Health care provider practices in diagnosis and treatment of malaria in rural communities in Kisumu County, Kenya

Improving transparency in malaria programme funds: A business case for connected diagnostics in Kenya

Background

International vertical health financing programmes risk functioning in parallel with existing domestic funding in low- and middle-income countries (LMICs), leading to inefficient service delivery and concomitant poor health outcomes.

Aim

We assessed the opportunities offered by digitalised diagnostics (ConnDx) to target and monitor health funds to those in need objectively and transparently.

Setting

ConnDx was rolled out in five private health facilities in Kisumu, Kenya.

Methods

The ConnDx process was codeveloped with the local Department of Health of Kisumu. We used the quantitative data generated by ConnDx. We also calculated the costs for ConnDx and standard care to assess potential cost reductions.

Results

In total, 2199 malaria cases were detected among 11 689 patients with fever. ConnDx allowed for the identification of malaria hotspots, semi-real-time assessment of patient health seeking behaviour across facilities and insights in doctor's prescription behaviours. Based on these insights, we estimated a 25% reduction in costs can be realised, while simultaneously better quality indicators can be monitored.

Conclusion

The concept of ConnDx can be used for any medical condition that can be diagnosed in a digital manner and linked to mobile payment systems. The generated data can contribute to better quality services for individual patients while at the same time support local health policy makers and managers for more targeted interventions.

Contribution

The ConnDx approach can help decision makers in LMICs to channel disease-specific funds to the right patients for the right disease at the right time, which can potentially accelerate the way to universal health coverage.

The feasibility and impact of deploying a four-tests panel at antenatal care in primary health care facilities of a developing country, Kenya

Introduction

Contracting HIV, syphilis, hepatitis B virus (HBV), and malaria during pregnancy significantly affects the health of the woman, the pregnancy, and the unborn child. The World Health Organization (WHO) recommends testing pregnant women for these infections to achieve triple elimination of mother-to-child transmissions. However, this goal has not been fully realized in low- to medium-income countries, primarily due to segmented testing practices. This study aimed to investigate the effect of introducing a four-tests panel on the quality of antenatal care (ANC) among pregnant women attending selected Primary Health Care facilities in Kenya.

Methods

Using a multi-design approach, we analyzed ANC medical records from 577 pregnant women attending eight facilities across four different counties. Blood from the women fingerpick was tested for HIV, Syphilis Hepatitis B Virus and Malaria using the four-tests panel and the results compared to those in the medical records.

Results

Out of 577 ANC women, only 8.3% had test results for all four infections available. The majority of the mothers had been tested for syphilis (93.7%), HIV (78.5%), and malaria (62.6%), only 19.5% had been tested for HBV. Testing the women using the 4-tests panel yielded positivity rates of 6.9% for HIV, 0.9% for syphilis, 1.9% for malaria, and 1.1% for HBV. Among those without previous test results, the positivity rate was 2.8% for syphilis, 13.8% for HIV (with 10.6% testing positive for recent p24 infections, F = 24.876, p < 0.001), 2.3% for malaria, and 4.5% for HBV, with 83.3% of these individuals having no prior test results. The mean positivity rate of those tested using the 4-tests panel compared to segmented single tests was significantly different. The panel was cost-effective and user-friendly for healthcare workers, and in facilities facing staff shortages, it reduced turnaround time and workloads by half. The use of the panel also improved the profiling of ANC mothers and enhanced data management for the four infections by 91.7%.

Conclusion

Adopting the 4-tests panel has the potential to improve test result outputs, enhance the quality-of-service delivery, and contribute significantly to the achievement of triple elimination goals.

Evaluating the gap in rapid diagnostic testing: insights from subnational Kenyan routine health data

BACKGROUND

Understanding diagnostic capacities is essential to addressing healthcare provision and inequity, particularly in low-income and middle-income countries. This study used routine data to assess trends in rapid diagnostic test (RDT) reporting, supplies and unmet needs across national and 47 subnational (county) levels in Kenya.

