Table 5.
Odds ratio (standard error) estimates for the association between multiple myeloma and IL-6ra using conditional likelihood for the individually matched and unconditional likelihood after breaking the match, Los-Angeles County, 1999–2002.
| Variable | Values | CCA | EM | MI | LLMI | LCMI-3 | LCMI-2 | ||
|---|---|---|---|---|---|---|---|---|---|
| Conditional likelihood: odds ratio and SE estimates | |||||||||
| IL6174R | DD | 1.00(–) | 1.00(–) | 1.00(–) | 1.00(–) | 1.00(–) | 1.00(–) | ||
| AD | 1.60(0.45) | 1.93(0.37) | 1.92(0.23) | 1.91(0.23) | 2.03(0.22) | 1.95(0.22) | |||
| AA | 1.12(0.71) | 1.70(0.62) | 1.25(0.39) | 1.25(0.40) | 1.23(0.37) | 1.51(0.36) | |||
| Gender | Male | 2.23(0.44) | 2.59(0.38) | 4.44(0.21) | 4.40(0.22) | 4.48(0.21) | 4.48(0.21) | ||
| Age | Age > 60 | 1.15(0.64) | 2.25(0.56) | 4.66(0.33) | 4.60(0.37) | 4.48(0.32) | 4.48(0.33) | ||
| BMI | BMI | 1.25(0.26) | 1.27(0.21) | 1.39(0.13) | 1.38(0.13) | 1.40(0.13) | 1.39(0.12) | ||
| EDUC | EDUC | 1.38(0.27) | 1.23(0.22) | 1.46(0.13) | 1.46(0.13) | 1.43(0.13) | 1.43(0.13) | ||
| Unconditional likelihood: odds ratio and SE estimates | |||||||||
| IL6174R | DD | 1.00(–) | 1.00(–) | 1.00(–) | 1.00(–) | 1.00(–) | 1.00(–) | ||
| AD | 1.46(0.32) | 1.48(0.28) | 1.39(0.14) | 1.39(0.14) | 1.40(0.13) | 1.38(0.13) | |||
| AA | 0.79(0.49) | 0.97(0.45) | 0.75(0.23) | 0.72(0.24) | 0.71(0.22) | 0.79(0.21) | |||
| Gender | Male | 1.67(0.29) | 1.72(0.26) | 1.63(0.12) | 1.63(0.12) | 1.63(0.12) | 1.65(0.12) | ||
| Race | Black | 1.01(0.36) | 1.04(0.32) | 0.99(0.15) | 1.00(0.16) | 0.97(0.14) | 0.98(0.14) | ||
| Age | Age > 60 | 1.09(0.29) | 1.12(0.26) | 1.15(0.12) | 1.16(0.12) | 1.15(0.12) | 1.15(0.12) | ||
| BMI | BMI | 1.12(0.18) | 1.17(0.17) | 1.16(0.08) | 1.16(0.09) | 1.16(0.08) | 1.16(0.07) | ||
| EDUC | EDUC | 1.12(0.18) | 1.08(0.16) | 1.05(0.07) | 1.06(0.07) | 1.05(0.07) | 1.04(0.07) | ||
SE=asymptotic standard error; CCA=complete case analysis using conditional logistic; MI=multiple imputation with discriminant method; LLMI= log-linear multiple imputation; EM= Expectation Maximization algorithm; LCMI-2=latent class multiple imputation with two classes; LCMI-3=latent class multiple imputation with three classes.