Fitri Kusumasa
Faculty of Health Science, Universidade Dili, Timor-Leste

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Face Detection-Enabled FAST Mobile Application for Stroke Risk Self-Screening in Hypertensive Patients Anita Fatarona; Hendra Cahyono; Zahra Saad; Fitri Kusumasa
Jurnal Kesehatan Masyarakat Vol. 22 No. 1 (2026)
Publisher : Universitas Negeri Semarang in collaboration with Ikatan Ahli Kesehatan Masyarakat Indonesia (IAKMI Tingkat Pusat) and Jejaring Nasional Pendidikan Kesehatan (JNPK)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/kemas.v22i1.38954

Abstract

Delayed pre-hospital recognition of stroke symptoms sustains high mortality and disability, especially where awareness and rapid access are limited. Objective: To evaluate a smartphone FAST application integrating face detection for self-screening among individuals with hypertension. Methods: This study was conducted from July to August 2025. Two expert raters assessed usability with the MAUQ; inter-rater reliability was assessed using Cohen’s kappa. A pre–post study (n=80) tested change in combined knowledge/behavior; normality was confirmed, and paired t-tests were applied. Post-intervention semi-structured interviews underwent thematic analysis. Results: Expert agreement was statistically significant (kappa = 0.400; p = 0.018), indicating fair-to-moderate reliability. Knowledge/behavior improved significantly from pre to post (t (79) = -21,334; p <0.001; mean difference −4.538; 95% CI -4,961 to -4,114). Themes included helpfulness (practical recognition, relevant education), internet dependence (limitations in low-coverage areas), perceived complexity (photo-capture burden), portability (use across settings), and targeted education (content aligned with needs). Conclusion: A smartphone-based FAST app with face detection is feasible and associated with significant gains in knowledge/behavior for early stroke recognition. Priorities include improving rater consistency, validating module accuracy against clinical standards, reducing dependence on connectivity, and streamlining image capture to enhance pre-hospital impact.