METHODS

We extracted facility-level RDT data for 19 tests (2018-2020) from the Kenya District Health Information System, linked to 13 373 geocoded facilities. Data quality was assessed for reporting completeness (ratio of reports received against those expected), reporting patterns and outliers. Supply assessment covered 12 RDTs reported by at least 50% of the reporting facilities (n=5251), with missing values imputed considering reporting trends. Supply was computed by aggregating the number of tests reported per facility. Due to data limitations, demand was indirectly estimated using healthcare-seeking rates (HIV, malaria) and using population data for venereal disease research laboratory test (VDRL), with unmet need computed as the difference between supply and demand.

RESULTS

Reporting completeness was under 40% across all counties, with RDT-specific reporting ranging from 9.6% to 89.6%. Malaria RDTs showed the highest annual test volumes (6.3-8.0 million) while rheumatoid factor was the lowest (0.5-0.7 million). Demand for RDTs varied from 2.5 to 11.5 million tests, with unmet needs between 1.2 and 3.5 million. Notably, malaria testing and unmet needs were highest in Turkana County, as well as the western and coastal regions. HIV testing was concentrated in the western and central regions, with decreasing unmet needs from 2018 to 2020. VDRL testing showed high volumes and unmet needs in Nairobi and select counties, with minimal yearly variation.

CONCLUSION

RDTs are crucial in enhancing diagnostic accessibility, yet their utilisation varies significantly by region. These findings underscore the need for targeted interventions to close testing gaps and improve data reporting completeness. Addressing these disparities is vital for equitably enhancing diagnostic services nationwide.

Adherence to national malaria treatment guidelines in private drug outlets: a cross-sectional survey in the malaria-endemic Kisumu County, Kenya

BACKGROUND

Malaria prevalence in Kenya is 6%, with a three-fold higher prevalence in western Kenya. Adherence to malaria treatment guidelines improves care for suspected malaria cases and can reduce unnecessary anti-malarial use. Data on adherence to guidelines in retail drug outlets (DOs) is limited, yet approximately 50% of people with fever access treatment first in these outlets. This study assessed adherence to the national malaria treatment guidelines among DOs in a high transmission area of Western Kenya.

METHODS

In a cross-sectional survey of DOs in Kisumu Central and Seme sub-counties in 2021, DO staff were interviewed using structured questionnaires to assess outlet characteristics (location, testing services), staff demographics (age, sex, training), and health system context (supervision, inspection). Mystery shoppers (research assistants disguised as clients) observed malaria management practices and recorded observations on a standardized tool. Adherence was defined as dispensing artemether-lumefantrine (AL) to patients with a confirmed positive test, accompanied by appropriate medication counseling. Logistic regression was used to test for association between adherence to guidelines and DO-related factors.

RESULTS

None of the 70 DOs assessed had a copy of the guidelines, and 60 (85.7%) were in an urban setting. Staff adhered to the guidelines in 14 (20%) outlets. The odds of adherence were higher among staff who had a bachelor's degree {odds ratio (OR) 6.0, 95% confidence interval (95% CI) 1.66-21.74}, those trained on malaria rapid diagnostic test (RDT) {OR 4.4, 95% CI 1.29-15.04}, and those who asked about patient's symptoms {OR 3.6, 95% CI 1.08-12.25}. DOs that had higher odds of adherence included those with functional thermometers {OR 5.3, 95% CI 1.46-19.14}, those recently inspected (within three months) by Pharmacy and Poisons Board (PPB) {OR 9.4, 95% CI 2.55-34.67}, and those with all basic infrastructure {OR 3.9, 95% CI 1.01-15.00}. On logistic regression analysis, recent PPB inspection {adjusted OR (AOR) 4.6, 95% CI 1.03-20.77} and malaria RDT-trained staff (aOR 4.5, 95% CI 1.02-19.84) were independently associated with adherence.

