Influencing factors on costs of care at baseline
| Variable | One-part model | Two-part model | ||||||||
| Total costs | Direct medical costs | Non-medical costs | Indirect costs | |||||||
| Utilisation probability | Costs per user | |||||||||
| Impact | P-value | Impact | P-value | Impact | P-value | OR | P-value | Impact | P-value | |
| Age (years) | 0.899 | <0.001 | 0.882 | <0.001 | 0.931 | <0.001 | 0.891 | <0.001 | 0.981 | 0.088 |
| Sex (female) | 1.16 | 0.371 | 1.22 | 0.481 | 1.61 | 0.171 | 1.12 | 0.309 | 0.905 | 0.363 |
| Category (B) | 1.56 | <0.001 | 1.77 | 0.014 | 1.52 | 0.053 | 1.39 | 0.002 | 1.02 | 0.846 |
| Other countries* | 0.713 | 0.002 | 0.672 | 0.051 | 0.901 | 0.607 | 0.448 | <0.001 | 0.916 | 0.643 |
The estimates (exp(β)) are the results of GLMs using gamma regression and log-link function, adjusted for age, gender, country of treatment and disease category.
The reference categories were defined as follows: female versus male, category (A) versus category (B), other countries versus Germany
*UK, Poland, Turkey, Italy, Denmark, Switzerland, Spain, Hungary and Austria.
GLM, generalised linear model.