patterns of adherence to oral hypoglycemic agents among patients with type 2 diabetes mellitus

    Fam Community Health

    (2015)
  • K. Pollard et al.

    The Appalachian Region: A Data Overview from the 2007–2011 American Community Survey

    (2013)
  • R.J. Wilson et al.

    Cancer incidence in Appalachia, 2004–2011

    Cancer Epidemiol Biomarkers Prev

    (2016)
  • R.L. Siegel et al.

    Cancer death rates in US congressional districts

    CA Cancer J Clin

    (2015)
  • N. Yao et al.

    Breast cancer mortality in Appalachia: reversing patterns of disparity over time

    J Health Care Poor Underserved

    (2012)
  • The Cancer Burden in Appalachia

    (2010)
  • H.I. Hall et al.

    Breast and cervical cancer screening among Appalachian women

    Cancer Epidemiol Biomarkers Prev

    (2002)
  • A.B. Freeman et al.

    Patterns of care with regard to surgical choice and application of adjuvant radiation therapy for preinvasive and early stage breast cancer in rural Appalachia

    Am J Clin Oncol

    (2012)
  • Cited by (16)

    • Geographic variation in and contextual factors related to biguanide adherence amongst medicaid enrolees with type 2 Diabetes Mellitus

      2022, SSM - Population Health
      Citation Excerpt :

      Knowledge about how medication adherence varies across geographic regions is extremely limited (Han et al., 2020). Studies that have examined geographic variation in medication adherence have adopted a variety of scales (U.S. regions, counties, ZIP codes, census tracts) for a variety of chronic health conditions (Couto et al., 2014; Erickson & Yuan-Nung, 2014; Han et al., 2020; Hoang, Kolenic, Kline-Rogers, Eagle, & Erickson, 2011; Pizzonia et al., 2019; Tan et al., 2017) making valid comparisons of relationships difficult due to issues including the modifiable areal unit problem (Openshaw, 1984). An early study of geographic variation demonstrated the feasibility of hot spot analysis to identify census tract-level clusters of medication non-adherence among Acute Coronary Syndrome patients (Hoang et al., 2011).

    • Examining factors associated with adherence to hormonal therapy in breast cancer patients

      2020, Research in Social and Administrative Pharmacy
      Citation Excerpt :

      Previous studies were restricted to people who were Medicare patients, or had commercial insurance. Consequently, they were not able to fully investigate the impact that insurance status has on adherence to hormonal therapy among breast cancer patients.11–15,24,28 MEPS data does include people with different types of insurance coverage and this study found that insurance status did not have a significant association with adherence to hormonal therapy.

    • How the relationships between preterm birth and ambient air pollution vary over space: A case study in Georgia, USA using geographically weighted logistic regression

      2018, Applied Geography
      Citation Excerpt :

      While a global logistic regression model relies on one regression equation for the entire dataset, GWLR, as a local spatial statistical technique, attempts to capture spatially varying relationships by calculating local regression results for each location of the observed cases (BioMedware, 2014). GWR and GWLR were also used to improve the reliabilities of the relationships among variables by taking into account spatial autocorrelations (Tu & Xia, 2008; Stojanova, Ceci, Appice, Malerba, & Džroski, 2013; Tan et al., 2017). To the best of our knowledge, this technique has never been reported to analyze the relationships between PTB and its risk factors including ambient air pollution.

    View all citing articles on Scopus
    View full text