Category for Intercept Differences for Mean HbA1c | Fixed Effects | β Coefficient | SE (β) | P Value | Linear Contrast P Values |
---|---|---|---|---|---|
Node | Grand mean at baseline (intercept)† | 7.768 | 0.172 | <.001 | <.001 |
Mean HbA1c level difference, node 2 vs node 1 | −0.577 | 0.113 | <.001 | ||
Mean HbA1c level difference, node 3 vs node 1 | −0.391 | 0.232 | .123 | ||
Mean HbA1c level difference, node 4 vs node 1 | −0.387 | 0.083 | <.001 | ||
Age (at baseline) | Each additional year of age | −0.014 | 0.002 | .001 | NA |
Charlson index score | Each additional Charlson index score point | −0.091 | 0.031 | .011 | NA |
Sex | Female vs male | −0.231 | 0.040 | <.001 | NA |
Race/ethnicity | Hispanic vs non-Hispanic white | 0.137 | 0.112 | .243 | <.001 |
Non-Hispanic black vs non-Hispanic white | 0.314 | 0.090 | .004 | ||
Non-Hispanic Asian or Pacific Islander vs non-Hispanic white | 0.015 | 0.087 | .865 | ||
Non-Hispanic other vs non-Hispanic white | 0.393 | 0.128 | .009 | ||
Insurance status | Continually uninsured vs continually insured | 0.375 | 0.125 | .010 | .002 |
Insurance gaps vs continually insured | 0.133 | 0.051 | .022 | ||
Primary language | Spanish vs English | 0.035 | 0.145 | .812 | .009 |
Cantonese or Mandarin vs English | −0.213 | 0.086 | .028 | ||
Other language vs English | 0.079 | 0.104 | .457 | ||
Health care use‡ | 3 or 4 Visits vs 1 or 2 visits | 0.012 | 0.081 | .889 | .296 |
≥5 Visits vs 1 or 2 visits | 0.092 | 0.101 | .383 |
↵* This multilevel model predicted mean HbA1c level as a function of diabetes status, primary language, and insurance status, adjusting for age, Charlson index score, sex, race/ethnicity, and node.
↵† Assuming all other covariates equal zero.
↵‡ Average number of annual primary care visits.
NA, not available; SE, standard error.