CONCLUSION

Most outlets didn't adhere to malaria guidelines. Regular interaction with regulatory bodies could improve adherence. Ministry of Health should enhance private sector engagement and train DOs on RDT use.

Perceptions and drivers of healthcare provider and drug dispenser practices for the treatment of malaria in pregnancy in the context of multiple first-line therapies in western Kenya: a qualitative study

BACKGROUND

Emergence of Plasmodium falciparum resistance to artemether-lumefantrine in Africa prompted the pilot introduction of multiple first-line therapies (MFT) against malaria in Kenya, potentially exposing women-of-childbearing-age (WOCBAs) to anti-malarials with unknown safety profiles in the first trimester. This qualitative study explored knowledge and perceptions among healthcare providers providing malaria treatment to WOCBAs and pregnant women.

METHODS

In-depth interviews were conducted with purposively selected public and private health facility (HF) and drug outlet (DO) providers within and outside the pilot-MFT area. County health managers were interviewed about their knowledge of the national treatment guidelines. Transcripts were coded by content analysis using the World Health Organization health system building blocks (leadership/governance, financing, health workforce, health information systems, access to medicines, and service delivery).

RESULTS

Thirty providers (HF:21, DO:9) and three health managers were interviewed. Eighteen providers were from HFs in the pilot-MFT area; the remaining three and all nine DOs were outside the pilot-MFT area. The analysis revealed that providers had not been trained in malaria case management in the previous twelve months. DO providers were unfamiliar with national treatment guidelines in pregnancy and reported having no pregnancy tests. Health managers were unable to supervise DOs due to resource limitations. Providers from HFs and DOs noted poor sensitivity of malaria rapid diagnostic tests (RDTs) and hesitancy among patients who associated malaria-RDTs with HIV testing. Almost all providers reported anti-malarial stock-outs, with quinine most affected. Patient preference was a major factor in prescribing anti-malarials. Providers in HFs and DOs reported preferentially using artemether-lumefantrine in the first trimester due to the side effects and unavailability of quinine.

CONCLUSION

Knowledge of malaria case management in drug outlets and health facilities remains poor. Improved regulation of DO providers is warranted. Optimizing treatment of malaria in pregnancy requires training, availability of malaria commodities, and pregnancy tests.

Healthcare provider and drug dispenser knowledge and adherence to guidelines for the case management of malaria in pregnancy in the context of multiple first-line artemisinin-based combination therapy in western Kenya

BACKGROUND

Concerns about emerging resistance to artemether-lumefantrine (AL) in Africa prompted the pilot introduction of multiple first-line therapies (MFT) in Western Kenya, potentially exposing women-of-childbearing-age (WOCBA) to anti-malarials with unknown safety profiles in the first trimester. The study assessed healthcare provider knowledge and adherence to national guidelines for managing malaria in pregnancy in the context of the MFT pilot.

METHODS

From March to April 2022, a cross-sectional study was conducted in 50 health facilities (HF) and 40 drug outlets (DO) using structured questionnaires to assess pregnancy detection, malaria diagnosis, and treatment choices by trimester. Differences between HF and DO providers and between MFT and non-MFT HFs were assessed using Chi-square tests.

RESULTS

Of 174 providers (77% HF, 23% DO), 56% were from MFT pilot facilities. Most providers had tertiary education; 5% HF and 20% DO had only primary or secondary education. More HF than DO providers had knowledge of malaria treatment guidelines (62% vs. 40%, p = 0.023), received training in malaria in pregnancy (49% vs. 20%, p = 0.002), and reported assessing for pregnancy in WOCBA (98% vs. 78%, p < 0.001). Most providers insisted on parasitological diagnosis, with 59% HF using microscopy and 85% DO using rapid diagnostic tests. More HF than DO providers could correctly name the drugs for treating uncomplicated malaria in the first trimester (oral quinine, or AL if quinine is unavailable) (90% vs. 58%, p < 0.001), second and third trimesters (artemisinin-based combination therapy) (84% vs. 70%, p = 0.07), and for severe malaria (parenteral artesunate/artemether) (94% vs. 60%, p < 0.001). Among HF providers, those in the MFT pilot had more knowledge of malaria treatment guidelines (67% vs. 49%, p = 0.08) and had received training on treatment of malaria in pregnancy (56% vs. 32%, p = 0.03). Few providers (10% HF and 12% DO) had adequate knowledge of malaria treatment in pregnancy, defined as the correct drug and dose for uncomplicated and severe malaria in all trimesters.

