Modelling malaria routine surveillance data to inform seasonal malaria chemoprevention strategy in Moissala, Southern Chad

This study utilized a climate-informed mathematical model calibrated with routine surveillance data from Moissala, Chad, to demonstrate that seasonal malaria chemoprevention reduced under-five malaria cases by 26% and identified a five-round schedule starting in mid-June as the optimal strategy for maximizing case reduction.

The Gist

Imagine malaria as a relentless, seasonal storm that sweeps through the southern part of Chad every year, bringing sickness and danger, especially to young children. For years, health workers have tried to build a "shield" against this storm using a medicine called Seasonal Malaria Chemoprevention (SMC). Think of SMC as a monthly dose of medicine given to children under five, acting like a temporary umbrella that keeps them dry during the rainy, high-risk months.

But here's the tricky part: Just like weather patterns change, the "storm" of malaria doesn't hit the same way every year or in every place. In the town of Moissala, the rainy season is long and starts early. The big question was: Is the current shield strong enough? Is it being put up at the right time? And what happens if we stop building it for a year?

To answer these questions, the researchers in this paper didn't just count sick kids; they built a digital "weather forecast" for malaria.

The Digital Crystal Ball

The scientists created a computer model—a sort of "malaria simulator." Imagine a giant, complex video game where they programmed in:

• The Weather: How much it rains and how hot it gets (since mosquitoes love heat and puddles).
• The People: How many children live there and how their immunity changes as they grow up.
• The Shield (SMC): How well the medicine works and how many kids actually get it.

They fed this simulator real data from 2018 to 2023. This allowed them to run "what-if" scenarios. They could pause the game, turn off the medicine, and see what would have happened if they hadn't given the shots. They could also try different schedules, like starting the medicine earlier or giving it for more months, to see which version saved the most lives.

What the Simulator Told Them

1. The Shield Works, But It's Not Perfect
The model showed that the SMC program is a hero. Between 2018 and 2023, it prevented about 14,400 cases of malaria every year in children under five. That's like saving a whole small town from getting sick annually. Without this shield, the number of sick kids would have been 26% higher.

2. The "Great Pause" of 2019
In 2019, the program was accidentally paused because of administrative rules. The simulator showed this was a disaster. When the shield was taken down, malaria cases spiked by 31%. It was like taking down the roof of a house right before a hurricane; the damage was immediate and severe. This proved that stopping the program, even for a year, has a huge cost.

3. Timing is Everything
The researchers tested different schedules. They found that the old plan (starting in July with 4 rounds) was good, but not the best.

• Adding a Round: Because the rainy season in Moissala is long, adding a 5th round of medicine helped reduce cases by another 7%.
• Starting Earlier: The rain starts earlier in Moissala than in other parts of the country. The model showed that starting the 5 rounds in mid-June (instead of July) was the "Goldilocks" strategy—just right. It caught the storm before it got too strong.

The Big Takeaway

This study is like a mechanic tuning a race car. They realized that while the car (the SMC program) was fast, the engine needed a tweak to handle the specific track (Moissala's climate).

The main lesson is that one size does not fit all. What works in a dry, short-rain season might fail in a long, wet one. By using this "digital crystal ball," health officials in Chad can now say with confidence: "We need to start our medicine in June and give it five times a year to keep our children safe."

Why This Matters for Everyone

This isn't just about Chad. The researchers made their "malaria simulator" free and open-source (like a free app anyone can download). This means health workers in other countries with different climates can use the same tool to figure out their own perfect schedule.

In a world where climate change is making weather patterns more unpredictable, having a tool that can predict the best way to fight disease is like having a superpower. It turns guesswork into a precise plan, ensuring that every dose of medicine goes exactly where it's needed most.

Technical Summary

1. Problem Statement

Seasonal Malaria Chemoprevention (SMC) is a critical intervention involving the monthly administration of antimalarial drugs (sulfadoxine-pyrimethamine plus amodiaquine) to children under five during high-transmission seasons. While historically standardized for the Sahel region, recent World Health Organization (WHO) guidelines encourage tailoring SMC strategies (timing and number of rounds) to local transmission patterns, particularly in areas with longer transmission seasons like Southern Chad.

The Moissala health district in Chad presents a complex scenario:

• Heterogeneous Transmission: It has a longer rainy season (May–October) and a sustained malaria transmission period compared to northern Sahelian districts.
• Programmatic Instability: SMC implementation has fluctuated (4 rounds 2013–2018; suspended in 2019; resumed 2020; expanded to 5 rounds in 2021; start date shifted to June in 2023).
• Evaluation Challenge: Simple year-to-year comparisons of malaria incidence are misleading due to significant inter-annual variability in climate (rainfall and temperature), which drives vector dynamics. There is a need for a robust analytical framework to disentangle the effects of climate variability from the impact of SMC interventions and to identify the optimal strategy for this specific setting.

2. Methodology

The authors developed a climate-informed, compartmental malaria transmission model calibrated to routine surveillance data.

