Logistic regression analysis of determinants for changing (starting and stopping) work participation after 2 years of tumor necrosis factor inhibitor (TNFi) treatment.
| Beta (SE) | OR | p | |
|---|---|---|---|
| Starting model (R2 = 0.143) | |||
| HAQ-DI baseline | −0.55 (0.33) | 0.58 | 0.094 |
| Age | −0.05 (0.02) | 0.96 | 0.005 |
| Rheumatoid factor + | 1.01 (0.44) | 2.73 | 0.023 |
| Constant | 1.26 (0.91) | 3.52 | 0.165 |
| Stopping model (R2 = 0.125) | |||
| HAQ-DI baseline | 1.01 (0.31) | 2.74 | 0.001 |
| Comorbidity | −0.98 (0.37) | 2.67 | 0.007 |
| HAQ-DI response* | −1.16 (0.38) | 0.31 | 0.002 |
| Constant | 2.22 (0.39) | 0.11 | 0 |
p < 0.10 are included in the model.
↵* HAQ response = HAQ baseline – HAQ at 6 month (a positive outcome means an improvement in functionality). R2 = percent explained variance final model (Nagelkerke). HAQ-DI: Health Assessment Questionnaire-Disability Index. The process from the starting model to the presented final model is available from the authors on request.