Table 3.
Summary of logistic regression models relative to exacerbation with coefficient odds ratios, 95% confidence intervals and other model measures. Each column represents the regression coefficients for each of the 3 models
| Variable | miR modela | Clinical model | miR + Clinical model |
|---|---|---|---|
| hsa-miR-206b | 0.64**$ (0.45,0.89) | __ | 0.65** (0.44, 0.92) |
| hsa-miR-146b | 0.72** (0.52,0.98) | __ | 0.67** (0.47,0.93) |
| hsa-miR-720 | 0.75 (0.51, 1.1) | __ | 0.70* (0.46,1.03) |
| Clinical score | __ | 1.36*** (1.14,1.64) | 1.38*** (1.15,1.69) |
| Age | __ | 1.01 (0.84,1.23) | 0.97 (0.80,1.20) |
| Sex | __ | 0.79 (0.36,1.73) | 0.70 (0.29,1.65) |
| Observations | 153 | 153 | 153 |
| Log Likelihood | −76.8 | −79.5 | −69.9 |
| Akaike Information Criterion (AIC) | 161.6 | 167 | 153.7 |
| Hosmer-Lemeshow Goodness of Fit Test^ | Χ2 = 9.15 p-value = 0.33 |
Χ2 = 8.45 p-value = 0.39 |
Χ2 = 7.07 p-value = 0.53 |
aThe miRNA model was determined by backwards stepwise selection. The clinical score model was adjusted for age and sex per original publication
bEach of the model coefficients presents the odds ratio and the 95% confidence interval in parenthesis
P-values are marked by an asterisk (*) with: *p < 0.1, **p < 0.05, ***p < 0.01
^10 bins used to calculate quantiles with degree of freedom = 8