W. Dale et al., PROSTATE-CANCER, RACE, AND SOCIOECONOMIC-STATUS - INADEQUATE ADJUSTMENT FOR SOCIAL-FACTORS IN ASSESSING RACIAL-DIFFERENCES, The Prostate, 29(5), 1996, pp. 271-281
BACKGROUND. This paper reviews the state of the art in analyzing race,
social factors, and economic factors in cancer research, with an emph
asis on prostate cancer and the role of socioeconomic status (SES) in
racial differences in mortality. It analyzes the quality of articles i
n the literature that assess the role of SES in cancer mortality. METH
ODS. English-language titles were identified using MEDLINE with public
ation dates from mid-1985 through July 1994. Articles in the reference
s of these articles were also included in the final selection, based o
riginally on title and ultimately on content, dating back to 1978. Art
icles that included SES information and distinguished between whites a
nd African-Americans were chosen, resulting in a final selection of 21
articles. Articles are summarized with consideration of five criteria
considered minimal requirements of a well-designed study of the role
of race in cancer mortality: (1) SES measure(s) should be on an indivi
dual level, not census level; (2) SES should be controlled for when ma
king comparisons between whites and blacks; (3) SES should include at
least (individual level) measures of income and education; (4) sample
sizes are sufficient for the relevant populations; and (5) specific ca
ncer sites should be studied separately. RESULTS. Of the articles revi
ewed, only two meet the minimum standards, neither of which studied pr
ostate cancer. It is not clear whether observed racial differences in
prostate cancer are directly attributable to race or reflect underlyin
g social factor differences between whites and African-Americans. CONC
LUSIONS. In the future, specific characteristics of SES should be meas
ured at the individual level; there is a need for expansion and standa
rdization of data in terms of social and economic content; other metho
dological advances are needed in modeling to take into consideration t
he influence of SES in outcome research related to cancer. (C) 1996 Wi
ley-Liss, Inc.