Unnecessarily long hospital stays are costly and inefficient. Studies have shown that the length of hospital stay (LOS) for spine surgical procedures is influenced by various disease-related or medical factors, but few have examined the role of socio-demographic/socio-economic (SDE) factors.
Methods
This was a retrospective analysis of data from 10,770 patients (5056 men, 5714 women; 62 ± 15 years) with degenerative spinal disorders, collected prospectively in an in-house database within the framework of EUROSPINE’s Spine Tango Registry. Surgeons completed the Tango surgery form (clinical history, demographics, surgical measures, complications), and patients, a baseline Core Outcome Measures Index. Stepwise linear regression analyses examined SDE predictors of LOS, controlling for potential medical/biological factors.
Results
The mean LOS was 7.9 ± 5.2 days. The final model accounted for 42% of variance in LOS, with SDE variables explaining 13% variance and medical/surgical predictors, 29%. In the final model, the SDE factors age and being female were significant independent predictors of LOS, whereas others were either non-significant (insurance status, being of Swiss nationality, being a smoker) or reached only borderline significance (p < 0.1) (BMI). Controlling for all other SDE and medical/surgical confounders, being female was associated with 1.11-day longer LOS (95% CI 0.96–1.27; p < 0.0001).
Conclusions
Patients of advanced age and female gender are at increased risk of longer hospital stay after surgery for degenerative spinal disorders. Further studies should seek to understand the reasoning behind the gender disparity, in order to minimise potentially unnecessary costs of prolonged LOS. Targeted preoperative discharge planning may improve the utilisation of hospital resources.
Graphic abstract
These slides can be retrieved under Electronic Supplementary Material.
Calculation of this score was only possible for the data collected since 2012 (see earlier), and therefore the model is based on a smaller, more recent sub-sample of patients (n = 5,409) (Fig. 1).
