Abstract
Background: Unique patient identification is necessary for detecting malaria recurrence and facilitates individualised treatment. Reliable patient identification in settings with poor information technology infrastructure is challenging. A before-and-after study was conducted to investigate whether the use of biometric fingerprinting alongside an existing physical 'malaria card' system improved unique identification of patients with malaria at a busy, publicly-funded clinic in Papua, Indonesia.
Methods: A three-phase study was conducted at Wania clinic in Timika over a 16-month period. In Phase 1 (the 'pre-intervention period'-3 months), patient identification practices using the standard 'malaria card' system were documented. In Phase 2 ('training'-6 months), fingerprinting was introduced, troubleshooting was undertaken, and biometric data were gradually accrued. In Phase 3 ('consolidation'-7 months), fingerprinting continued to be incorporated into routine clinic practice. The main outcome of interest was the proportion of malaria patients visiting the clinic's 'Malaria Corner' who could be linked to a pre-existing malaria card number in Phase 1 versus Phase 3. Analyses were descriptive.
Results: During the 16-month study period, 7471 patients with malaria visited the Malaria Corner at Wania clinic, 1487 in Phase 1 (80.3%), 3228 in Phase 2 (59.6%), and 2756 in Phase 3 (59.1%). The proportion of patients who attended the Malaria Corner with a malaria card was 33.1% (492/1487) in Phase 1, 36.3% (663/1828) in Phase 2, and 46.9% (938/2001) in Phase 3. Overall, 56.6% (1828/3228) of patients attending the Malaria Corner had biometric fingerprinting in Phase 2 and 72.6% (2001/2756) in Phase 3. The proportion of all patients attending the Malaria Corner who could be linked to a pre-existing malaria card number increased from 44.4% (660/1487) in Phase 1 to 48.9% (1348/2756) in Phase 3 (difference = 4.5%, 95% Confidence Interval (CI) 1.4-7.7%). In Phase 3, 55.8% (1117/2001) of patients who had fingerprinting were linked to a pre-existing malaria card number (difference compared with Phase 1 = 11.4%, 95% CI 8.1-14.8%). Of the 2714 patients who were offered biometric fingerprinting for the first time, 0.39% (6/1556) refused in Phase 2 and 0.26% (3/1158) refused in Phase 3. Challenges in implementation included unreadable fingerprints, particularly among children.
Conclusion: This study demonstrates the potential use of biometrics to improve patient identification in resource-limited settings and to streamline workflows. Expanding biometric systems to include complementary methods, such as facial recognition, could further address challenges in uniquely identifying specific patient groups, such as young children.