Table 5. Multivariate analysis of the effects of demographic and clinical variables and treatment strategy on other health resource use costs over 4 years.
| Complete case analysis(n = 548) | MCMC imputation model(n = 777) | |||
| Variables | Rate Ratio (95% CI) | P-value | Rate Ratio (95% CI) | P-value |
| Later year biologic use | 1.64 (1.32–2.05) | <0.0001 | 1.73 (1.28–2.18) | 0.0016 |
| HAQ score at baseline: >2 | 2.43 (1.87–3.17) | <0.0001 | 2.66 (1.87–3.45) | <0.0001 |
| HAQ score at baseline: 1 to ≤2 | 1.48 (1.24–1.76) | <0.0001 | 1.51 (1.22–1.81) | 0.0007 |
| Variation in HAQ score (per 0.25-point increase) between baseline and 6 month visit | 1.06 (1.03–1.10) | 0.000 | 1.08 (1.04–1.11) | <0.0001 |
| Doctor's evaluation of RA diagnostic <50% at baseline | 0.76 (0.64–0.92) | 0.004 | 0.82 (0.67–0.96) | 0.0114 |
| Hospitalisation at baseline | 1.38 (1.16–1.64) | 0.000 | 1.29 (1.06–1.51) | 0.0113 |
| Lives with a partner at baseline | 0.65 (0.55–0.77) | <0.0001 | 0.72 (0.60–0.84) | <0.0001 |
| Age (per 10-year increase) at baseline | – | – | 1.10 (1.03–1.17) | 0.0033 |
| Proportion of variance explained by model | 22% | 17% (16%–19%)1 | ||
| ICC | 4.85% | 5.85% (4.21%–8.02%)1 | ||
The reported Rate Ratios describe the variations in costs expressed as a multiplicative factor for patients presenting the associated characteristic compared to those who did not, all other things being equal.
MCMC = Markov Chain Monte Carlo; CI = Confidence Intervalle; HAQ = Health Assessment Questionnaire; RA = Rheumatoid Arthritis; ICC = Intraclass Correlation Coefficient.
1
Median and range for all 24 imputed datasets.