Estimating malaria incidence from routine health facility-based surveillance data in Uganda

Decline in malaria test positivity rates following capacity building and archiving of malaria rapid diagnostic test cassettes in Oyo State, Nigeria: a retrospective review of records

BACKGROUND

The malaria test positivity rate (TPR) is a key indicator for evaluating the effectiveness of malaria interventions. In Nigeria, routine data from January to June 2021 reported consistently high TPRs, ranging from 73 to 82%, while Oyo State reported TPRs of 70% to 74% during the same period. These figures were inconsistent with malaria therapeutic efficacy studies conducted between October 2009 and November 2010, which reported a much lower TPR of 35%. This discrepancy raised concerns about data quality, increased malaria incidence, or inaccuracies in malaria diagnosis.

METHODS

This study assessed the effect of two interventions aimed at improving the accuracy of TPR data using secondary quantitative data from the National District Health Information System (NDHIS) for both Primary Healthcare Facilities (PHFs) and Secondary Health Facilities (SHFs). The interventions included (1) facility-level audits of used malaria Rapid Diagnostic Test (RDT) cassettes archived at 733 PHFs, initiated in September 2021, and (2) a 10-day basic malaria microscopy training (BMMT) for Laboratory Scientists at 17 SHFs, completed in September 2021.

RESULTS

At PHFs, the RDT positivity rate declined from 71% in October 2021 to 53% in December 2022. A period review from January to September revealed a decrease in TPR from 62 to 53% in 2022, compared to no difference in TPR for the same period in 2021 with an average TPR of 77%. A paired t-test comparing the mean TPR for each period showed a statistically significant decline of 19.56 (t = 18.081, p < 0.01, CI (17.06-22.05). At SHFs, microscopy-based TPR decreased from 40% in October 2021 to 18% in December 2022. A review of January to September 2021 showed a TPR decline from 53 to 50%, while in 2022, TPR decreased from 25 to 18%. A paired t-test revealed a statistically significant decline of 19.33 in mean TPR at SHFs (t = 8.14, p < 0.01, CI 13.86-24.81).

CONCLUSION

This study highlights the critical role of auditing used RDT cassettes and recommends scaling up this approach in PHFs. It also underscores the value of basic malaria microscopy training in improving the quality and accuracy of microscopy-based diagnosis. One limitation of this study is the absence of comparative data from other states in Nigeria where the interventions were not implemented.

A scoping review of the methods used to estimate health facility catchment populations for child health indicators in sub-Saharan Africa

BACKGROUND

Evidence indicating persistent geographic inequalities in health outcomes signifies a need for routine subnational monitoring of health-related Sustainable Development Goal targets in sub-Saharan Africa. Health facilities may be an appropriate subnational unit for monitoring purposes, but a lack of suitable demographic data complicates the production of baseline facility-level population denominators against which progress can be reliably measured. This scoping review aimed to map the methods and data sources used to estimate health facility catchment areas and translate them to population denominators for child health indicators in the region.

METHODS

Peer-reviewed research publications and grey literature reports were identified by searching bibliographic databases and relevant organisational websites. The inclusion criteria required that studies were conducted in sub-Saharan Africa since January 2000, described quantitative method(s) for estimating health facility catchment areas and/or population denominators, and focussed on children as the population of interest. Following title/abstract then full text screening of search results, relevant data were extracted using a standard form. Thematic analysis was undertaken to extract themes and present a narrative synthesis.

RESULTS

Overall, 33 research publications and 3 grey literature reports were included. Of these, only 7 research studies and 1 technical guidance document outlined aims explicitly framed around methods development and/or evaluation. Studies increasingly estimated catchment areas using complex geostatistical or travel time-based modelling approaches rather than simpler proximity metrics, and produced denominators by intersecting catchment boundaries with gridded population surfaces rather than aggregating area-based administrative counts. Few studies used data produced by or describing health facilities to link estimation methods to service utilisation patterns, inter-facility competition or facility characteristics.

CONCLUSION

There is a need for catchment population estimation methods that can be scaled to national-level facility networks and replicated across the region. This could be achieved by leveraging routinely collected health data and other readily available and nationally consistent data sources. Future methodological development should emphasise modern geostatistical approaches drawing upon the relative strengths of multiple data sources and capturing the range of spatial, supply-side, individual-level and environmental factors with potential to influence catchments' extent, shape and demographic composition.

