Table 6 Adjusted Odds ratios from binary logistic regression models of obesity and diabetes in sensitivity analyses to time change between 2011 and 2017.

From: Population mobility data provides meaningful indicators of fast food intake and diet-related diseases in diverse populations

 

Full sample

 

Method 1

 

Method 2

 

Model

AOR of obesity (95% CI)

AOR of diabetes (95% CI)

AOR of obesity (95% CI)

AOR of diabetes (95% CI)

AOR of obesity (95% CI)

AOR of diabetes (95% CI)

FF visits/time

1.16 (1.12, 1.21)

1.15 (1.08, 1.21)

1.17 (1.12, 1.22)

1.17 (1.09, 1.24)

1.17 (1.12, 1.22)

1.16 (1.09, 1.23)

FF visits/food

1.13 (1.10, 1.17)

1.11 (1.06, 1.16)

1.14 (1.09, 1.18)

1.12 (1.06, 1.17)

1.14 (1.10, 1.18)

1.12 (1.06, 1.17)

  1. P < 0.001 for all estimated odds ratios. Each model includes the variable listed in the Model column as the primary independent variable. All models adjusted for age group, gender, race and ethnicity, educational level, and household income level. Sensitivity tests fit regression models to the analytic sample subtracting out respondents living in outlier census tracts demonstrating the largest change in demographic variables between 2011 and 2017 according to the two methods. Method 1: Outliers are identified by the Tukey method as values >1.5 times the interquartile range from each of the quartiles for a variable, i.e. upper outliers are values of the distribution >Q3 + 1.5 × IQR and lower outliers are values <Q1–1.5 × IQR. Method 2: Outliers are identified as values of a variable above and below 2 standard deviations of the mean. For each method, we remove from the LACHS sample all respondents living in outlier census tracts. We then identified the union over outlier census tracts across the three measures, as some census tracts overlapped. Regression model results fit to the full analytic sample (full sample) are provided for comparison. AOR adjusted odds ratio, CI confidence interval, FF fast food.
Back to article page