Predicting the risk of postoperative constipation in middle-aged and elderly patients with lower limb fractures using machine learning algorithms
Description
This dataset contains de-identified clinical data from a cohort of middle-aged and elderly patients who underwent surgery for lower limb fractures. The data were collected to support the development and validation of machine learning models for predicting the risk of postoperative constipation. The variables encompass patient demographics, pre-operative comorbidities, surgical details, medication use, and postoperative outcomes, with the target variable being the occurrence of constipation within a specified postoperative period. The dataset is provided in a structured format (e.g., .excel) to facilitate reuse in predictive analytics and clinical research.
Files
Steps to reproduce
This was a retrospective cohort study conducted at the Affiliated Traditional Chinese Medicine Hospital of Southwest Medical University. The study protocol was reviewed and approved by the Ethics Committee of the Affiliated Traditional Chinese Medicine Hospital of Southwest Medical University (Approval No.: BY2024031). The requirement for informed consent was waived due to the retrospective nature of the study. We screened the electronic medical records of all patients aged [e.g., ≥45 years] who were admitted for surgical treatment of lower limb fractures (including [e.g., femoral neck, intertrochanteric, or tibial plateau fractures]) between [Start Date: e.g., January 1, 2020] and [End Date: e.g., December 31, 2023] Inclusion Criteria: (1) Aged 45 years or older; (2) Underwent elective or emergency surgery for a lower limb fracture; (3)Had a complete medical record for the relevant postoperative period. Exclusion Criteria: (1) Pre-existing chronic constipation or gastrointestinal diseases (e.g., inflammatory bowel disease); (2) Spinal cord injury or neurological disorders affecting bowel function; (3) Patients who were unable to communicate or died during the postoperative period. Data were manually extracted from the hospital's Electronic Medical Record system by two trained research assistants using a standardized data collection form. Any discrepancies were resolved through discussion and by referring to the original records.