Evaluating the impact of two next-generation long-lasting insecticidal nets on malaria incidence in Uganda: an interrupted time-series analysis using routine health facility data

INTRODUCTION

Malaria remains a significant public health challenge globally, particularly in sub-Saharan Africa, where progress has stalled in recent years. Long-lasting insecticidal nets (LLINs) are a critical preventive tool against malaria. This study investigated the effectiveness of newer-generation LLINs following a universal coverage campaign in Uganda.

METHODS

Health facility data collected 36 months prior to LLIN distribution and 24 months after LLIN distribution were used from 64 sites that took part in a cluster-randomised trial comparing two newer-generation LLINs (pyrethroid-piperonyl butoxide and pyrethroid-pyriproxyfen). Using an interrupted time-series approach, we compared observed malaria incidence with counterfactual scenarios if no LLINs were distributed, adjusting for precipitation, vegetation, seasonality and care-seeking behaviour. Analyses were also stratified by LLIN type and study-site level estimates of transmission intensity.

RESULTS

Overall, malaria incidence decreased from 827 cases per 1000 person-years in the predistribution period to 538 per 1000 person-years in the postdistribution period. Interrupted time-series analyses estimated a 23% reduction in malaria incidence (incidence rate ratio [IRR]=0.77, 95% CI 0.65 to 0.91) in the first 12 months following distribution relative to what would be expected had no distribution occurred, which was not sustained in the 13-24 month post-distribution period (IRR=0.97, 95% CI 0.75 to 1.28). Findings were similar when stratified by LLIN type. In the first 12 months following distribution, LLIN effectiveness was greater in the high-transmission sites (IRR=0.67, 95% CI 0.54 to 0.86) compared with the medium- (IRR=0.74, 95% CI 0.59 to 0.92) and low-transmission sites (IRR=0.87, 95% CI 0.56 to 1.32).

CONCLUSION

This study demonstrated a modest reduction in malaria incidence following the distribution of newer-generation LLINs that was sustained for only 12 months, highlighting the need for improved strategies to maintain net effectiveness. Adjusting the frequency of universal coverage campaigns based on local malaria transmission intensity may enhance control efforts.

LLIN Evaluation in Uganda Project (LLINEUP2) - Effect of long-lasting insecticidal nets (LLINs) treated with pyrethroid plus pyriproxyfen vs LLINs treated with pyrethroid plus piperonyl butoxide in Uganda: A cluster-randomised trial

Long-lasting insecticidal nets (LLINs) are the cornerstone of malaria control, but their effectiveness is threatened by pyrethroid resistance. We embedded a pragmatic, cluster-randomised trial into Uganda's national LLIN distribution campaign in 2020-2021, comparing pyrethroid-piperonyl butoxide (PBO) LLINs to pyrethroid-pyriproxyfen LLINs. Target communities surrounding public health facilities (clusters, n=64), covering 32 districts were included. Clusters were randomised 1:1 in blocks of two by district to receive: (1) pyrethroid-PBO LLINs (PermaNet 3.0, n=32) or (2) pyrethroid-pyriproxyfen LLINs (Royal Guard, n=32). LLINs were delivered from 7 November 2020 to 26 March 2021. Malaria surveillance data were collected from health facilities from 1 November 2019 until 31 March 2023. Cluster-level estimates of malaria incidence in residents of all ages (primary outcome) were generated from enhanced health facility surveillance data. Cross-sectional community surveys were conducted in randomly selected households (at least 50 per cluster) at 12-months (24 November 2021 to 1 April 2022) and 24-months (23 November 2022 to 21 March 2023) post-LLIN distribution. Overall, 186,364 clinical malaria episodes were diagnosed in cluster residents during 398,931 person-years of follow-up. At 24-months, malaria incidence was lower than baseline in both arms (pyrethroid-PBO: 465 vs 676 episodes per 1000 person-years; pyrethroid-pyriproxyfen: 469 vs 674 episodes per 1000 person-years); but there was no evidence of a difference between the arms (incidence rate ratio 1.06, 95% confidence interval [CI] 0.91-1.22, p=0.47). Two years post-distribution, ownership of at least one LLIN for every two household residents was low in both arms (41.1% pyrethroid-PBO vs 38.6% pyrethroid-pyriproxyfen). Parasite prevalence in children aged 2-10 years was no different between the arms in either survey (24-months: 26.1% pyrethroid-PBO; 29.5% pyrethroid-pyriproxyfen; odds ratio 1.29 [95% CI: 0.81-2.05], p=0.29). The effectiveness of pyrethroid-PBO LLINs and pyrethroid-pyriproxyfen LLINs was no different in Uganda, but two years after mass distribution, LLIN coverage was inadequate. Trial registration: NCT04566510. Registered 28 September 2020, https://clinicaltrials.gov/ct2/show/NCT04566510.

