Reducing errors in identification of von Willebrand disease: The experience of the Royal College of Pathologists of Australasia Quality Assurance Program

Regular multilaboratory surveys of laboratories derived primarily from Australia, New Zealand, and Southeast Asia have been conducted during the last 8 years to evaluate testing proficiency in the diagnosis of von Willebrand disease (vWD). We summarize and update the findings of these surveys with a particular emphasis on diagnostic errors and error rates associated with particular tests or test panel limitations. A total of 43 plasma samples have been dispatched to survey participants. These have included 13 normal samples, five type 1 vWD samples, eight type 2 vWD samples (three 2A, three 213, one 2M, and one 2N), and four type 3 vWD samples. In addition to numerical test results, participant laboratories (currently, n = 49) were asked to provide diagnostic interpretations regarding results, and whether or not vWD was suggested, and if so, a probable subtype. Although laboratories usually provided correct interpretative responses, diagnostic errors occurred in a substantial number of cases. On average, type 1 vWD plasma was misidentified as type 2 vWD in 13.3% of cases, and laboratories performing von Willebrand factor (vWF):ristocetin cofactor activity (RCo) without vWF:collagen-binding activity (CB) were seven times more likely to make such an error compared with those performing vWF:CB. Similarly, type 2 vWD plasma was misidentified as type 1 or type 3 vWD in an average of 20.1% of cases, and laboratories performing vWF:RCo without vWF:CB were three times more likely to make such an error compared with those performing vWF:CB. Finally, normal plasma was misidentified as vWD in an average of 6.7% of cases, and laboratories performing vWF:RCo without vWF:CB were four times more likely to make such an error compared with those performing vWF:CB. We conclude that although laboratories are generally proficient in tests for vWD, diagnostic errors do occur and error rates are substantially reduced when test panels are more comprehensive and include the vWF:CB.