Malaria Data-to-Decisions Initiative - Sproxil
Strengthening evidence-driven decision making and tailored interventions at the subnational level across Nigeria.
This portal presents findings from a private sector–focused Monitoring, Learning and Evaluation (MLE) survey in Kano and Kaduna States, Nigeria. The study supports Nigeria's Subnational Tailoring (SNT) approach to malaria — matching interventions to local need — by building evidence on how malaria is diagnosed and treated outside formal health facilities, where many people first seek care. It was funded by the Gates Foundation, led by Solina Centre for International Development and Research (SCIDaR), and implemented by Sproxil.
Gates Foundation
SCIDaR — Solina Centre for International Development and Research
How people recognise, test for and treat malaria, and their awareness of national treatment guidelines.
Services offered, testing practices, medicine availability and malaria case management among private providers.
Malaria positivity drawn from AI-interpreted images of used rapid-test (mRDT) cassettes.
Patients who verified an antimalarial on Sproxil's platform were invited by SMS, or called by a multilingual agent, to complete a short survey in English, Hausa or Pidgin.
Registered PPMVs and pharmacies completed a questionnaire via a unique link, with agent support where needed.
Providers tested patients and sent a photo of each result; Sproxil's AI verified the test kit and read the outcome.
Participation was voluntary and consented. Personal data was handled under the Nigeria Data Protection Act 2023, and no identifiable data is published on this portal.
Evidence
Headline indicators from each survey instrument. Estimates are released with confidence intervals and state-level breakdowns as data is finalised.
95% CI 85.5–95.2 · 119 / 130 responses
Pregnant subgroup.
95% CI 88.4–97.1 · 112 / 119 responses
Pregnant subgroup who attended ANC.
95% CI 72.3–86.0 · 104 / 130 responses
Denominator = women pregnant now or in last 2 yrs.
95% CI 88.6–92.4 · 798 / 880 responses
95% CI 92.3–95.4 · 837 / 890 responses
Self-report; not a diagnostic-confirmed rate.
95% CI 85.8–90.1 · 748 / 849 responses
Among net owners answering.
95% CI 76.6–81.3 · 901 / 1140 responses
Among those who sought treatment and answered (excl. don't know).
Self-reported
56,713 / 84,245 · PPMV three-month recalled service volumes
AI-validated
157 / 2,066 · Logged mRDT outcomes
Self-reported and AI-validated rates use different denominators and population bases and are not directly equivalent.
Explore LGA
Switch streams to recolour the map. Click any LGA to filter the Results table above.
Survey insights
Six illustrative findings from the SNT survey streams, paired with the underlying statistic from the live data.
Providers' own three-month tallies imply roughly two-thirds of fever cases are positive, but the 2,066 logged rapid tests put validated positivity at 7.6%. The two use different denominators, yet a gap this large points to recall and reporting bias in routine provider data.
7.6% validated vs 67.3% self-reported
Among patients who sought care and answered, 79% were tested before receiving treatment — a solid base for test-and-treat, though one in five still treats presumptively.
79%
About 44% of providers ran out of ACTs in the previous three months, so a patient can reach a stocked-looking shop and still leave without first-line care.
44.2%
Fewer than half of providers keep records of the malaria services they deliver, leaving most private-sector testing and treatment invisible to routine surveillance.
43.6%
Within the small group of women pregnant in the last two years, antenatal attendance and IPTp delivery were both high. Note this is a small subgroup (n=130), so treat the figure as indicative.
94.1%
Supply readiness diverges sharply — 78% of Kano providers stock mRDTs versus 55% in Kaduna, with SP availability splitting similarly. The differences are statistically significant; Kaduna's private sector is materially less equipped.
Kano 78.4% vs Kaduna 55.4%
Findings in context
How our findings sit alongside national surveys and the wider literature. Population prevalence (how many people are infected) and test positivity (how many of those tested are positive) are different measures — each comparison below is labelled accordingly.
Our finding
Across Kano and Kaduna, AI-validated rapid-test positivity was about 7.6%, against roughly 67% positivity that providers reported from their own three-month tallies.
What the literature shows
A 2026 Kano hospital study found expert re-checking of routine slides gave 20.9% positivity versus 64.9% recorded in routine facility data, with routine results correct only about 30% of the time when they read positive. A 2025 Oyo State study reported routine positivity of 73–82% nationally (Jan–Jun 2021) against roughly 35% in therapeutic-efficacy work.[1][2]
Takeaway
Independent studies repeatedly show routinely reported positivity overstates validated positivity in Nigeria. Our finding adds private-sector, barcode-tracked evidence to that pattern.
Our finding
The size of the reported-versus-validated gap points to presumptive over-reporting in private-sector records, which AI validation of each test image helps correct.
Takeaway
Higher apparent positivity from rapid tests than from microscopy is well documented even in gold-standard surveys, which is why image-validated testing matters.
Our finding
Our survey measures test positivity among symptomatic people and provider behaviour, not population infection — so it complements, rather than restates, prevalence surveys.
Takeaway
Kano and Kaduna are high-burden states, but population prevalence is a different quantity from test positivity among the symptomatic. This portal keeps the two distinct.
Our finding
About 79% of patients who sought care reported being tested before treatment, though provider readiness varied — Kano providers were better stocked with rapid tests and SP than Kaduna.
Takeaway
Reported test-before-treat in our private-sector sample exceeds historic public-facility compliance — a policy-relevant contrast worth probing further.
Our finding
Our state-level evidence supports the case for subnational tailoring: the north-west remains high-burden even as the national picture improves.
Takeaway
Long-run national gains coexist with a stubborn north-west burden, which is the rationale for tailoring interventions to states and LGAs.
Why Sproxil
This portal was built from surveys that reach communities in days - powered by a network millions of Nigerians already use, and AI-assisted analysis that turns raw responses into verified findings fast.
respondents in under a week
A single national fielding across all 36 states and the FCT.
Nigerians reached each year
Through our product verification network - the audience our surveys draw on.
faster than traditional fielding
On-site surveys typically take a month or more to reach this many respondents; we did it in a week.
AI-assisted analysis
AI-assisted tooling speeds data processing and analysis; every figure stays human-verified and traceable to source.
Put this evidence to work
Access the data behind this portal, partner on subnational tailoring, or get in touch with the team.
De-identified data and metadata available on request.
Explore MLE and subnational tailoring collaborations.