CONCLUSIONS

Knowledge of national malaria in pregnancy treatment guidelines among providers in Western Kenya is suboptimal. Robust training on appropriate anti-malarial and dosage is needed, particularly given the recent change in recommendation for artemether-lumefantrine use in the first trimester. Supervision of DO and HF practices is essential for correct treatment of malaria in pregnancy in the context of MFT programmes.

Availability of malaria diagnostic tests, anti-malarial drugs, and the correctness of treatment: a systematic review and meta-analysis

BACKGROUND

Health facilities' availability of malaria diagnostic tests and anti-malarial drugs (AMDs), and the correctness of treatment are critical for the appropriate case management, and malaria surveillance programs. It is also reliable evidence for malaria elimination certification in low-transmission settings. This meta-analysis aimed to estimate summary proportions for the availability of malaria diagnostic tests, AMDs, and the correctness of treatment.

METHODS

The Web of Science, Scopus, Medline, Embase, and Malaria Journal were systematically searched up to 30th January 2023. The study searched any records reporting the availability of diagnostic tests and AMDs and the correctness of malaria treatment. Eligibility and risk of bias assessment of studies were conducted independently in a blinded way by two reviewers. For the pooling of studies, meta-analysis using random effects model were carried out to estimate summary proportions of the availability of diagnostic tests, AMDs, and correctness of malaria treatment.

RESULTS

A total of 18 studies, incorporating 7,429 health facilities, 9,745 health workers, 41,856 febrile patients, and 15,398 malaria patients, and no study in low malaria transmission areas, were identified. The pooled proportion of the availability of malaria diagnostic tests, and the first-line AMDs in health facilities was 76% (95% CI 67-84); and 83% (95% CI 79-87), respectively. A pooled meta-analysis using random effects indicates the overall proportion of the correctness of malaria treatment 62% (95% CI 54-69). The appropriate malaria treatment was improved over time from 2009 to 2023. In the sub-group analysis, the correctness of treatment proportion was 53% (95% CI 50-63) for non-physicians health workers and 69% (95% CI 55-84) for physicians.

CONCLUSION

Findings of this review indicated that the correctness of malaria treatment and the availability of AMDs and diagnostic tests need improving to progress the malaria elimination stage.

Community case management of malaria in Western Kenya: performance of community health volunteers in active malaria case surveillance

BACKGROUND

In western Kenya, not all malaria cases are reported as stipulated in the community case management of malaria (CCMm) strategy. This underreporting affects the equity distribution of malaria commodities and the evaluation of interventions. The current study aimed to evaluate the effectiveness of community health volunteers' active case detection and management of malaria in western Kenya.

METHODS

Cross-sectional active case detection (ACD) of malaria survey was carried out between May and August 2021 in three eco-epidemiologically distinct zones in Kisumu, western Kenya: Kano Plains, Lowland lakeshore and Highland Plateau. The CHVs conducted biweekly ACD of malaria household visits to interview and examine residents for febrile illness. The Community Health Volunteers (CHVs) performance during the ACD of malaria was observed and interviews done using structured questionnaires.

