Health Disparities Interest Group August 2009 Seminar

Introduction

• Dr. Diane Solomon, Division of Cancer Prevention, called the meeting to order and asked participants to introduce themselves.
• Dr. Solomon introduced the guest speaker, Dr. Ann L. Coker, who presented "What's Behind Observed Disparities in Cervical Cancer Survival?"

Summary of Dr. Coker's Presentation/Discussion

Background

• The University of Kentucky's Center for Research on Violence Against Women, where Dr. Coker holds the Verizon Wireless Endowed Chair, focuses on research to prevent violence. It is the only center to focus on this particular topic.
• Dr. Coker began her career as a cancer epidemiologist. Her work has focused on the etiology of cervical neoplasia, HPV, active and passive smoking, contraceptive use, and stress. With regard to health disparities in cervical cancer incidence and survival, her work has focused on race, socioeconomic status (SES), stress, support, and using cancer registries.
• Currently, Dr. Coker's work emphasizes the epidemiology of violence against women (VAW), especially intimate partner violence (IPV) and its health effects, including:
  • VAW as a stressor on chronic and acute women's health;
  • VAW's effects on squamous intraepithelial lesion (SIL) risk, screening frequency, cervical cancer incidence, and the cancer care continuum;
  • How VAW and SES modify the cancer care continuum.

