TABLE 3

Results of Adjusted Logistic Regression Models Predicting Trends in Health Care Access for US Children, NHIS, 2000–2014

No Well-Child Visit, n= 176 524No Doctor Office Visit, n= 176 014No Dental Visit, n= 166 155No Usual Source of Care, n= 177 794Unmet Health Care Need, n= 177 578
Average Marginal Effect for Year (SE)aAverage Marginal Effect for Year (SE)aAverage Marginal Effect for Year (SE)aAverage Marginal Effect for Year (SE)aAverage Marginal Effect for Year (SE)a
Model 1bModel 2bModel 1bModel 2bModel 1bModel 2bModel 1bModel 2bModel 1bModel 2b
Overall−0.89c (0.03)−0.77c (0.03)−0.34c (0.02)−0.24c (0.02)−0.73c (0.03)−0.60c (0.03)−0.20c (0.02)−0.10c (0.02)−0.11c (0.02)−0.03 (0.02)
White−0.22c (0.03)−0.40c (0.04)−0.34c (0.04)−0.07c (0.02)−0.04 (0.02)
Black−0.37c,d (0.06)−1.06c,d (0.08)−0.94c,d (0.08)−0.19c,d (0.03)−0.11c (0.04)
Hispanic−0.59c,d (0.05)−1.43c,d (0.06)−1.14c,d (0.06)−0.44c,d (0.03)−0.18c,d (0.03)
Other−0.44c (0.09)−0.52c (0.11)−0.35c (0.11)−0.28c,d (0.06)−0.12c (0.06)
Poor−0.47c,e (0.05)−1.15c,e (0.07)−0.89c,e (0.06)−0.34c,e (0.04)−0.24c,e (0.04)−0.07e (0.04)
Near poor−0.44c,e (0.05)−0.97c,e (0.07)−0.76c,e (0.07)−0.28c,e (0.03)−0.21c,e (0.04)−0.09c,e (0.04)
Not poor−0.25c (0.03)−0.47c (0.04)−0.43c (0.04)−0.09c (0.02)0.00 (0.02)0.02 (0.02)
  • —, no significant interaction.

  • a Coefficients represent the average percentage point change per year adjusted for covariates. A quadratic term for year was included if nested models showed improvement in model fit via F tests. Year*race/ethnicity and year*income interactions were tested in separate models. Results by race/ethnicity and income are shown only if a global test of the interaction was significant (P < .05).

  • b Model 1 includes controls for child age, child sex, child race/ethnicity, family income, number of parents, household education, and global child health status. Model 2 includes these covariates plus child health insurance status.

  • c P < .05.

  • d Additional analyses from the statistical interaction model (not shown) revealed that this rate change was statistically different (P < .05) from the rate change for white children.

  • e Additional analyses from the statistical interaction model (not shown) revealed that this rate change was statistically different (P < .05) from the rate change for children in families classified as not poor.