Trock B, Rimer BK, King E, Balshem A, Cristinzio CS, Engstrom PF.
Impact of an HMO-based intervention to increase mammography intervention.
Cancer Epidemiol Biomarkers Prev 1993;2(2):151-6.

This quasi-experimental study assessed the impact on an HMO population of an educational intervention regarding mammography use. All HMO members were offered the intervention, and a randomly selected sample of HMO members were selected to participate in the study. Additionally, a random sample of non-HMO members from the community was selected as the "non-exposed" comparison group (18% of this group had no health insurance at all).

For two years after the introduction of the mammography intervention, HMO members were 1.4 times more likely to obtain a mammography than the non-HMO group. In multivariate analysis, while blacks and women with income less than $31,000 were more likely to obtain a mammogram, the effect of the intervention on obtaining a mammogram was lower for these groups. That is, the odds ratio (OR) for obtaining a mammogram among HMO members versus those not in the HMO was 3.6 for lower income whites and 1.6 for higher income whites; the OR was 1.5 for lower income blacks and 0.7 for higher income blacks.

However, the authors made more accurate estimates of the race and income effects by calculating relative risks, showing that, while the trends were similar, the differences were not statistically significant. That is, the RR for obtaining a mammogram among HMO members versus those not in the HMO was 1.9 for lower income whites (95% confidence interval=1.6, 2.2) and 1.2 for higher income whites (95% confidence interval=1.0, 1.4). The RR was 1.2 for lower income blacks (95% confidence interval=0.8, 1.7) and 1.0 for higher income blacks (95% confidence interval=0.6, 1.6).

The authors state that "within income strata, no differences in confounding factors were found to explain the different impact of the intervention between blacks and white. Factors examined included age, income, education, marital status, and concerns related to mammogram cost or radiation."

The authors suggested that the black-white differences might be due to "proximity to HMO mammography sites or physician communications." They provided data to demonstrate that "more white HMO members than black HMO members reported a previous physician recommendation for mammography (68% versus 55%), but no racial difference was seen for control women." The authors then concluded, "the identification of factors contributing to lower adherence in black women could permit the development of new intervention tailored to address those specific barriers (the current intervention was not so tailored)." Although all HMO members were sent packages advising them to obtain regular mammograms, no data were provided regarding potential race differences in recommendations/advice given by HMO physicians themselves.