Figure 3—

Mean excess medical expenditures during the 10 years before and after the index date by medical expenditure component. The expenditures in follow-up year 1 also include the cost incurred on the day of diagnosis if diagnosed in outpatient setting or the cost incurred for whole admission if diagnosed in inpatient setting. The control cohort includes people not diagnosed with diabetes during the study period. Excess = average expenditures for diabetes cohort – average expenditures for control cohort. All the excess medical expenditures are statistically significant (P < 0.05, paired Student t test), except for the outpatient and inpatient expenditures in follow-up year −10.