Clinical Scorecard: Creation and assessment of a predictive model for preoperative outcomes in elderly patients undergoing emergency general surgery

At a Glance

Key Highlights

• Elderly patients constitute over 40% of emergency department visits in China, with increasing numbers undergoing emergency surgery.
• Existing scoring systems (POSSUM, SRS, CCI) are not specifically designed for elderly emergency surgery patients and have limitations.
• A novel geriatric emergency surgery score (GES) was developed using retrospective data and logistic regression to predict in-hospital mortality.

Guideline-Based Recommendations

Diagnosis

• Collect comprehensive preoperative data including demographics, comorbidities, ASA classification, laboratory values, and imaging.
• Calculate GNRI and estimate GFR using CKD-EPI equation to assess nutritional and renal status.
• Exclude patients with incomplete data, trauma, or surgeries under local/regional anesthesia.

Management

• Use the newly developed GES model to stratify mortality risk preoperatively in elderly emergency surgery patients.
• Incorporate independent predictors identified by multivariate logistic regression to guide surgical decision-making.

Monitoring & Follow-up

• Monitor postoperative complications and mortality rates to validate and refine predictive models.
• Track hospital stay duration and final diagnosis for outcome assessment.

Risks

• Higher risk of severe postoperative complications and mortality in elderly patients undergoing emergency surgery due to physiological decline.
• Limitations of existing scoring systems may lead to inaccurate risk prediction if used without adjustment.

Patient & Prescribing Data

Elderly patients admitted for emergency general surgery at a tertiary hospital in China between 2017 and 2022.

Retrospective data analysis enabled identification of key preoperative predictors of mortality, facilitating development of a tailored risk score.

Clinical Best Practices

• Perform thorough preoperative assessment including nutritional and renal function evaluation using GNRI and GFR.
• Apply statistical modeling (univariate and multivariate logistic regression) to identify independent mortality predictors.
• Develop and use simplified scoring systems based on weighted odds ratios for practical clinical application.
• Exclude non-relevant patient groups (trauma, local anesthesia cases) to improve model specificity.

References

• 2022 China Health Statistical Yearbook
• CKD-EPI Equation for GFR Estimation