Similar differences in treatment by age have been reported from other studies (Farrow et al., 1992; Greenfield et al., 1987; Lazovich et al., 1991).
Not all the evidence points to underuse of treatment in older people, however. One of the more thorough studies, of postmenopausal women with early breast cancer treated in 1993 in Minnesota, found that most (92 percent) women with node-positive breast cancer received some form of adjuvant therapy (Guadagnoli et al., 1997). The likelihood of treatment with adjuvant therapy did decline slightly with age, but the decline was not statistically significant. The use of adjuvant therapy was less frequent in women with node-negative breast cancer and did decline with age, but the age-associated differences were not significant after adjusting for various demographic and disease-associated factors.
Differences in treatment by race have been well documented: African-American patients are less likely than white patients to undergo surgical resection for colorectal cancer (Cooper et al., 1996), to receive bone marrow transplantation for leukemia or lymphoma (Mitchell et al., 1997), to receive radical prostatectomy and radiation for localized prostate cancer (Harlan et al., 1995), or to have breast conserving surgery (BCS) for breast cancer or receive radiation therapy following BCS (Farrow et al., 1992; Muss et al., 1992; Nattinger et al., 1992). However, it appears that these effects may actually be more closely related to social class than to race.
In one study, elderly residents of areas characterized by low compared to high educational attainment were more likely to have received no treatment for non-small-cell lung cancer and, when treated, to receive radiation instead of surgical therapy, despite having similar clinical profiles (Smith et al., 1995). In another study, low educational attainment and a high percentage of the population with poverty-level incomes were associated with lower rates of BCS for women with breast cancer (Samet et al., 1994).
There are few large differences in survival among cancer patients by gender and studies of patterns of treatment by gender for bladder or colorectal cancer, leukemia, and lymphoma do not suggest any differences in care for men and women (Harris et al., 1997; Mitchell et al., 1997).
Physicians' treatment recommendations are influenced by a number of factors, including physician age (Liberati et al., 1987), gender (GIVIO, 1988), specialty (Deber and Thompson, 1987), and belief in efficacy of care (Liberati et al., 1987). The content of physician communication also varies according to patient characteristics, including age, income, education, race or ethnicity, and expected prognosis (Waitzkin, 1985).
Variations in the use of breast conserving surgery (BCS) instead of mastectomy for patients with early breast cancer may, in part, be explained by physician specialty, training, and experience. When asked about treatment preferences, medical oncologists were more likely than surgeons to prefer BCS (Deber and Thompson, 1987), surgeons were more likely than primary care physicians to prefer BCS, and among surgeons, those with postgraduate specialty training in