Development and validation of a risk-based algorithm for preoperative type and screen testing in spine surgery.

BACKGROUND CONTEXT

With improvements in surgical techniques and perioperative management, transfusion rates after spine surgery have decreased over time. Given this trend, routine preoperative ABO/Rh type and antibody screen (T&S) laboratory testing may not be warranted in all patients undergoing spine surgery.

PURPOSE

The aim of the current study is to evaluate risk factors for intra/postoperative transfusion in patients undergoing a variety of spine procedures and to develop an algorithm for selectively ordering preoperative T&S testing in appropriate patients.

STUDY DESIGN/SETTING: This is a single institution, retrospective observational study of patients undergoing emergent or elective spine surgery. External validation of the algorithm was performed on a national sample of patients undergoing spine surgery from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) national database.

PATIENT SAMPLE

A total of 5,947 surgeries from January 1, 2016 to December 31, 2019 at a single institution, and 166,113 surgeries from the 2016 to 2018 ACS-NSQIP database.

OUTCOME MEASURES

The primary outcome measure was performance of intraoperative or postoperative transfusion.

METHODS

Using the institutional sample, univariate statistics (chi-square tests, fisher's exact test, 2-sided independent sample tests) were performed to compare demographics, comorbidities, and surgical details (case type, number of levels treated, etc.) between patients who did and did not require intra- or postoperative transfusion. Transfusion rates were calculated and compared across procedure types. Multivariate logistic regression was performed to identify independent predictors of transfusion and the model's accuracy was evaluated using the area under the curve (AUC) of the receiver operating characteristics (ROC) curve. A risk-based algorithm suggesting no preoperative T&S in low transfusion risk procedures, routine preoperative T&S in high-risk procedures, and further assessment in medium risk thoracolumbar fusion procedures was created. The algorithm's sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were evaluated when it was applied to both the institutional and national samples. Potential cost savings from reducing T&S orders were calculated.

RESULTS

In the institutional sample, 120 patients (2.0%) required intraoperative or postoperative transfusion. The highest rates of transfusion were found in corpectomy (10.5%) and anterior/posterior cervical fusion (6.9%) procedures. In the multivariate logistic regression model, the presence of a preoperative coagulation defect or hemorrhagic condition (OR: 7.149, p<.001) and 6+ level surgery (OR: 7.511, p<.001) were the strongest predictors of transfusion. Overall, the model generated an AUC of 0.882, indicating excellent predictive accuracy. When applied to the institutional cohort, the risk-based algorithm had a sensitivity of 78.3%, specificity of 80.5%, PPV of 7.6%, and NPV of 99.4% for evaluating likelihood of transfusion. Using the algorithm 4,717 T&S tests would have been eliminated (79.3%), resulting in a cost savings of $179,246. Application of the model to the ACS-NSQIP cohort resulted in a sensitivity of 61.9%, specificity of 84.6%, PPV of 15.6%, and NPV of 98.0%.

CONCLUSIONS

The routine use of preoperative ABO/Rh type and antibody screen testing does not appear to be warranted in patients undergoing spine surgery. A risk-based approach to preoperative type and screen testing may eliminate unnecessary tests and generate significant cost savings with minimal disruption to clinical care.