Comparison of a mobile phone-based malaria reporting system with source participant register data for capturing spatial and temporal trends in epidemiological indicators of malaria transmission collected by community health workers in rural Zambia

BACKGROUND: Timeliness, completeness, and accuracy are key requirements for any surveillance system to reliably monitor disease burden and guide efficient resource prioritization. Evidence that electronic reporting of malaria cases by community health workers (CHWs) meet these requirements remains limited. METHODOLOGY: Residents of two adjacent rural districts in Zambia were provided with both passive and active malaria testing and treatment services with malaria rapid diagnostic tests (RDTs) and artemisinin-based combination therapy by 42 CHWs serving 14 population clusters centred around public sector health facilities. Reference data describing total numbers of RDT-detected infections and diagnostic positivity (DP) were extracted from detailed participant register books kept by CHWs. These were compared with equivalent weekly summaries relayed directly by the CHWs themselves through a mobile phone short messaging system (SMS) reporting platform. RESULTS: Slightly more RDT-detected malaria infections were recorded in extracted participant registers than were reported in weekly mobile phone summaries but the difference was equivalent to only 19.2% (31,665 versus 25,583, respectively). The majority (81%) of weekly SMS reports were received within one week and the remainder within one month. Overall mean [95% confidence limits] difference between the numbers of register-recorded and SMS-reported RDT-detected malaria infections per CHW per week, as estimated by the Bland Altman method, was only −2.3 [−21.9, 17.2]. The mean [range] for both the number of RDT-detected malaria infections (86 [0, 463] versus 73.6 [0, 519], respectively)) and DP (22.8% [0.0 to 96.3%] versus 23.2% [0.4 to 75.8%], respectively) reported by SMS were generally very consistent with those recorded in the reference paper-based register data and exhibited similar seasonality patterns across all study clusters. Overall, mean relative differences in the SMS reports and reference register data were more consistent with each other for DP than for absolute numbers of RDT-detected infections, presumably because this indicator is robust to variations in patient reporting rates by location, weather, season and calendar event because these are included in both the nominator and denominator. DISCUSSION/CONCLUSION: The SMS reports captured malaria transmission trends with adequate accuracy and could be used for population-wide, continuous, longitudinal monitoring of malaria transmission.