Quality of routine malaria data captured at primary health facilities in the Hohoe Municipality, Ghana

Malaria is a major public health concern and requires quality data management system for effective preventive measures. The District Health Management Information System (DHMIS II) has been used to routinely capture data at health facilities. However, little is known about the quality of routine malaria data captured on the DHIMS II database in Community-Based Health Planning and Service (CHPS) compounds. The study therefore determined the quality of routine malaria data captured on the DHIMS II database in CHPS compounds in the Hohoe Municipality, Ghana. A retrospective cross-sectional analysis of Out Patient Department (OPD) malaria indicators was conducted using data from January 2018 to December 2022 at CHPS compounds in the Hohoe Municipality. Data were collected from three sources: the DHIMS II, monthly morbidity report forms, and consulting room registers. The study assessed three (3) malaria indicators: suspected malaria cases, tested malaria cases, and confirmed OPD malaria cases. A data validation tool was developed to determine the quality of malaria indicators measuring availability, completeness (percentage of missing data), and accuracy. The data was analysed descriptively using Microsoft excel. Out of the four (4) health facilities, 50% (2/4) met the suggested target of (≥ 90%) in 2018 and 2020 whiles all the (4) facilities met the recommended target in 2021 and 2022 for the availability of monthly OPD morbidity reports. For the availability of monthly data returns on anti-malarial, none of the facilities met the recommended target from 2018 to 2022. All 4 facilities met the recommended target in 2021 and 2022. For completeness of source data, 25% of the facility had complete data that met the required target in specific years (2021-2022). For accuracy, 50% of the facilities showed accurate reporting with a Good (± 5%) accuracy level. The remaining 50% underreported data, resulting in a Poor (± 11-20%) accuracy level. The study finds that while half of the facilities had reliable and complete malaria data in their source registers, there are inconsistencies with the DHIMS II database regarding the standards of data quality. Most facilities faced significant issues like unavailability of data, uncompleted data and underreporting of data, making it not-advisable to rely on DHIMS II for critical health decisions. Although, half of the facilities showed evidence of good data quality, there is still a need for improvement in the capturing of routine malaria data.

Time series analysis of malaria cases to assess the impact of various interventions over the last three decades and forecasting malaria in India towards the 2030 elimination goals

BACKGROUND

Despite the progress made in this decade towards malaria elimination, it remains a significant public health concern in India and many other countries in South Asia and Asia Pacific region. Understanding the historical trends of malaria incidence in relation to various commodity and policy interventions and identifying the factors associated with its occurrence can inform future intervention strategies for malaria elimination goals.

METHODS

This study analysed historical malaria cases in India from 1990 to 2022 to assess the annual trends and the impact of key anti-malarial interventions on malaria incidence. Factors associated with malaria incidence were identified using univariate and multivariate linear regression analyses. Generalized linear, smoothing, autoregressive integrated moving averages (ARIMA) and Holt's models were used to forecast malaria cases from 2023 to 2030.

RESULTS

The reported annual malaria cases in India during 1990-2000 were 2.38 million, which dropped to 0.73 million cases annually during 2011-2022. The overall reduction from 1990 (2,018,783) to 2022 (176,522) was 91%. The key interventions of the Enhanced Malaria Control Project (EMCP), Intensified Malaria Control Project (IMCP), use of bivalent rapid diagnostic tests (RDT-Pf/Pv), artemisinin-based combination therapy (ACT), and involvement of the Accredited Social Health Activists (ASHAs) as front-line workers were found to result in the decline of malaria significantly. The ARIMA and Holt's models projected a continued decline in cases with the potential for reaching zero indigenous cases by 2027-2028. Important factors influencing malaria incidence included tribal population density, literacy rate, health infrastructure, and forested and hard-to-reach areas.

