Table 5.
Zero-inflated negative binomial model for the relationships between cPTG and clinical outcomes.
| Part 1: logistic part for excessive zero (i.e., having no SAE) | ||||
| Parameter | Estimate | Wald 95% Confidence Interval | P value | |
| Age at last Bevacizumab PX | 0.03 | (-0.01, 0.06) | 0.1153 | |
| Female vs Male | -0.78 | (-1.4, -0.10) | 0.0242 < 0.05 | |
| Race/Ethnicity | ||||
| Hispanic vs NHW | -0.55 | (-1.37, 0.27) | 0.1882 | |
| NHB vs NHW | -0.35 | (-1.12, 0.42) | 0.3767 | |
| Other vs NHW | 0.12 | (-4.09, 4.33) | 0.956 | |
| Unknown vs NHW | 0.53 | (-0.22, 1.28) | 0.1638 | |
| Follow up day | 0.00 | (-0.01, 0.00) | 0.1882 | |
| Number of Bevacizumab PXs | -0.05 | (-0.10, -0.01) | 0.0262 < 0.05 | |
| cPTG score | 1.30 | (-0.48, 3.08) | 0.1519 | |
| Part 2: negative binomial part | ||||
| Parameter | Estimate | Wald 95% Confidence Interval | P value | |
| Age at last Bevacizumab PX | -0.010 | (-0.028, 0.009) | 0.3038 | |
| Female vs Male | -0.225 | (-0.586, 0.137) | 0.223 | |
| Race/Ethnicity | ||||
| Hispanic vs NHW | -0.262 | (-0.713, 0.189) | 0.2542 | |
| NHB vs NHW | -0.078 | (-0.521, 0.364) | 0.729 | |
| Other vs NHW | 0.068 | (-2.834, 2.970) | 0.9633 | |
| Unknown vs NHW | 0.057 | (-0.450, 0.564) | 0.8249 | |
| Follow up day | 0.003 | (-0.001, 0.006) | 0.1047 | |
| Number of Bevacizumab PXs | 0.024 | (0.012, 0.037) | 0.0002 < 0.05 | |
| cPTG score | -1.079 | (-1.996, -0.162) | 0.0211 < 0.05 | |
| Dispersion | 5.421 | (3.561, 8.253) | ||