RESULTS

Of the total 28,800 surveyed, 2597 (9%) had fever and associated malaria symptoms. Eco-epidemiological zones, gender, age group, axillary body temperature, bed net use, travel history, and survey month all had a significant association with malaria febrile illness (p < 0.05). The qualification of the CHV had a significant influence on the quality of their service. The number of health trainings received by the CHVs was significantly related to the correctness of using job aid (χ2 = 6.261, df = 1, p = 0.012) and safety procedures during the ACD activity (χ2 = 4.114, df = 1, p = 0.043). Male CHVs were more likely than female CHVs to correctly refer RDT-negative febrile residents to a health facility for further treatment (OR = 3.94, 95% CI = 1.85-5.44, p < 0.0001). Most of RDT-negative febrile residents who were correctly referred to the health facility came from the clusters with a CHV having 10 years of experience or more (OR = 1.29, 95% CI = 1.05-1.57, p = 0.016). Febrile residents in clusters managed by CHVs with more than 10 years of experience (OR = 1.82, 95% CI = 1.43-2.31, p < 0.0001), who had a secondary education (OR = 1.53, 95% CI = 1.27-1.85, p < 0.0001), and were over the age of 50 (OR = 1.44, 95% CI = 1.18-1.76, p < 0.0001), were more likely to seek malaria treatment in public hospitals. All RDT positive febrile residents were given anti-malarial by the CHVs, and RDT negatives were referred to the nearest health facility for further treatment.

CONCLUSIONS

The CHV's years of experience, education level, and age had a significant influence on their service quality. Understanding the qualifications of CHVs can assist healthcare systems and policymakers in designing effective interventions that assist CHVs in providing high-quality services to their communities.

A Systematic Review of Applications of Machine Learning and Other Soft Computing Techniques for the Diagnosis of Tropical Diseases

This systematic literature aims to identify soft computing techniques currently utilized in diagnosing tropical febrile diseases and explore the data characteristics and features used for diagnoses, algorithm accuracy, and the limitations of current studies. The goal of this study is therefore centralized around determining the extent to which soft computing techniques have positively impacted the quality of physician care and their effectiveness in tropical disease diagnosis. The study has used PRISMA guidelines to identify paper selection and inclusion/exclusion criteria. It was determined that the highest frequency of articles utilized ensemble techniques for classification, prediction, analysis, diagnosis, etc., over single machine learning techniques, followed by neural networks. The results identified dengue fever as the most studied disease, followed by malaria and tuberculosis. It was also revealed that accuracy was the most common metric utilized to evaluate the predictive capability of a classification mode. The information presented within these studies benefits frontline healthcare workers who could depend on soft computing techniques for accurate diagnoses of tropical diseases. Although our research shows an increasing interest in using machine learning techniques for diagnosing tropical diseases, there still needs to be more studies. Hence, recommendations and directions for future research are proposed.

Impact of Environmental Modifications on the Ecology, Epidemiology, and Pathogenesis of Plasmodium falciparum and Plasmodium vivax Malaria in East Africa

Food insecurity, recurrent famine, and poverty threaten the health of millions of African residents. Construction of dams and rural irrigation schemes is key to solving these problems. The sub-Saharan Africa International Center of Excellence for Malaria Research addresses major knowledge gaps and challenges in Plasmodium falciparum and Plasmodium vivax malaria control and elimination in malaria-endemic areas of Kenya and Ethiopia where major investments in water resource development are taking place. This article highlights progress of the International Center of Excellence for Malaria Research in malaria vector ecology and behavior, epidemiology, and pathogenesis since its inception in 2017. Studies conducted in four field sites in Kenya and Ethiopia show that dams and irrigation increased the abundance, stability, and productivity of larval habitats, resulting in increased malaria transmission and a greater disease burden. These field studies, together with hydrological and malaria transmission modeling, enhance the ability to predict the impact of water resource development projects on vector larval ecology and malaria risks, thereby facilitating the development of optimal water and environmental management practices in the context of malaria control efforts. Intersectoral collaborations and community engagement are crucial to develop and implement cost-effective malaria control strategies that meet food security needs while controlling malaria burden in local communities.

