Table 6.

NSABP DMP-1: Final multivariate models of participants' characteristics associated with the decision about taking SERMs for breast cancer risk reduction

CharacteristicsORa(95% CI)PbORa(95% CI)PbORa(95% CI)Pb
Recommendation from HCP
 No recommendation/not to take a SERM<0.001<0.001<0.001
 To take a SERM14.00 (8.39–23.37)17.73 (6.27–50.14)13.32 (7.43–23.90)
Nervous about taking a drug that no one she knows has taken
 Knows someone who took drug<0.0010.002<0.001
 Not nervous1.43 (0.83–2.48)0.80 (0.28–2.30)1.78 (0.93–3.40)
 Nervous somehow1.81 (0.91–3.60)0.88 (0.23–3.40)2.62 (1.18–5.80)
 Nervous a lot0.31 (0.17–0.56)0.16 (0.06–0.43)0.37 (0.18–0.79)
Attitudes about taking medication scalec1.05 (1.02–1.09)0.0041.06 (0.996–1.13)0.071.05 (1.00–1.09)0.03
Worried about BC in the next 5 years
 Not worried<0.0010.10<0.001
 Worried somehow1.56 (0.85–2.86)2.52 (0.78–8.16)1.40 (0.70–2.82)
 Worried a lot3.61 (2.04–6.38)3.14 (1.09–9.02)4.15 (2.14–8.05)
Trust in the HCP scale1.06 (1.02–1.10)0.0031.05 (0.99–1.13)0.111.05 (1.00–1.10)0.053
Knows someone with good experience who took SERMs
 Yes2.06 (1.30–3.26)1.02 (0.43–2.41)2.11 (1.21–3.68)
Knows someone with bad experience who took SERMs
 Yes0.38 (0.22–0.64)0.39 (0.14–1.08)0.27 (0.14–0.52)
Anyone in the family ever had blood clotting problems?
 Yes0.59 (0.38–0.92)0.27 (0.11–0.69)0.67 (0.40–1.13)
“Good” traitsd,e
 Yes7.57 (3.41–16.78)62.19 (7.4–520.84)
Talked about testing results
 Yes1.80 (1.13–2.86)2.40 (1.07–5.41)
 No breast biopsies0.010.310.20
 No breast biopsies with AH or status of AH unknown0.66 (0.35–1.27)1.30 (0.39–4.33)0.61 (0.29–1.30)
 At least one biopsy with AH1.48 (0.85–2.57)2.07 (0.75–5.75)1.09 (0.56–2.11)
Anyone in the family ever had endometrial cancer?
 Yes0.79 (0.39–1.63)0.18 (0.04–0.87)1.25 (0.56–2.79)
Had you ever thought about taking SERM before?
 Yes1.14 (0.73–1.77)2.49 (1.15–5.40)1.03 (0.59–1.77)
“Good” traitsd,e/Talked about testing results
 No  No0.006
    Yes0.17 (0.03–0.96)
 Yes  No
    Yes2.19 (1.25–3.83)
  • Abbreviations: AH, Atypical hyperplasia; BC, breast cancer.

  • aOR, the relative odds for SERM uptake, “—” identifies the baseline level of the covariate.

  • bP value is based on the likelihood ratio test, unknowns were excluded for calculation of P values.

  • cItalics indicate that these covariates were not in the final model, however, their effect and significance were tested by adding them into the model one at a time.

  • d“Good” traits = Average woman who takes a drug to prevent breast cancer perceived as: brave, or smart, or healthy, or takes good care of her health.

  • eThe postmenopausal model interaction between “Good traits” and “Talked about testing results” was found to be statistically significant. The corresponding “main effects” section of the table for these two covariates is shown by dashes to indicate that they are not applicable for the postmenopausal model. The corresponding “interaction” section of the table for these two variables is shown by dashes to indicate that it is not applicable for the overall and premenopausal models.