Predictors of Prolonged Hospital Length of Stay in Patients With Odontogenic Infections in Ghana
Source: PubMed Central Open Access, NCBI / U.S. National Library of Medicine
Background Poor outcomes of odontogenic infections usually increase the length of stay (LOS) in hospitals, and the cost of treatment increases substantially. The LOS of patients with odontogenic infections is not set in stone. In clinical practice, it is observed that cost, certain medications and treatments, age and a plethora of factors influence this. However, it is unclear which factors have direct effects on it. As such, evidence‐based interventions become difficult. sec-0001 Methodology The study utilised a retrospective observational approach and a total population sampling technique to investigate 286 out of the 811 patients admitted at the allied ward of STH from 2021 to 2025. Data was extracted from the Lightwave Health Information Management System and analysed with IBM SPSS 27, Claude (Anthropic, version Sonnet 4.6) and Python (Version 3.12). sec-0002 Results A total of 286 patients were included, with a mean hospital length of stay (HLOS) of 9.28 ± 4.21 days. Necrotising fasciitis and Ludwig′s angina were associated with the longest admissions. On proportional odds ordinal logistic regression, severe infection classification (OR 25.39, 95% CI: 4.21–153.32) and necrotising fasciitis (OR 9.36, 95% CI: 3.87–22.61) were the strongest independent predictors of prolonged HLOS (all< 0.001). HIV/AIDS, diabetes mellitus, hypertension, Ludwig′s angina and male sex were also significant independent predictors. The model demonstrated strong explanatory power (Nagelkerke= 0.6
Abstract
Background Poor outcomes of odontogenic infections usually increase the length of stay (LOS) in hospitals, and the cost of treatment increases substantially. The LOS of patients with odontogenic infections is not set in stone. In clinical practice, it is observed that cost, certain medications and treatments, age and a plethora of factors influence this. However, it is unclear which factors have direct effects on it. As such, evidence‐based interventions become difficult. sec-0001 Methodology The study utilised a retrospective observational approach and a total population sampling technique to investigate 286 out of the 811 patients admitted at the allied ward of STH from 2021 to 2025. Data was extracted from the Lightwave Health Information Management System and analysed with IBM SPSS 27, Claude (Anthropic, version Sonnet 4.6) and Python (Version 3.12). sec-0002 Results A total of 286 patients were included, with a mean hospital length of stay (HLOS) of 9.28 ± 4.21 days. Necrotising fasciitis and Ludwig′s angina were associated with the longest admissions. On proportional odds ordinal logistic regression, severe infection classification (OR 25.39, 95% CI: 4.21–153.32) and necrotising fasciitis (OR 9.36, 95% CI: 3.87–22.61) were the strongest independent predictors of prolonged HLOS (all< 0.001). HIV/AIDS, diabetes mellitus, hypertension, Ludwig′s angina and male sex were also significant independent predictors. The model demonstrated strong explanatory power (Nagelkerke= 0.686,< 0.001). All predictor variance inflation factors were below 2.5, indicating no multicollinearity concerns. sec-0003 Conclusion Infection severity, primary diagnosis, immunocompromising comorbidities and surgical interventions were the principal independent determinants of prolonged HLOS. Multispace involvement showed a crude association with extended HLOS but did not emerge as an independent predictor in the adjusted ordinal regression model. Early diagnosis and prompt, multidisciplinary management are crucial to reducing hospitalisation and improving patient outcomes. sec-0004