TABLE 2

Ordered Logit Models for Use of DBS and Teas Among Infants As a Function of Mother's Characteristics (n = 2653)

VariableDBS and Teas Use, Adjusted Odds RatioP95% Confidence Interval
Intercept 1<.0001
Intercept 2<.0001
Maternal herbal use before or during pregnancy3.69<.00012.78–4.89
Parity
    Primiparousa
    Multiparous0.74.0840.53–1.04
Breastfeeding duration, wk1.02<.00011.01–1.02
Mother's age
    18–24a
    25–291.44.1150.92–2.28
    30–341.21.4550.73–1.99
    ≥351.52.1240.89–2.59
Mother's education
    High school or lessa
    Some college1.19.4170.78–1.82
    College graduate or more0.87.5420.55–1.38
Poverty level
    Under 100a
    100–1841.20.4940.71–2.01
    185–3491.50.1170.90–2.47
    ≥3501.69.0660.97–2.97
Race
    Non-Hispanic whitea
    Non-Hispanic black1.19.6370.58–2.46
    Non-Hispanic other0.90.7470.46–1.74
    Hispanic1.85.0161.12–3.06
Married1.30.2240.85–2.00
Population density
    Rurala
    Suburban0.98.9170.66–1.45
    Urban1.02.9460.66–1.55
Region
    Northeasta
    Midwest0.92.6990.58–1.44
    South1.22.3650.79–1.88
    West1.39.1430.89–2.16
        Model χ2161.46
        df20
        Score testb20.20
  • Data were obtained from the Infant Feeding Practices Study II.

  • a Reference categories.

  • b The score test evaluates the null hypothesis of no difference in the regressor effects for the ordinal categories using a χ2 distribution. A value of 20.20 with 20 degrees of freedom (df) fails to reject the null hypothesis, indicating that the proportional odds assumption is reasonable.