CONCLUSIONS

Studies aimed at assessing the impact of major commodity and policy interventions on the incidence of disease and studies of disease forecasting will inform programmes and policymakers of steps needed during the last mile phase to achieve malaria elimination. It is proposed that these time series and disease forecasting studies should be performed periodically using granular (monthly) and meteorological data to validate predictions of prior studies and suggest any changes needed for elimination efforts at national and sub-national levels.

Impact of the Introduction of a Package of Diagnostic Tools, Diagnostic Algorithm, and Training and Communication on Outpatient Acute Fever Case Management at 3 Diverse Sites in Uganda: Results of a Randomized Controlled Trial

BACKGROUND

Increasing trends of antimicrobial resistance are observed around the world, driven in part by excessive use of antimicrobials. Limited access to diagnostics, particularly in low- and middle-income countries, contributes to diagnostic uncertainty, which may promote unnecessary antibiotic use. We investigated whether introducing a package of diagnostic tools, clinical algorithm, and training-and-communication messages could safely reduce antibiotic prescribing compared with current standard-of-care for febrile patients presenting to outpatient clinics in Uganda.

METHODS

This was an open-label, multicenter, 2-arm randomized controlled trial conducted at 3 public health facilities (Aduku, Nagongera, and Kihihi health center IVs) comparing the proportions of antibiotic prescriptions and clinical outcomes for febrile outpatients aged ≥1 year. The intervention arm included a package of point-of-care tests, a diagnostic and treatment algorithm, and training-and-communication messages. Standard-of-care was provided to patients in the control arm.

RESULTS

A total of 2400 patients were enrolled, with 49.5% in the intervention arm. Overall, there was no significant difference in antibiotic prescriptions between the study arms (relative risk [RR]: 1.03; 95% CI: .96-1.11). In the intervention arm, patients with positive malaria test results (313/500 [62.6%] vs 170/473 [35.9%]) had a higher RR of being prescribed antibiotics (1.74; 1.52-2.00), while those with negative malaria results (348/688 [50.6%] vs 376/508 [74.0%]) had a lower RR (.68; .63-.75). There was no significant difference in clinical outcomes.

CONCLUSIONS

This study found that a diagnostic intervention for management of febrile outpatients did not achieve the desired impact on antibiotic prescribing at 3 diverse and representative health facility sites in Uganda.

Mapping malaria incidence using routine health facility surveillance data in Uganda

INTRODUCTION

Maps of malaria risk are important tools for allocating resources and tracking progress. Most maps rely on cross-sectional surveys of parasite prevalence, but health facilities represent an underused and powerful data source. We aimed to model and map malaria incidence using health facility data in Uganda.

METHODS

Using 24 months (2019-2020) of individual-level outpatient data collected from 74 surveillance health facilities located in 41 districts across Uganda (n=445 648 laboratory-confirmed cases), we estimated monthly malaria incidence for parishes within facility catchment areas (n=310) by estimating care-seeking population denominators. We fit spatio-temporal models to the incidence estimates to predict incidence rates for the rest of Uganda, informed by environmental, sociodemographic and intervention variables. We mapped estimated malaria incidence and its uncertainty at the parish level and compared estimates to other metrics of malaria. To quantify the impact that indoor residual spraying (IRS) may have had, we modelled counterfactual scenarios of malaria incidence in the absence of IRS.

RESULTS

Over 4567 parish-months, malaria incidence averaged 705 cases per 1000 person-years. Maps indicated high burden in the north and northeast of Uganda, with lower incidence in the districts receiving IRS. District-level estimates of cases correlated with cases reported by the Ministry of Health (Spearman's r=0.68, p<0.0001), but were considerably higher (40 166 418 cases estimated compared with 27 707 794 cases reported), indicating the potential for underreporting by the routine surveillance system. Modelling of counterfactual scenarios suggest that approximately 6.2 million cases were averted due to IRS across the study period in the 14 districts receiving IRS (estimated population 8 381 223).

CONCLUSION

Outpatient information routinely collected by health systems can be a valuable source of data for mapping malaria burden. National Malaria Control Programmes may consider investing in robust surveillance systems within public health facilities as a low-cost, high benefit tool to identify vulnerable regions and track the impact of interventions.

