Method agreement analysis and interobserver reliability of the ISTH proposed definitions for effective hemostasis in management of major bleeding.

ABSTRACT

Essentials In 2016 the SSC proposed definitions for effective hemostasis in management of major bleeding. To validate these definitions, we studied the use in three large anticoagulant-reversal studies. Method agreement analysis and interobserver reliability showed at least acceptable agreement. Recommendations were made, advising use of the definition in hemostatic effectiveness studies. SUMMARY: Introduction In 2016 the Scientific and Standardization Subcommittee (SSC) on Control of Anticoagulation of the International Society on Thrombosis and Haemostasis (ISTH) proposed criteria to evaluate the effectiveness of anticoagulant reversal in major bleeding management. Testing and validation of these criteria are required. Objective To investigate the method agreement, interobserver reliability and applicability of the ISTH proposed definitions for hemostatic effectiveness. Methods Patient data from three anticoagulant-antidote studies were used for hemostatic effectiveness assessment using the ISTH-proposed definitions and clinical opinion. For every patient a case document was produced. For each cohort, four adjudicators were asked to assess the hemostatic effectiveness independently on a case-by-case basis. Agreement between the two methods of hemostatic effectiveness assessment was calculated using Cohen's kappa (κ), with a calculated sample size of at least 73 cases. Results The full dataset consisted of 116 cases, resulting in 464 assessments. Method agreement in outcome was observed in 364 of 464 assessments (78.5%), resulting in κ of 0.634 (95% CI 0.575-0.694), or "substantial agreement." Interobserver reliability analysis of the proposed definitions computed an overall agreement of 54.2% with κ of 0.312 ("fair agreement"). Discussion Method agreement analysis shows that the conclusions drawn using the ISTH definitions have "substantial agreement" with clinical opinion. Interobserver reliability analysis demonstrated acceptable agreement. In-depth analysis provided minor opportunities for further improvement and correct application of the definition. The definition is recommended to be used in all future studies evaluating hemostatic effectiveness, taking the suggested recommendations into account.