A combined multivariate model associating PhA and SAPSII improves 28-day mortality predictability.
Sensitivity, specificity, PPV and NPV of a multivariate model associating d1 PhA < 3.49° and SAPSII ≥ 46 for 28-day mortality were 77, 66, 32 and 93%%, respectively (AUC = 0.78 [0.78–0.78]).
Confounding factors were kept in the multivariate model when associated with the outcome at a significance level lower than 20%.
In the multivariate model, factors associated with consulting a GP were mobility problems (OR 2.62 (1.64-4.17)), urban living (OR 2.40 (1.14-5.04) and pain severity (1.28 (1.28-1.28)).
In the multivariate model, factors associated with unprotected sex in the 6 months following the assault included unprotected sex in the 6 months prior to the assault (OR = 6.46, 95% CI = 3.04, 13.74), time since the assault (per 30-day increments, OR = 1.33, 95% CI = 1.12, 1.57) and age (per 5-year increments OR = 1.30, 95% CI = 1.08, 1.57, Table 6).
We chose variables for multivariate modeling if associated with mortality in univariate Cox regression analyses (P < 0.05) or if a factor had been previously reported to be associated with the risk of death in cystic fibrosis, e.g., sex and S. aureus.
The strongest reduction in risk, which could be observed in both univariate and multivariate models, was associated with a smoking history of 40 or more years compared with having never smoked (RR = 0.37, 95% CI = 0.15 0.90).
In multivariate model, independent variables associated with anemia were moderate disease activity and CRP.
It is of note that ventilatory monitored variables included in the multivariate model were not associated with mortality.
The moderate discriminative value (AUC = 0.69) of the multivariate model describing factors associated with precautionary behaviour however suggests that other, unmeasured factors may be of additional importance.