East Africa International Center of Excellence for Malaria Research: Impact on Malaria Policy in Uganda

Malaria is the leading cause of disease burden in sub-Saharan Africa. In 2010, the East Africa International Center of Excellence for Malaria Research, also known as the Program for Resistance, Immunology, Surveillance, and Modeling of Malaria (PRISM), was established to provide a comprehensive approach to malaria surveillance in Uganda. We instituted cohort studies and a robust malaria and entomological surveillance network at selected public health facilities that have provided a platform for monitoring trends in malaria morbidity and mortality, tracking the impact of malaria control interventions (indoor residual spraying of insecticide [IRS], use of long-lasting insecticidal nets [LLINs], and case management with artemisinin-based combination therapies [ACTs]), as well as monitoring of antimalarial drug and insecticide resistance. PRISM studies have informed Uganda's malaria treatment policies, guided selection of LLINs for national distribution campaigns, and revealed widespread pyrethroid resistance, which led to changes in insecticides delivered through IRS. Our continuous engagement and interaction with policy makers at the Ugandan Ministry of Health have enabled PRISM to share evidence, best practices, and lessons learned with key malaria stakeholders, participate in malaria control program reviews, and contribute to malaria policy and national guidelines. Here, we present an overview of interactions between PRISM team members and Ugandan policy makers to demonstrate how PRISM's research has influenced malaria policy and control in Uganda.

Establishing a Public Health Emergency Operations Center in an Outbreak-Prone Country: Lessons Learned in Uganda, January 2014 to December 2021

Uganda is highly vulnerable to public health emergencies (PHEs) due to its geographic location next to the Congo Basin epidemic hot spot, placement within multiple epidemic belts, high population growth rates, and refugee influx. In view of this, Uganda's Ministry of Health established the Public Health Emergency Operations Center (PHEOC) in September 2013, as a central coordination unit for all PHEs in the country. Uganda followed the World Health Organization's framework to establish the PHEOC, including establishing a steering committee, acquiring legal authority, developing emergency response plans, and developing a concept of operations. The same framework governs the PHEOC's daily activities. Between January 2014 and December 2021, Uganda's PHEOC coordinated response to 271 PHEs, hosted 207 emergency coordination meetings, trained all core staff in public health emergency management principles, participated in 21 simulation exercises, coordinated Uganda's Global Health Security Agenda activities, established 6 subnational PHEOCs, and strengthened the capacity of 7 countries in public health emergency management. In this article, we discuss the following lessons learned: PHEOCs are key in PHE coordination and thus mitigate the associated adverse impacts; although the functions of a PHEOC may be legalized by the existence of a National Institute of Public Health, their establishment may precede formally securing the legal framework; staff may learn public health emergency management principles on the job; involvement of leaders and health partners is crucial to the success of a public health emergency management program; subnational PHEOCs are resourceful in mounting regional responses to PHEs; and service on the PHE Strategic Committee may be voluntary.

Malaria outbreak facilitated by engagement in activities near swamps following increased rainfall and limited preventive measures: Oyam District, Uganda

In April 2019, the District Health Office of Oyam District, Uganda reported an upsurge in malaria cases exceeding expected epidemic thresholds, requiring outbreak response. We investigated the scope of outbreak and identified exposures for transmission to inform control measures. A confirmed case was a positive malaria rapid diagnostic test or malaria microscopy from 1 January-30 June 2019 in a resident or visitor of Acaba Sub-county, Oyam District. We reviewed medical records at health facilities to get case-patients. We conducted entomological and environmental assessments to determine vector density, and identify aquatic Anopheles habitats, conducted a case-control study to determine exposures associated with illness. Of 9,235 case-patients (AR = 33%), females (AR = 38%) were more affected than males (AR = 20%) (p<0.001). Children <18 years were more affected (AR = 37%) than adults (p<0.001). Among 83 case-patients and 83 asymptomatic controls, 65 (78%) case-patients and 33 (40%) controls engaged in activities <500m from a swamp (ORMH = 12, 95%CI 3.6-38); 18 (22%) case-patients and four (5%) controls lived <500m from rice irrigation sites (ORMH = 8.2, 95%CI 1.8-36); and 23 (28%) case-patients and four (5%) controls had water pools <100m from household for 3-5 days after rainfall (ORMH = 7.3, 95%CI 2.2-25). Twenty three (28%) case-patients and four (5%) controls did not sleep under bed nets the previous night (ORMH = 20, 95%CI 2.7-149); 68 (82%) case-patients and 43(52%) controls did not wear long-sleeved clothes during evenings (ORMH = 9.3, 95%CI 2.8-31). Indoor resting vector density was 4.7 female mosquitoes/household/night. All Anopheles aquatic habitats had Anopheles larvae. Weekly rainfall in 2019 was heavier (6.0±7.2mm) than same period in 2018 (1.8±1.8mm) (p = 0.006). This outbreak was facilitated by Anopheles aquatic habitats near homes created by human activities, following increased rainfall compounded by inadequate use of individual preventive measures. We recommended awareness on use of insecticide-treated bed nets, protective clothing, and avoiding creation of Anopheles aquatic habitats.

