(431) Patient-Similarity Tool Enhancing Decision-Making on Whether to Place Cervical Cerclage

Decision-making for cervical cerclage placement is complex and requires consideration of specific patient characteristics. We sought to develop a patient similarity tool to augment clinical decision-making and enhance shared decision-making based on the clinical course of similar patients. Patient similarity identifies patients with similar clinical characteristics to predict outcomes and personalize treatment plans.

Study Design:

We used a retrospective cohort of all pregnancies at a tertiary center from 2010-24, including all individuals with discussions considering cervical cerclage. The list of patients and their clinical variables served as the analytic DATABASE for the patient similarity tool. We defined 16 variables for similarity assessment that were available in the electronic health record. An additional seven variables will require manual collection: cervical dilation, membranes bulging, cervical length, amniotic fluid sludge, uterine activity, WBC, and CRP. Based on the index patient’s similarity parameters, the tool detects similar patients and describes their maternal and neonatal outcomes with and without cerclage.

Results:

Among 138,439 deliveries during the study period, 684 clinical scenarios considering cervical cerclage were captured. Cerclage was performed in 343 (50.1%) of these cases. Table 1 compares individuals with and without cerclage based on the variables selected for similarity assessment. Significant differences were observed in maternal age, gravidity, previous painless dilation, previous cerclage, last pregnancy gestational age (GA) at delivery, spontaneous conception, multifetal pregnancy, and GA at decision time. Figure 1 illustrates an index patient and presents similar patients and their outcomes with and without cerclage.

Conclusion:

Developing a patient similarity tool to augment decision-making and decision sharing regarding cervical cerclage placement is feasible. We speculate that our proposed tool may improve data-driven decision-making and empower patient participation in shared decision-making.