Skip to main content
. 2005 Feb;95(2):305–311. doi: 10.2105/AJPH.2002.012567

TABLE 4—

Logistic Regression Analysis for Completion of Immunization in Infants Aged 1 Year in the National Maternal and Infant Health Survey Sample Before and After Implementation of National Health Insurance (NHI): Taiwan, 1989 and 1996

Pre-NHIa Post-NHIb
Variable Coefficient P Value Odds Ratio Coefficient P Value Odds Ratio
Constant 0.4270 0.273 . . . 0.7512 0.024** . . .
Infant’s characteristics
    Gender 0.0340 0.768 1.0346 −0.1545 0.160 0.8568
    Gestation period −0.0747 0.827 0.9281 0.4814 0.015** 1.6183
    Birth order 0.2607 0.044** 1.2978 0.5001 0.001*** 1.6489
    Hospital birth −0.2405 0.054* 0.7863 0.2352 0.039** 1.2652
    Well-baby care −0.1367 0.257 0.8722 0.4169 0.001*** 1.5172
    Principal caregiver
        Grandparents −0.0136 0.930 0.9865 0.2890 0.839 1.0293
        Babysitter −0.3831 0.048** 0.6817 −0.2677 0.122 0.7651
Maternal characteristics
    Age, y
        25–34 0.1687 0.217 1.1837 0.0356 0.803 1.0363
        ≥35 −0.1284 0.673 0.8795 −0.1867 0.408 0.8297
    Education
        Junior high 0.2365 0.192 1.2669 0.2448 0.325 1.2774
        Senior high 0.2800 0.105 1.3232 0.3760 0.110 1.4564
        Junior college −0.2284 0.381 0.7958 0.3368 0.224 1.4005
        College and above −0.2882 0.358 0.7496 0.6395 0.047** 1.8955
    Occupation
        Government 0.6462 0.023** 1.9083 0.0454 0.824 1.0464
        Laborer 0.2904 0.120 1.3369 0.1302 0.392 1.1390
        Farmerc 0.3495 0.622 1.4184 . . . . . . . . .
        Businesswoman 0.3339 0.101 1.3964 0.0144 0.932 1.0145
        Other 0.2524 0.334 1.2871 −0.0026 0.992 0.9974
Geographic characteristics
    Central 0.1603 0.277 1.1739 −0.0144 0.918 0.9857
    South −0.1412 0.345 0.8683 −0.1033 0.465 0.9019
    East 0.3129 0.228 1.3674 −0.1780 0.388 0.8369

aLog likelihood = −884.1923; LR χ2 (21) = 33.9; Prob > χ2 = .036; pseudo R2 = .018; n = 1398.

bLog likelihood = −1154.7057; LR χ2 (20) = 67.1; Prob > χ2 = .0001; pseudo R2 = .028; n = 3171.

c In the post-NHI period farmer ≠ 0 predicts success perfectly. The dummy variable-farmer was dropped and 13 observations were not used.

*P < .1; **P < .05; ***P < .01.