Data Sources (2018–2023)

• Incidence: Weekly malaria cases in children <5 years from health facilities in Moissala.
• Coverage: Annual household survey data (Epicentre/MSF) estimating SMC coverage (using verbal confirmation + card presentation).
• Climate: Daily rainfall (CHIRPS) and monthly temperature (ERA5-Land), smoothed and lagged to reflect biological delays in transmission.
• Demographics: Population estimates adjusted for growth rates and age structure (under-5 vs. 5+).

Model Structure

• Framework: A deterministic compartmental model (ODEs) adapted from Ukawuba and Shaman, stratified into two age groups: <5 years (target of SMC) and ≥5 years.
• Compartments: Susceptible (S), Exposed (E), Infected Untreated (IU), Infected Treated (IT), and Asymptomatic Infected (IA).
• Climate Drivers: The Entomological Inoculation Rate (EIR) is modeled as a function of:
  • Human Infectiousness: A saturating function of prevalence.
  • Temperature: An asymmetric Gaussian function allowing transmission persistence at high temperatures (>32°C), with a lag effect.
  • Rainfall: A logistic function representing the relationship between rainfall and vector abundance, with a lag effect.
• SMC Effect: Modeled as a time-varying reduction in the force of infection ($\omega_t$) for the under-5 population, dependent on coverage, intrinsic efficacy, and a decay factor (drug efficacy waning over time).

Inference and Simulation

• Calibration: A Bayesian inference framework using Adaptive Metropolis–Hastings (via the mcstate R package) to estimate 11 parameters, including SMC intrinsic efficacy ($eff_{SMC}$), climate lag parameters, and observation noise.
• Counterfactual Analysis: The calibrated model was used to simulate:
  1. Overall Effectiveness: Comparing the implemented strategy vs. a "No SMC" scenario.
  2. Strategy Changes: Assessing the impact of the 2019 suspension, the 2021 addition of a 5th round, and the 2023 shift to a June start date.
  3. Optimization: Simulating 8 scenarios (varying rounds: 4 vs. 5; start dates: June 1, June 15, July 1, July 15) to find the strategy minimizing cases.

3. Key Results

Model Calibration

The model successfully captured the overall trends of weekly malaria cases in children under five, with all observations falling within the 95% prediction intervals. It accounted for the 2019 spike (due to suspension and high rainfall) and the subsequent reductions.

SMC Effectiveness (2018–2023)

• Intrinsic Efficacy: The estimated reduction in the force of infection due to SMC ($eff_{SMC}$) was 68% (95% CI: 60–76%).
• Overall Impact: SMC reduced malaria cases in children under five by 26% (95% CI: 21–31%) compared to a no-SMC scenario.
• Cases Averted: Approximately 14,400 cases were averted annually during the high-transmission season.

Impact of Specific Strategy Changes

• 2019 Suspension: Halting SMC led to an estimated 31% increase in cases (approx. 13,600 additional cases), highlighting the intervention's critical role even amidst high rainfall.
• 2021 Expansion (4 to 5 rounds): Adding a fifth round reduced cases by 7% (approx. 3,000 fewer cases/year) compared to a 4-round schedule.
• 2023 Timing Shift (July to June): Starting the 5-round schedule in June instead of July resulted in an additional 5% reduction (approx. 2,100 fewer cases/year).

Optimal Strategy

Simulation of various start dates and round counts identified the optimal strategy as 5 rounds of SMC beginning on June 15th. This strategy yielded the greatest reduction in cases (4,300 fewer cases/year) compared to the historical baseline of 4 rounds starting July 15th.

4. Key Contributions

1. Methodological Innovation: The study demonstrates the utility of climate-informed mechanistic modeling using routine surveillance data to evaluate public health interventions. This approach effectively controls for climatic confounding, which simple statistical comparisons cannot do.
2. Operational Guidance: It provides evidence-based recommendations for the Moissala district (and similar Southern Chad regions) to adopt a 5-round SMC schedule starting in mid-June, aligning with the longer transmission season.
3. Open-Source Tooling: The authors released the model as an open-source R package (malclimsim), enabling other researchers and national programs to adapt the framework to their local contexts using routine data and climate inputs.
4. Real-World Effectiveness: The study quantifies the gap between efficacy (RCTs, 73% reduction) and real-world effectiveness (26% reduction), attributing the difference to operational factors like coverage and adherence, thereby emphasizing the need for routine monitoring.

5. Significance

This work addresses a critical gap in malaria control: how to optimize SMC in regions with longer transmission seasons where standard Sahelian protocols are suboptimal. By validating that an earlier start date and an additional round significantly reduce the disease burden, the study supports the WHO's shift toward flexible, context-specific SMC guidelines. Furthermore, it establishes a replicable framework for using routine data and climate modeling to guide dynamic public health decision-making in the face of climate variability and changing transmission dynamics. The findings directly informed the National Malaria Control Programme (PNLP) and implementing partners (MSF) in Chad to refine their SMC strategy.