Research

• Research Question: What factors explain observed racial/ethnic differences in cervical cancer survival?
  • In this context, "explain" means that observed racial differences in survival disappear when adjusting for a factor. Factors include: SES, urban/rural residence (access), treatment received (access), comorbid conditions (competing risks of death), age and stage at diagnosis (primary predictors of survival), and others (including partner violence against women).
• Texas studies
  • Dr. Coker conducted an NCI-funded study in a large, ethnically diverse population of nearly 7,000 cervical cancer cases with confounder data on neighborhood (not individual) SES and urbanicity. The investigators used Census data; geocoding of addresses; and random matching of zip code, race, age, and sex to estimate neighborhood urbanicity and SES.
  • Findings included that, compared to white women with cervical cancer, non-Hispanic black and Hispanic women with cervical cancer were more likely to have the lowest SES, live in an urban environment, and have their cancer diagnosed at a later stage.
  • In addition, compared to white women, black women with cervical cancer were found to be about 30 percent more likely to die of cervical cancer; Hispanic women with cervical cancer were found to be about 30 percent less likely to die of cervical cancer. (Reference: Eggleston KS, Coker AL, Williams M, Tortolero-Luna G, Martin JB, Tortolero SR. Socioeconomic status as a predictor of cervical cancer survival in Texas, 1995-2001. J Women's Health 2006;15(8):941-51.)
  • Similar racial/ethnic differences remained after adjusting for SES, urbanicity, age, cell type, and treatment, but only among those diagnosed at the later or missing stage. (Reference: Coker AL, DeSimone CP, Eggleston KS, White AL, Williams M. Ethnic disparities in cervical cancer survival among Texas women. J Women's Health, in press.)
  • After adjusting for comorbid conditions, treatment received, SES, urbanicity, etc., the same racial/ethnic pattern of lower survival rates for black women and higher survival rates for Hispanic women compared to white women remained. Medicare and Medicaid treatment data were used in this analysis.
  • In a smaller study of women with Medicare, Hispanic women remained less likely to die of cervical cancer than did white women. The study had limited power of analysis for black women. Further studies are needed to explain the "Hispanic paradox." (Reference: Coker, AL, Eggleston, KS, Du XL, Ramondetta L. Ethnic disparities in cervical cancer survival among Medicare eligible women in a multiethnic population. Intl J Gynecol Cancer 2009;19(1):13-20.)
• Kentucky study
  • A study using Kentucky Cancer Registry (KCR) data explored the quality of smoking data. (The population did not include sufficient numbers to explore racial/ethnic differences.) Women with cervical cancer who were current smokers at the time of diagnosis were significantly more likely to die of cervical cancer (and of any cause) than were other women. In addition, women who had no insurance or were self-payers were significantly more likely to die of cervical cancer. (Reference: Coker AL, DeSimone CP, Eggleston KS, Hopenhayn C, Nee J, Tucker T. Smoking and survival among Kentucky women diagnosed with invasive cervical cancer: 1995-2005. Gynecol Oncol 2009 Feb;112(2):365-9.)
• VAW among women with cancer
  • The only study on this topic found that nearly 50 percent of 101 women treated for breast, cervical, endometrial, or ovarian cancer reported a history of physical or sexual violence by a partner, and 25 percent reported childhood abuse. Current abuse was rarely reported. For the majority of women, violence was chronic (e.g., occurred more than 10 times). Women with cancer who had experienced IPV were less likely to have had a Pap test or mammography. A history of VAW was associated with currently being divorced, having less education, and being less likely to have private health insurance. Lifetime VAW was strongly associated with being diagnosed at a later cancer stage. (Reference: Modesitt SC, Gambrell AC, Cottrill HM, Hays LR, Walker R, Shelton BJ, Jordan CE, Ferguson JE 2nd. Adverse impact of a history of violence for women with breast, cervical, endometrial, or ovarian cancer. Obstet Gynecol 2006 Jun;107(6):1330-6.)
  • The American Psychological Association commissioned a two-volume monograph, currently in press, on VAW and Children: Consensus, Critical Analyses, and Emergent Priorities, Volumes 1 and 2. Dr. Coker contributed a chapter that reviewed peer-reviewed studies of IPV and cervical dysplasia or cancer with at least 500 subjects. The three studies that addressed IPV and having an abnormal Pap test noted a significant association, as did the three studies that investigated IPV and invasive cervical cancer.
  • The Kentucky Women's Health Registry is a Kentucky-based health survey begun in 2006 for women 18 years of age or older. Its purposes are to: (1) understand how risk and protective behavioral factors may differentially affect women's health, and (2) give interested women the chance to participate in medical research. The 10-year goal is to enroll more than 25,000 Kentucky women. To date, nearly 10,000 women have completed at least one survey. The survey includes items on VAW and cancers.
  • An in-press analysis by Dr. Coker and colleagues of Kentucky Women's Health Registry data showed that women who had experienced any IPV were about twice as likely to be diagnosed with cervical cancer.
• What is known about cervical cancer survival?
  • Race and more likely SES play a role in survival -- either directly or through access to treatment.
  • Treatment received is a strong predictor of survival; SES and race are predictors of treatment.
  • Smoking and perhaps other substance use affect survival.
  • VAW may be associated with cervical cancer through sexual abuse or through partner behaviors that affect the ability to obtain recommended care.
  • More subtle partner behaviors (e.g., partner-controlling behaviors, abuse of power) also may influence care.
• Partner influence on cancer care
  • The partners of women who currently are experiencing abuse almost always accompany those women when they seek care. This may prevent a woman from receiving needed care.
• Partner interference and health care (not just cancer care)
  • This topic rarely has been studied in large populations. Studies to date have shown that currently/recently abused women are more likely to report partner interference with health care than nonabused women.
  • Women with interfering partners also were more likely to report poorer health, even after controlling for recent violence and demographic characteristics. (Reference: McCloskey LA, Williams CM, Lichter E, Gerber M, Ganz ML, Sege R. Abused women disclose partner interference with health care: an unrecognized form of battering. J Gen Intern Med 2007 Aug;22(8):1067-72.)
• Effect of psychosocial factors on cancer care
  • Cancer outcomes of both partners can be affected by abuse and interfering/controlling behaviors. Multiple factors, including personality, social support, SES, personal dynamics, and spirituality, may be involved.
• Measuring VAW and partner-controlling behaviors
  • Validated measures of VAW exist, however more information is needed on current and lifetime VAW by type (e.g., psychological, sexual, physical, stalking).
  • No measures of partner-controlling/-interfering behaviors exist currently. Dr. Coker and colleagues are in the process of developing and evaluating such a measure.
• Next steps
  • Study to address VAW as a possible reason for disparities in cancer outcomes.
  • Qualitative assessment of partner-interfering and -controlling behaviors in gynecologic oncology patients.
  • Psychometric assessment of partner-controlling/-interfering scale.
  • Followback studies with the KCR.

Conceptual Model for VAW, Effects on Life Trajectories, and Cancer Care Continuum in a Social Environment Fostering Disparities

[D]

Discussion

• This model shows contributing factors but does not control for screening history. Screening history information will be sought in the KCR followback studies.
• HPV infection cannot be studied with cancer registry data. Kentucky Women's Health Registry data do show an association between violence and HPV.
• These studies parallel HIV findings.
• Future studies will examine data from Appalachia in greater detail.

Next HDIG Meeting

• Dr. Cathy J. Bradley will discuss "Inadequate Access to Surgeons: Cause for Disparate Care Among Dually Eligible Cancer Patients?" on Monday, September 21, 1 p.m. to 2:30 p.m., room 405, 6116 Executive Plaza, Rockville, MD.