Fig. 1From: Expenditure variations analysis using residuals for identifying high health care utilizers in a state Medicaid programQuantile-to-quantile plot (Q-Q plot) of standard normal distribution and residuals obtained from the linear regression model for year 2014. In the left panel, the dependent variable, per-member per-month expenditures, is not transformed. Due to its strong skewness, the residuals significantly deviate from normal distribution. In the right panel, we log-transform the dependent variable with base 10 and re-run the model. The residuals distribution better fits a normal distribution. The right and left ends are heavy-tailed, indicating over/under utilizationBack to article page