Modeling DREAMS impact: trends in new HIV diagnoses among women attending antenatal care clinics in DREAMS countries

OBJECTIVES

To understand the impact of United States President's Emergency Plan for AIDS Relief (PEPFAR's) DREAMS (Determined, Resilient, Empowered, AIDS-Free, Mentored, and Safe) Partnership on new HIV diagnoses among women in antenatal care (ANC) settings in 10 African countries from 2015 to 2020.

DESIGN

We modeled spatiotemporal changes in new HIV diagnoses among women in ANC settings using PEPFAR data. Statistical tests were performed in R to compare differences in new diagnoses rates between DREAMS and non-DREAMS subnational units (SNUs) and to explore predictors of new diagnoses declines within DREAMS SNUs.

METHODS

We used a predictive geospatial model to forecast the rate of new diagnoses for each time period in a 5 km grid cell (n = 861 SNUs). Linear model analyses were conducted using predictor variables: urbanicity, DREAMS geographic footprint, 'layering' proxy, and community-level male viral load suppression.

RESULTS

New HIV diagnoses in ANC from 2015 to 2020 declined in nearly all SNUs. 'Always' DREAMS SNUs reported declines of 45% while 'Never' DREAMS SNUs reported a decline of only 37% (F = 8.1, 1 and 829 DF, P < 0.01). Within Always DREAMS SNUs, greater declines were seen in areas with a higher number of minimum services in their DREAMS primary package (t = 2.77, P < 0.01).

CONCLUSION

New HIV diagnoses among women are declining in both DREAMS and non-DREAMS SNUs; mirroring HIV incidence decreases and reflecting increasing community viral load suppression and voluntary male medical circumcision rates. DREAMS programming may have contributed to accelerated declines of new HIV diagnoses in DREAMS SNUs compared with non-DREAMS SNUs. Increased progress is needed to further reduce the disparities between adolescent girls and young women (AGYW) and young men to achieve epidemic control.

Stratification of malaria incidence in Papua New Guinea (2011-2019): Contribution towards a sub-national control policy

Malaria risk in Papua New Guinea (PNG) is highly heterogeneous, between and within geographical regions, which is operationally challenging for control. To enhance targeting of malaria interventions in PNG, we investigated risk factors and stratified malaria incidence at the level of health facility catchment areas. Catchment areas and populations of 808 health facilities were delineated using a travel-time accessibility approach and linked to reported malaria cases (2011-2019). Zonal statistics tools were used to calculate average altitude and air temperature in catchment areas before they were spatially joined with incidence rates. In addition, empirical Bayesian kriging (EBK) was employed to interpolate incidence risk strata across PNG. Malaria annual incidence rates are, on average, 186.3 per 1000 population in catchment areas up to 600 m, dropped to 98.8 at (800-1400) m, and to 24.1 cases above 1400 m altitude. In areas above the two altitudinal thresholds 600m and 1400m, the average annual temperature drops below 22°C and 17°C, respectively. EBK models show very low- to low-risk strata (<100 cases per 1000) in the Highlands, National Capital District and Bougainville. In contrast, patches of high-risk (>200 per 1000) strata are modelled mainly in Momase and Islands Regions. Besides, strata with moderate risk (100-200) predominate throughout the coastal areas. While 35.7% of the PNG population (estimated 3.33 million in 2019) lives in places at high or moderate risk of malaria, 52.2% (estimated 4.88 million) resides in very low-risk areas. In five provinces, relatively large proportions of populations (> 50%) inhabit high-risk areas: New Ireland, East and West New Britain, Sandaun and Milne Bay. Incidence maps show a contrast in malaria risk between coastal and inland areas influenced by altitude. However, the risk is highly variable in low-lying areas. Malaria interventions should be guided by sub-national risk levels in PNG.