Influence of landscape heterogeneity on entomological and parasitological indices of malaria in Kisumu, Western Kenya

Background

Identification and characterization of larval habitats, documentation of Anopheles spp. composition and abundance, and Plasmodium spp. infection burden are critical components of integrated vector management. The present study aimed to investigate the effect of landscape heterogeneity on entomological and parasitological indices of malaria in western Kenya.

Methods

A cross-sectional entomological and parasitological survey was conducted along an altitudinal transect in three eco-epidemiological zones: lakeshore along the lakeside, hillside, and highland plateau during the wet and dry seasons in 2020 in Kisumu County, Kenya. Larval habitats for Anopheles mosquitoes were identified and characterized. Adult mosquitoes were sampled using pyrethrum spray catches (PSC). Finger prick blood samples were taken from residents and examined for malaria parasites by real-time PCR (RT-PCR).

Results

Increased risk of Plasmodium falciparum infection was associated with residency in the lakeshore zone, school-age children, rainy season, and no ITNs (χ² = 41.201, df = 9, P < 0.0001). Similarly, lakeshore zone and the rainy season significantly increased Anopheles spp. abundance. However, house structures such as wall type and whether the eave spaces were closed or open, as well as the use of ITNs, did not affect Anopheles spp. densities in the homes (χ² = 38.695, df = 7, P < 0.0001). Anopheles funestus (41.8%) and An. arabiensis (29.1%) were the most abundant vectors in all zones. Sporozoite prevalence was 5.6% and 3.2% in the two species respectively. The lakeshore zone had the highest sporozoite prevalence (4.4%, 7/160) and inoculation rates (135.2 infective bites/person/year). High larval densities were significantly associated with lakeshore zone and hillside zones, animal hoof prints and tire truck larval habitats, wetland and pasture land, and the wet season. The larval habitat types differed significantly across the landscape zones and seasonality (χ² = 1453.044, df = 298, P < 0.0001).

Conclusion

The empirical evidence on the impact of landscape heterogeneity and seasonality on vector densities, parasite transmission, and Plasmodium infections in humans emphasizes the importance of tailoring specific adaptive environmental management interventions to specific landscape attributes to have a significant impact on transmission reduction.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s13071-022-05447-9.

Asymptomatic and submicroscopic Plasmodium infections in an area before and during integrated vector control in Homa Bay, western Kenya

BACKGROUND

Long-lasting insecticidal nets (LLINs) have been the primary vector control strategy until indoor residual spraying (IRS) was added in Homa Bay and Migori Counties in western Kenya. The objective of this study was to evaluate the impact of LLINs integrated with IRS on the prevalence of asymptomatic and submicroscopic Plasmodium infections in Homa Bay County.

METHODS

A two-stage cluster sampling procedure was employed to enroll study participants aged ≥ 6 months old. Four consecutive community cross-sectional surveys for Plasmodium infection were conducted in residents of Homa Bay county, Kenya. Prior to the start of the study, all study households received LLINs, which were distributed between June 2017 and March 2018. The first (February 2018) and second (June 2018) surveys were conducted before and after the first round of IRS (Feb-Mar 2018), while the third (February 2019) and fourth (June 2019) surveys were conducted before and after the second application of IRS (February-March 2019). Finger-prick blood samples were obtained to prepare thick and thin smears for microscopic determination and qPCR diagnosis of Plasmodium genus.

RESULTS

Plasmodium spp. infection prevalence by microscopy was 18.5% (113/610) before IRS, 14.2% (105/737) and 3.3% (24/720) after the first round of IRS and 1.3% (11/849) after the second round of IRS (p < 0.0001). Submicroscopic (blood smear negative, qPCR positive) parasitaemia reduced from 18.9% (115/610) before IRS to 5.4% (46/849) after IRS (p < 0.0001). However, the proportion of PCR positive infections that were submicroscopic increased from 50.4% (115/228) to 80.7% (46/57) over the study period (p < 0.0001). Similarly, while the absolute number and proportions of microscopy positives which were asymptomatic decreased from 12% (73/610) to 1.2% (9/849) (p < 0.0001), the relative proportion increased. Geometric mean density of P. falciparum parasitaemia decreased over the 2-year study period (p < 0.0001).

