Physician-Assessed Vs. Patient-Reported Bleeding Scores in an Undiagnosed Clinic Population

Background: The International Society of Thrombosis and Haemostasis (ISTH) has proposed that a physician or other trained healthcare provider administer a bleeding assessment tool (ISTH-BAT) in the evaluation of bleeding symptoms. ISTH-BAT scores may then guide further testing or subspecialty referral. However, the ISTH-BAT is lengthy to administer and requires a trained assessor to interpret and categorize patient responses. A version of the tool that patients may complete independently (Self-BAT) has been developed and tested in von Willebrand Disease; however, it has not been compared to the ISTH-BAT in undifferentiated patients in the context of a clinic encounter.

Objective: We sought to compare the results of the ISTH-BAT vs. Self-BAT in adult patients presenting to hematology clinic for evaluation of a potential bleeding disorder.

Methods: Sequential adult patients presenting for evaluation of a potential bleeding disorder completed an online-adapted Self-BAT using an iPad. An attending hematologist or supervised hematology fellow completed an ISTH-BAT as part of the routine clinical encounter, aided by an electronic medical record-embedded template. The physician did not have access to the Self-BAT scores. Reliability was assessed using intraclass correlation coefficients (ICC) and Krippendorff’s alpha (α). The ISTH-BAT and Self-BAT were considered normal or abnormal using established cutoff values (>5 for women or >3 for men).

Results: Fifty-one patients were enrolled, 49 of whom had a complete ISTH-BAT and Self-BAT. Eighty-eight percent of patients were female, 12% were male; 69% of patients were White, 18% Black, 6% Asian, and 2% American Indian. One quarter of patients had a prolonged Platelet Function Analyzer (PFA-100) test in either the collagen/epinephrine or collagen/ADP cartridges. Clinical and laboratory evaluation yielded the following diagnoses: Bleeding Disorder of Unknown Cause (36%), platelet function defect (22%), von Willebrand Disease (10%), acquired bleeding disorder (4%), no bleeding disorder (24%), and other (4%). The mean ISTH-BAT score was 8.08 (SD=4.20) and the mean Self-BAT score was 9.72 (SD=6.06). Five patients with a positive ISTH-BAT had a negative Self-BAT and 5 patients with a negative Self-BAT had a positive ISTH-BAT. Intraclass correlation using a one-way random-effects model with absolute agreement yielded an individual ICC of 0.64 (95% CI 0.45 – 0.78). Agreement between the two measures was moderate to high for post-partum hemorrhage (α = 0.84), epistaxis (α =0.76), and menstrual bleeding (α =0.66). Agreement was lowest for muscle hematoma (α =0.47), cutaneous bleeding (α =0.46), bleeding from minor wounds (α =0.45), hematuria (α =0.41), GI bleeding (α =0.35), and hemarthrosis (α =0.10). The Self-BAT was slightly more sensitive and specific than the ISTH-BAT in the diagnosis of any hemostatic disorder (sensitivity 87% vs. 78%, specificity 72% vs. 66%): 1 patient with von Willebrand Disease and 1 patient with a platelet function defect had normal ISTH-BAT and abnormal Self-BAT scores.

Conclusion: Moderate overall reliability is seen between the ISTH-BAT and Self-BAT. Self-BAT scores were on average higher than ISTH-BAT score and demonstrated slightly increased sensitivity for the diagnosis of a bleeding disorder using established cut-off values. Assessment of bleeding severity in certain domains may be more subjective (e.g. extent of bruising or bleeding from minor wounds) or unfamiliar to patients (hemarthrosis, muscle hematoma), leading to discrepancy. More research is necessary on how to best integrate patient self-report with clinician assessment in the evaluation of bleeding disorders.