The Utility of the TWIST Score in Diagnosis and Risk Stratification of Patients with Testicular Torsion: Results of a Quality Improvement Study
Jackson Joseph Sanden Cabo, MD, Mark C. Adams, MD, Stacy Tanaka, MD, John W. Brock, III, MD, Douglass Clayton, MD, Abby S. Taylor, MD, John C. Pope, IV, MD, John Thomas, MD.
Vanderbilt University Med Center, Nashville, TN, USA.
Background:
Testicular torsion is a frequently encountered urologic emergency. The TWIST score is a 5-component physical examination score used to aid in efficient diagnosis but had not been used historically at our institution. In a retrospective analysis of 133 children treated for torsion at our facility (2008-2017), we applied individual TWIST score components such as firm testicle on exam and found a correlation with need for orchiectomy. We subsequently introduced the TWIST score into Emergency Medicine (ED) and Urology practice, and a prospective study of score compliance was undertaken.
Methods:
As a two-stage quality improvement initiative, ED staff were educated about the TWIST score and asked to complete the 5-component assessment (testicular swelling, hard testicle, nausea/vomiting, absent cremasteric reflex, and high riding testicle) for patients presenting with acute scrotal pain. Simultaneously, an electronic medical record-based dot phrase was introduced for urology trainees to complete an independent TWIST evaluation. The primary objective was to achieve 100% compliance amongst the groups. Secondary goals were correlation of accuracy between the ED and urology service, need for orchiectomy, time from ED to operating room (OR), and assessment of repeat pre-op imaging. Differences in compliance were determined using Fisher’s exact and Wilcoxon tests. All statistical tests were 2-tailed with alpha=0.05.
Results: 90 patients presented to the emergency room from 3/2018 to 11/2020 with a complaint of acute scrotal pain; 37 were diagnosed with torsion and underwent operative intervention.As compared to our retrospective cohort, the documentation rate of complete TWIST score components on physical exam rose from 10.4% to 100% (P<0.001) on ED evaluation and 16.9% to 78.3% on urology evaluation (P<0.001). Rates of repeat US for patient’s transferred between facilities was similar (57.5% vs. 62.5%) as was median time to OR (159 minutes vs. 139 minutes; p=0.577). Using cutoffs of <2 (low) and >5 (high) for TWIST scores yielded a sensitivity of 94.5 % (NPV 92.5%) and specificity of 94.3% (PPV 85%) for a diagnosis of torsion. Median TWIST scores were similar in patients requiring orchiectomy compared to those where salvage was performed (ED Scores: 6 points vs. 5 points; P=0.46; Urology Score: 6 vs. 5; P=0.22). However, TWIST specific components did correlate with need for orchiectomy—100% of patients who required orchiectomy had a firm testicle on examination by a urologist, compared to 52% of those without firmness noted on exam (P=0.006).
Conclusions:
Utilization of the quality improvement process of Plan, Do, Study, Act (PDSA), we were able to significantly improve utilization of the TWIST score by ED and urology staff for workup of patients with acute testicular pain. We were able to confirm the high sensitivity and specificity of the TWIST score to diagnose testicular torsion according to risk stratification described by Barbossa. However, utilization of US and time to OR were comparable to our retrospective cohort without complete TWIST scores. On prospective analysis, we found that testicular firmness on exam remained predictive of need for orchiectomy. Future work will aim at instituting more judicious use of ultrasound to facilitate quicker time to OR.
Back to 2021 Abstracts