TABLE 3.
12-Month Direct Health Care Cost Differences by Anti-CCP Categorya
Type of Encounter | Label | Beta Estimate | SE | P Value | Estimated Mean in Anti-CCP-Positive Patients, $ | Estimated Mean in Anti-CCP-Negative Patients, $ | Estimated Mean Differences in Anti-CCP-Positive and Anti-CPP-Negative Patientsb, $ |
---|---|---|---|---|---|---|---|
Office visit | Unadjusted | 0.05 | 0.04 | 0.176 | |||
PS weighted | 0.06 | 0.04 | 0.138 | 338 | 326 | 12 | |
Rheumatologist visit | Unadjusted | 0.26 | 0.07 | 0.0001 | |||
PS weighted | 0.29 | 0.07 | < 0.0001 | 298 | 222 | 76c | |
Laboratory testing | Unadjusted | 0.22 | 0.04 | < 0.0001 | |||
PS weighted | 0.23 | 0.04 | < 0.0001 | 398 | 215 | 183c | |
Prescription | Unadjusted | 0.43 | 0.07 | < 0.0001 | |||
PS weighted | 0.46 | 0.08 | < 0.0001 | 6,671 | 4,172 | 2,499c | |
ED | Unadjusted | -0.11 | 0.21 | 0.607 | |||
PS weighted | 0.06 | 0.23 | 0.809 | 1 | -3 | 3 | |
Hospitalization | Unadjusted | -0.20 | 0.28 | 0.467 | |||
PS weighted | -0.26 | 0.31 | 0.406 | 52 | 654 | -602 | |
Total | Unadjusted | 0.24 | 0.08 | 0.003 | |||
PS weighted | 0.25 | 0.08 | 0.001 | 7,465 | 5,302 | 2,163c |
Note: These results are from the generalized linear models using difference-in-difference unadjusted and propensity score methods. Variables used in the propensity score models and regression models were age, gender, race/ethnicity, and Elixhauser Comorbidity Index score.
a Cost differences are shown in 2015 U.S. dollars.
b Estimated mean changes were calculated using recycled prediction methods after the generalized linear models using gamma distribution, log link.
c Statistically significant findings (P < 0.05).
CCP = cyclic citrullinated peptide; ED = emergency department; PS = propensity score; SE = standard error.