CONCLUSIONS

These data suggest that two annual rounds of IRS integrated with LLINs significantly reduced the prevalence of Plasmodium parasitaemia, while the proportion of asymptomatic and submicroscopic infections increased. To reduce cryptic P. falciparum transmission and improve malaria control, strategies aimed at reducing the number of asymptomatic and submicroscopic infections should be considered.

Risk associations of submicroscopic malaria infection in lakeshore, plateau and highland areas of Kisumu County in western Kenya

BACKGROUND

Persons with submicroscopic malaria infection are a major reservoir of gametocytes that sustain malaria transmission in sub-Saharan Africa. Despite recent decreases in the national malaria burden in Kenya due to vector control interventions, malaria transmission continues to be high in western regions of the country bordering Lake Victoria. The objective of this study was to advance knowledge of the topographical, demographic and behavioral risk factors associated with submicroscopic malaria infection in the Lake Victoria basin in Kisumu County.

METHODS

Cross-sectional community surveys for malaria infection were undertaken in three eco-epidemiologically distinct zones in Nyakach sub-County, Kisumu. Adjacent regions were topologically characterized as lakeshore, hillside and highland plateau. Surveys were conducted during the 2019 and 2020 wet and dry seasons. Finger prick blood smears and dry blood spots (DBS) on filter paper were collected from 1,777 healthy volunteers for microscopic inspection and real time-PCR (RT-PCR) diagnosis of Plasmodium infection. Persons who were PCR positive but blood smear negative were considered to harbor submicroscopic infections. Topographical, demographic and behavioral risk factors were correlated with community prevalence of submicroscopic infections.

RESULTS

Out of a total of 1,777 blood samples collected, 14.2% (253/1,777) were diagnosed as submicroscopic infections. Blood smear microscopy and RT-PCR, respectively, detected 3.7% (66/1,777) and 18% (319/1,777) infections. Blood smears results were exclusively positive for P. falciparum, whereas RT-PCR also detected P. malariae and P. ovale mono- and co-infections. Submicroscopic infection prevalence was associated with topographical variation (χ2 = 39.344, df = 2, p<0.0001). The highest prevalence was observed in the lakeshore zone (20.6%, n = 622) followed by the hillside (13.6%, n = 595) and highland plateau zones (7.9%, n = 560). Infection prevalence varied significantly according to season (χ2 = 17.374, df = 3, p<0.0001). The highest prevalence was observed in residents of the lakeshore zone in the 2019 dry season (29.9%, n = 167) and 2020 and 2019 rainy seasons (21.5%, n = 144 and 18.1%, n = 155, respectively). In both the rainy and dry seasons the likelihood of submicroscopic infection was higher in the lakeshore (AOR: 2.71, 95% CI = 1.85-3.95; p<0.0001) and hillside (AOR: 1.74, 95% CI = 1.17-2.61, p = 0.007) than in the highland plateau zones. Residence in the lakeshore zone (p<0.0001), male sex (p = 0.025), school age (p = 0.002), and living in mud houses (p = 0.044) increased the risk of submicroscopic malaria infection. Bed net use (p = 0.112) and occupation (p = 0.116) were not associated with submicroscopic infection prevalence.

CONCLUSION

Topographic features of the local landscape and seasonality are major correlates of submicroscopic malaria infection in the Lake Victoria area of western Kenya. Diagnostic tests more sensitive than blood smear microscopy will allow for monitoring and targeting geographic sites where additional vector interventions are needed to reduce malaria transmission.