ABSTRACT
OBJECTIVE: Post-operative length of hospital stay (LOS) is a valuable measure for monitoring quality of care provision, patient recovery, and guiding hospital resource management. But the impact of patient ethnicity, socio-economic deprivation as measured by the indices of multiple deprivation (IMD), and pre-existing health conditions on LOS post-anterior cervical decompression and fusion (ACDF) is under-researched in public healthcare settings. METHODS: From 2013 to 2023, a retrospective study at a single center reviewed all ACDF procedures. We analyzed 14 non-clinical predictors-including demographics, comorbidities, and socio-economic status-to forecast a categorized LOS: short (≤2 days), medium (2-3 days), or long (>3 days). Three machine learning (ML) models were developed and assessed for their prediction reliability. RESULTS: 2033 ACDF patients were analyzed; 79.44 % had a LOS ≤ 2 days. Significant predictors of LOS included patient sex (HR:0.81[0.74-0.88], p < 0.005), IMD decile (HR:1.38[1.24-1.53], p < 0.005), smoking (HR:1.24[1.12-1.38], p < 0.005), DM (HR:0.70[0.59-0.84], p < 0.005), and COPD (HR:0.66, p = 0.01). Asian patients had the highest mean LOS (p = 0.003). Testing on 407 patients, the XGBoost model achieved 80.95 % accuracy, 71.52 % sensitivity, 85.76 % specificity, 71.52 % positive predictive value, and a micro F1 score of 0.715. This model is available at: https://acdflos.streamlit.app. CONCLUSIONS: Utilizing non-clinical pre-operative parameters such as patient ethnicity, socio-economic deprivation index, and baseline comorbidities, our ML model effectively predicts postoperative LOS for patient undergoing ACDF surgeries. Yet, as the healthcare landscape evolves, such tools will require further refinement to integrate peri and post-operative variables, ensuring a holistic decision support tool.
ABSTRACT
Background: Over the past decade, neurosurgical interventions have experienced changes in operative frequency and postoperative length of stay (LOS), with the recent COVID-19 pandemic significantly impacting these metrics. Evaluating these trends in a tertiary National Health Service center provides insights into the impact of surgical practices and health policy on LOS and is essential for optimizing healthcare management decisions. Methods: This was a single tertiary center retrospective case series analysis of neurosurgical procedures from 2012 to 2022. Factors including procedure type, admission urgency, and LOS were extracted from a prospectively maintained database. Six subspecialties were analyzed: Spine, Neuro-oncology, Skull base (SB), Functional, Cerebrospinal fluid (CSF), and Peripheral nerve (PN). Mann-Kendall temporal trend test and exploratory data analysis were performed. Results: 19,237 elective and day case operations were analyzed. Of the 6 sub-specialties, spine, neuro-oncology, SB, and CSF procedures all showed a significant trend toward decreasing frequency. A shift toward day case over elective procedures was evident, especially in spine (P < 0.001), SB (tau = 0.733, P = 0.0042), functional (tau = 0.156, P = 0.0016), and PN surgeries (P < 0.005). Over the last decade, decreasing LOS was observed for neuro-oncology (tau = -0.648, P = 0.0077), SB (tau = -0.382, P = 0.012), and functional operations, a trend which remained consistent during the COVID-19 pandemic (P = 0.01). Spine remained constant across the decade while PN demonstrated a trend toward increasing LOS. Conclusion: Most subspecialties demonstrate a decreasing LOS coupled with a shift toward day case procedures, potentially attributable to improvements in surgical techniques, less invasive approaches, and increased pressure on beds. Setting up extra dedicated day case theaters could help deal with the backlog of procedures, particularly with regard to the impact of COVID-19.
ABSTRACT
Background: The aim of this study was to develop natural language processing (NLP) algorithms to conduct automated identification of incidental durotomy, wound drains, and the use of sutures or skin clips for wound closure, in free text operative notes of patients following lumbar surgery. Methods: A single-centre retrospective case series analysis was conducted between January 2015 and June 2022, analysing operative notes of patients aged >18 years who underwent a primary lumbar discectomy and/or decompression at any lumbar level. Extreme gradient-boosting NLP algorithms were developed and assessed on five performance metrics: accuracy, area under receiver-operating curve (AUC), positive predictive value (PPV), specificity, and Brier score. Results: A total of 942 patients were used in the training set and 235 patients, in the testing set. The average age of the cohort was 53.900 ± 16.153 years, with a female predominance of 616 patients (52.3%). The models achieved an aggregate accuracy of >91%, a specificity of >91%, a PPV of >84%, an AUC of >0.933, and a Brier score loss of ≤0.082. The decision curve analysis also revealed that these NLP algorithms possessed great clinical net benefit at all possible threshold probabilities. Global and local model interpretation analyses further highlighted relevant clinically useful features (words) important in classifying the presence of each entity appropriately. Conclusions: These NLP algorithms can help monitor surgical performance and complications in an automated fashion by identifying and classifying the presence of various intra-operative elements in lumbar spine surgery.
