Molecular Point-of-Care Diagnostics for Vector-Borne Pathogens: Biotechnological Innovations and Challenges in Field Implementation

mPOC Diagnostics for Vector-Borne Diseases

Abstract

Background: Centralized diagnostics can delay detection of vector-borne pathogens. Molecular point-of-care (mPOC) technologies may address this gap, but performance and field readiness remain uncertain.
Methods: This rapid systematic review followed Cochrane Diagnostic Test Accuracy guidance. PubMed, Scopus and Web of Science (Clarivate) were searched for 2020–2025 publications (13 September 2025). Two-by-two data were extracted or reconstructed, and sensitivity and specificity were recalculated with 95% Wilson confidence in­ter­vals. Studies were classified by setting, assessed using the real-time connectivity, ease of specimen collection, afford­able, sensitive, specific, user-friendly, rapid and robust, equipment-free and deliverable to end-users (REAS­SURED) criteria, and appraised with Quality Assessment of Diagnostic Accuracy Studies–2. Heterogeneity led to narrative synthesis.
Results: 1,964 records yielded 22 studies: 11 malaria, 6 leishmaniasis and 5 arboviral infections. Twenty-one stud­ies provided two-by-two data, yielding 26 non-overlapping analyses. One study reported sensitivity only (86.4%, 95% con­fidence interval 75.5–93.0%) and one analysis was not reconstructable. Median sample size was 100 (in­terquartile range 76.5–161.5; range 13–490). Thirteen analyses reported 100% sensitivity or specificity, but 9 had lower confi­dence lim­its below 90%. Four studies evaluated point-of-care tests. Based on lower confidence limits, 8 met sensi­tivity and 5 met specificity REASSURED thresholds; 6 met real-time connectivity and 7 met ease of spec­imen col­lection. Seventeen (77.3%) reported no cost data.
Conclusions: mPOC technologies are promising, but evidence is dominated by laboratory or near-laboratory stud­ies with imprecise estimates. Multicenter field evaluations, standardized reporting including costs and complete two-by-two data are needed before integration.