BIOANALYSISPAPER
VectorCam, a phone based species identification tool built for community health workers, is being tested on whether the national program can absorb faster data, not on camera accuracy.
VectorCam, a phone- and camera-based mosquito identification tool built to be operated by community health workers, enters Uganda's national malaria program on a 12-district test designed to measure how a low-resource public health system uses faster surveillance data. The protocol, published this month in JMIR Research Protocol (DOI 10.2196/99717), is an implementation-effectiveness study rather than a model benchmark.
VectorCam runs species identification on a phone or tablet so that a trained community health worker can produce a result that today requires a microscopist with years of entomology experience. Uganda's routine vector surveillance depends on a small pool of those specialists, and the resulting reporting delays can stretch long enough to outlast the window in which a control response is useful. The VectorCam implementation package was co-designed with Uganda's Ministry of Health through a human-centered design process that worked with the program staff and health workers who would actually run it, before the rollout began, with the goal of pushing identification out of centralized labs and into the districts where malaria is endemic.
Mosquito species identification matters because the malaria control response depends on which species are present in a given district. Surveillance data feeds directly into district-level decisions, and a district that learns a week after collection which vector species it is dealing with cannot target its response to the local ecology on the ground. The protocol's premise is that faster identification, pushed down to community health workers in endemic districts, can shorten the loop between field collection and program decision.
The trial will unfold across 12 districts in Northern and Eastern Uganda assigned to four sequential rollout waves following a 2-district pilot. The implementation window runs for one year. District-level restricted randomization, rather than individual randomization, controls for the operational fact that surveillance data, malaria program decisions, and program staffing all happen at the district scale, and it lets the researchers compare waves against each other as the implementation package is refined. A two-district pilot phase runs first so the team can stress-test the rollout machinery, including the data pipeline and the supervision model, before the four waves begin.
The protocol measures four implementation outcomes: acceptability (whether health workers and program staff will use the tool in routine practice), feasibility (whether the system survives field conditions), fidelity (whether the package is delivered as designed), and cost. These are captured through qualitative interviews with health workers and program staff, system usage logs from the tool itself, field observations of how identification feeds into routine reporting, and program documentation of how the data reaches decision-makers. Effectiveness outcomes cover the technical performance of mosquito species identification, time efficiency, and whether surveillance data is operationally used for decisions. Accuracy is benchmarked against molecular identification, the reference standard for species confirmation, alongside additional classification performance metrics.
Adaptive learning stitches the two halves together. The protocol calls for iterative refinement of the implementation package as field experience comes in, with mixed-methods triangulation and longitudinal analyses across districts and rollout waves. The researchers plan to use what the qualitative interviews and field observations surface, including bottlenecks in reporting, data use, and worker training, to revise the package before each subsequent wave. Each wave gets a revised tool, shaped by the friction the prior wave's interviews and observations surfaced.
The design's stress test sits at the seam between data and decision. A faster identification result only matters if a district malaria surveillance team can act on it inside the same transmission season, and the protocol's measurement of operational use is built around that question rather than the camera's accuracy. The pilot phase results, which the protocol does not yet report, will be the first signal of how the absorption is going before the four rollout waves begin.