PB1222 - Cut-off score for identifying patients at risk of Hemophilia using a Hemophilia self-screening tool in resource-constrained settings

Background: Hemophilia diagnosis remains a persistent challenge in resource-constrained countries, particularly in Indonesia, where only around 10% of patients with hemophilia are diagnosed compared to the estimated prevalence. In response, we developed an affordable, user-friendly hemophilia self-screening tool consisting of 17 multiple-choice questions.

Aims: This research aims to determine the cut-off score for identifying hemophilia risk using the self-screening tool in Indonesia. The obtained score serves as a diagnostic tool, indicating either a high or low risk of hemophilia.

Methods: The self-screening tool comprises 17 questions related to bleeding and hemophilia, with answers ranging from none to heavy bleeding, and 1 question on family history. A cross-sectional study involving children under 18 years old was conducted from June 2021 to July 2022 in South Kalimantan, Indonesia. Participants included both children with hemophilia and healthy children who consented to participate. A total of 214 children were enrolled. Receiver Operating Characteristic (ROC) curve analysis was utilized to determine the cut-off score.

Results: Two hundred fourteen (214) children participated. The optimal cut-off score was determined to be 26.5, demonstrating 90.5% of sensitivity. This score serves as a baseline for categorizing hemophilia risk; scores above 26.5 indicate a high risk, while scores below 26.5 indicate a low risk. A positive response to a question on family history of hemophilia significantly influences the categorization of individuals into high-risk. The Area under the Curve (AUC) reached a value of 0.988, indicating high true positive and low false negative.

Conclusion(s): The cut-off score for patients at risk of hemophilia using the hemophilia self-screening tool was 26.5. Patients with a family history of hemophilia will remain at high-risk, despite the score falling below 26.5. This tool holds promise in efficiently identifying at-risk individuals, exhibiting high sensitivity, and enabling accessible implementation to improve diagnostics in resource-constrained settings.