OBJECTIVE To explore the influencing factors and risk prediction model for prolonged hospital stay in patients with maxillofacial and cervical space infection (MCSI), providing a reference basis for clinical management, diagnosis and treatment.

METHODS A total of 386 MCSI patients admitted to the First Affiliated Hospital of Chongqing Medical University from Jan. 2014 to Dec. 2024 were retrospectively analyzed. They were randomly divided into a training set (271 cases) and a validation set (115 cases) at a ratio of 7∶3. Based on the 75th percentile of hospital stay, patients were categorized into a short hospital stay group and a long hospital stay group. General baseline characteristics and biochemical test indicators obtained upon admission were compared between the two groups. Lasso regression was used to compress variables, and logistic regression analysis was employed to screen for risk factors of prolonged hospital stay. A prediction model for the risk factors of prolonged hospital stay in MCSI patients was constructed, and its discriminative ability, calibration and clinical net benefit were evaluated based on the receiver operating characteristic curve (ROC), calibration curve and decision curve, respectively.

RESULTS Multi-space infection, timing of surgery, age, albumin and C-reactive protein were identified as risk factors for prolonged hospital stay in MCSI patients (P < 0.05). Based on these findings, a prediction model was constructed. The area under the ROC curve (AUC) for the training set and validation set were 0.852 (95%CI: 0.802 to 0.902) and 0.793 (95%CI: 0.700 to 0.887), respectively. The Hosmer-Lemeshow test yielded χ²=7.860, P=0.447 (training set) and χ²=6.215, P=0.623 (validation set). The calibration curve indicated good goodness-of-fit, and the decision curve demonstrated favorable net benefit values when the threshold probability ranged from 0.01 to 0.64.

CONCLUSIONS Multi-space infection, timing of surgery, age, albumin and C-reactive protein are independent risk factors related to the prolonged hospital stay of MCSI patients. The prediction model constructed based on these factors has certain clinical predictive value.