Figure 1.

The flowchart of this study. Clinical diagnoses: upon admission, the patient was diagnosed by the clinician with organ infection or sepsis. SIRS: systemic inflammatory response syndrome, defined as the presence of two or more of the following: (1) temperature <36 °C or >38 °C; (2) heart rate >90 beats/min; (3) respiratory rate >20 breaths/min or PaCO2 < 32 mmHg; and (4) white blood cell count≥12,000 cells/mm3 or≤4,000 cells/mm3. SOFA: based on six different scores, one for each of the respiratory, cardiovascular, hepatic, coagulation, renal and neurological systems and allocates a score of 0-4 with an increasing score reflecting worsening organ dysfunction.

Table 1.

Logistics regression analysis of predictive markers for sepsis

Figure 2.

The nomogram of the presdictive model of sepsis occurrence. The nomogram was developed by visualizing graphical representations, wherein each factor was assigned a score based on its respective value. Subsequently, the total score was calculated by summing the individual scores, enabling the assessment of clinical risks. For instance, a patient admitted to the hospital presented with a lactate level of 1.94 mmol/L and a procalcitonin (PCT) level of 23 ng/mL without mechanical ventilation. Additionally, coagulation parameters revealed that fibrin degradation products (FDP) levels at 8.1 mg/L and D-dimer levels at 3 mg/L. These parameters were assigned β (X-m) scores of 0, 0.66, 0, -1.85 and -2.52, respectively. The final total score calculated amounted to -3.71 which corresponds to a predicted probability of sepsis at approximately 82.8%.

Figure 3.

Receiver operating characteristic curve analysis of the sepsis prediction model. AUC: area under the curve; CI: confidence intervals.

Figure 4.

The calibration plot of the sepsis prediction model. The dot line represents the model’s predicted fit, while the solid line represents the bias-corrected fit from 40 bootstrap resamples. The proximity of these curves to the reference line (dashed line) indicates strong alignment between predicted and actual values.

Figure 5.

Decision curve analysis (DCA) curve of the sepsis prediction model. The X-axis represents threshold probability, while the Y-axis represents the standard net benefit. The All line represents the net benefit of predicting that all patients will develop sepsis, while the None line represents the net benefit of predicting that none of the patients will develop sepsis. The decision curve for the sepsis prediction model lies between the All and None line and above the All line, indicating that the model provides an enhanced clinical decision-making.

Figure 6.

Receiver operating characteristic curves for 10-fold cross-calidation of the sepsis prediction model. AUC: area under the curve.

Table 2.

Cox regression analysis for the mortality of sepsis