Ethnic minority status and breast cancer screening practices:

1
Ethnic minority status and breast cancer
screening practices A meta-analytic review
Table 2. Relationship Between Race and
Mammography Screening
Rebecca Purc-Stephenson Kevin M. Gorey,
University of Windsor
Abstract Regular use of breast cancer screening
has been shown to be associated with reduced
breast cancer incidence among women. Although
rates of cancer screening have increased
dramatically in the United States and Canada,
ethnic minority women are at particular risk for
not adhering to recommended cancer screening
guidelines. While it is widely cited that ethnic
minority women are less likely be screened for
breast cancer than their non-Hispanic white
counterparts, estimates of breast cancer
screening among ethnic minority women show a
great deal of variability across studies. This
study explores the association of breast cancer
screening with ethnic minority status in women by
integrating the findings of existing studies
using meta-analytic techniques. A total of 72
conceptually and empirically relevant studies
that compared breast cancer screening practices
among ethnic minority (i.e., African American,
Hispanic and/or Asian) women and non-Hispanic
white women were selected for systematic review
from social scientific and bio-medical research
literature databases (1975-2005). Preliminary
analyses indicate differential rates of
mammography compliance among African American,
Hispanic, Asian/Pacific Islander, and Native
American/Alaskan Native and non-Hispanic white
women aged 40 years and older. In addition to
observation of such main effects, this review
explores their moderation by key characteristics
of study participants, methods and contexts.
Table 1 Descriptive Characteristics of the 72
Studies Included in the Meta-Analysis Percentage
Distributions
Introduction Breast cancer is the most common
form of cancer affecting women worldwide and is
the major killer of women between the ages of 35
and 55 (Baum, 1988). However, vulnerable
populations such as racial ethnic minorities, the
elderly, and the poor experience a
disproportionate burden of breast cancer (Glanz,
Croyle, Chollette, Pinn, 2003 Gotzsche
Olsen, 2000). Research over the past two decades
has shown that whereas ethnic minority women have
a lower overall incidence of breast cancer, they
experience later stage at diagnosis, greater
incidence for multiple cancer sites, differential
treatments, greater mortality and morbidity than
do their white non-Hispanic counterparts
(Ashing-Giwa, 2005 Heeden White, 2001 Lannin,
et al., 1998 Randolph, Goodwin, Mahnken,
Freeman, 2002), presumably due to less screening
mammography (Heeden White, 2001 Lannin et al.,
1998). The objective of this review is to
empirically investigate the relationship between
screening mammography and ethnicity by
integrating the findings of existing studies
using meta-analytic techniques.

• Method
• In February 2006, the following peer-reviewed
  research literature databases were searched
  (1975-2005) ERIC, PsychINFO, Medline, Social
  Services Abstracts, and Sociological Abstracts.
• We used the key word search terms (ethnic,
  minorit, race) and (breast cancer, breast
  neoplasm, breast carcinoma) and (screen,
  preventi, mammography), augmented by
  bibliographic reviews of retrieved manuscripts
  and previous published reviews 123 conceptually
  relevant articles were uncovered.
• After applying our inclusion and exclusion
  criteria, 72 relevant studies were retained.
• Meta-Analysis
• The selected sample of 72 studies approximated a
  case-control search design, with 17 studies
  observing if respondents had ever had a
  mammography, and 69 studies observing if
  respondents had obtained a mammography in the
  past two years (14 studies included both
  observations).
• The odds ratio, which is an estimate of the
  relative odds of an ethnic minority women having
  a mammography versus white non-Hispanic women
  having a mammography, was selected as this
  studys central meta-analytic statistic (Cooper,
  1998 Greenland, 1987).
• The Mantel-Haenszel procedure was used to
  calculate a weighted OR from the individual
  studies, with 95 percent confidence intervals
  around the primary study as well (Mantel
  Haenszel, 1959).
• Separate pooled analyses were conducted for each
  ethnic group comparison for each aspect of
  mammography under study (ever had a mammogram
  and had a mammogram in the past two years)
  African American (14 and 69 outcomes), Hispanic
  (13 and 39 outcomes), Asian/Pacific Islander (5
  and 22 outcomes), and Native American/Alaskan
  Native (0 and 5 outcomes).
• As a control for publication bias, one-tailed
  fail-safe Ns were computed for each domain.
• Tests of Moderation
• Heterogeneity analyses were conducted using
  Cochrans Q.
• If significant heterogeneity was observed (p lt
  .05), its potential sources were systematically
  explored.

• Results
• Sample Description
• Descriptive characteristics are presented in
  Table 1.
• Of the 72 studies, 57 compared non-Hispanic
  whites to African Americans, 37 to Hispanics, 16
  to Asian/Pacific Islanders, and 5 to Native
  American/Alaskan Natives.
• 36.1 of the studies used population-based
  surveys (36.1), with the National Health
  Interview Survey the most common type (17.8).
• 50.8 of the studies collected data from
  urban/metropolitan regions few focused on women
  from rural regions (8.2).
• Meta Analysis
• The results from the meta-analysis are presented
  in Table 2.
• The results show that, compared to non-Hispanic
  white women, fewer ethnic minority women have
• ever had a screening mammography
• had a mammogram in the past two years.
• These differences tend to
• be larger across age categories
• be larger in Southern US compared to the Midwest
  or Northeast US.
• The differences in screening appear to decrease
  over time.

• Discussion
• Systematically integrating 72 studies on
  screening mammography, this meta-analysis found
  consistent and meaningful racial differences.
• Overall, the results found that ethnic minority
  women get screening mammography less often
  compared to non-Hispanic white women.
• There was significant heterogeneity in most of
  our pooled samples which we attempted to identify
  through tests of moderation.

Note. AA African American His Hispanic
As/PI Asian/Pacific Islander Na/AN Native
American/Alaskan Native. OR Odds ratio CI
Confidence intervals. p-value for heterogeneity
between studies, Cochrans Q.