Prediction of cardiovascular risk factors from retinal fundus photographs: Validation of a deep learning algorithm in a prospective non-interventional study in Kenya.

AIM: Hypertension and diabetes mellitus (DM) are major causes of morbidity and mortality, with growing burdens in low-income countries where they are underdiagnosed and undertreated. Advances in machine learning may provide opportunities to enhance diagnostics in settings with limited medical infrastructure. MATERIALS AND METHODS: A non-interventional study was conducted to develop and validate a machine learning algorithm to estimate cardiovascular clinical and laboratory parameters. At two sites in Kenya, digital retinal fundus photographs were collected alongside blood pressure (BP), laboratory measures and medical history. The performance of machine learning models, originally trained using data from the UK Biobank, were evaluated for their ability to estimate BP, glycated haemoglobin, estimated glomerular filtration rate and diagnoses from fundus images. RESULTS: In total, 301 participants were enrolled. Compared with the UK Biobank population used for algorithm development, participants from Kenya were younger and would probably report Black/African ethnicity, with a higher body mass index and prevalence of DM and hypertension. The mean absolute error was comparable or slightly greater for systolic BP, diastolic BP, glycated haemoglobin and estimated glomerular filtration rate. The model trained to identify DM had an area under the receiver operating curve of 0.762 (0.818 in the UK Biobank) and the hypertension model had an area under the receiver operating curve of 0.765 (0.738 in the UK Biobank). CONCLUSIONS: In a Kenyan population, machine learning models estimated cardiovascular parameters with comparable or slightly lower accuracy than in the population where they were trained, suggesting model recalibration may be appropriate. This study represents an incremental step toward leveraging machine learning to make early cardiovascular screening more accessible, particularly in resource-limited settings.

Optimal control analysis of a transmission interruption model for the soil-transmitted helminth infections in Kenya.

Kenya is among the countries endemic for soil-transmitted helminthiasis (STH) with over 66 subcounties and over 6 million individuals being at-risk of infection. Currently, the country is implementing mass drug administration (MDA) to all the at-risk groups as the mainstay control strategy. This study aimed to develop and analyze an optimal control (OC) model, from a transmission interruption model, to obtain an optimal control strategy from a mix of three strategies evaluated. The study used the Pontryagin's maximum principle to solve, numerically, the OC model. The analysis results clearly demonstrated that water and sanitation when implemented together with the MDA programme offer the best chances of eliminating these tenacious and damaging parasites. Thus, we advocate for optimal implementation of the combined mix of the two interventions in order to achieve STH elimination in Kenya, and globally, in a short implementation period of less than eight years.

Factors associated with viral load suppression among orphans and vulnerable children and adolescents living with HIV in Kenya.

While tremendous progress has been made on attaining HIV treatment goals (95-95-95), children's viral load suppression remains a challenge particularly among the orphans and vulnerable groups. In Sub Saharan Africa, there is limited evidence of specific interventions in orphans and vulnerable children (OVC) programs to support children and adolescents living with HIV (CALHIV) to attain durable viral load suppression. Through a large OVC cohort, the study sought to identify correlates of optimal viral load suppression among CALHIV in the Kenya OVC program. This cross-sectional study utilized data on CALHIV below the age of 18 years who were enrolled in the OVC program and actively receiving HIV care and treatment services from ART clinics across Kenya and with documented VL results between October 2019 and September 2020. To obtain a nationally representative sample, data was retrieved from USAID implementing partners' databases across the country. Association between selected variables and VL suppression (outcome of interest) were assessed using a multivariate mixed effect logistic regression model, using glmer function in the LME4 package in R. Factors associated with VL suppression included child's education status (aOR = 1.33; 95% CI: 1.07, 1.65), membership of a psychosocial support group (aOR = 1.258; 95% CI: 1.15, 1.38), and membership of a voluntary savings and lending association (VSLA) (aOR = 1.226; 95% CI: 1.129, 1.33). In addition, child's sex (aOR = 0.88; 95% CI: 0.83, 0.94), caregiver sex (aOR = 0.909; 95% CI: 0.839, 0.997) and "high" status for caregiver household vulnerability (aOR = 0.81; 95% CI: 0.71, 0.924), had an inverse relationship with VL suppression. CALHIV characteristics including child's sex, child's education status (whether currently active in school or inactive) and child's membership in a psychosocial support group were key determinants of VL suppression. Similarly, caregiver sex and membership in a voluntary savings and lending association also influence VL suppression.

Primary rectal melanoma in an African female: a case report.

Primary malignant melanoma occurs frequently on the skin and is rare in people of African ancestry. The rectal region is an unusual site for non-cutaneous melanoma. We report a case of a 58-year- old African woman presenting to a Kenyan hospital with lower abdominal pain and per rectal bleeding for three months, who underwent a colonoscopy that showed a rectal polypoid mass at the anorectal region. Histology of the mass showed pigmented pleomorphic cells which had positive stains for melanoma markers. Staging workup performed, including magnetic resonance imaging (MRI) of the pelvis and positron emission tomography (PET) - computed tomography (CT), showed regional lymph node involvement but no evidence of distant metastases. Surgery was recommended to the patient but she died eight months after the diagnosis. The case illustrates that primary rectal melanoma, though rare in Africans, is an aggressive disease which can be easily misdiagnosed as hemorrhoids or rectal adenocarcinoma.

Sensitivity Analysis of a Transmission Interruption Model for the Soil-Transmitted Helminth Infections in Kenya.

As the world rallies toward the endgame of soil-transmitted helminths (STH) elimination by the year 2030, there is a need for efficient and robust mathematical models that would enable STH programme managers to target the scarce resources and interventions, increase treatment coverage among specific sub-groups of the population, and develop reliable surveillance systems that meet sensitivity and specificity requirements for the endgame of STH elimination. However, the considerable complexities often associated with STH-transmission models underpin the need for specifying a large number of parameters and inputs, which are often available with considerable degree of uncertainty. Additionally, the model may behave counter-intuitive especially when there are non-linearities in multiple input-output relationships. In this study, we performed a global sensitivity analysis (GSA), based on a variance decomposition method: extended Fourier Amplitude Sensitivity Test (eFAST), to a recently developed STH-transmission model in Kenya (an STH endemic country) to; (1) robustly compute sensitivity index (SI) for each parameter, (2) rank the parameters in order of their importance (from most to least influential), and (3) quantify the influence of each parameter, singly and cumulatively, on the model output. The sensitivity analysis (SA) results demonstrated that the model outcome (STH worm burden elimination in the human host) was significantly sensitive to some key parameter groupings: combined effect of improved water source and sanitation (ϕ), rounds of treatment offered (τ), efficacy of the drug used during treatment (h), proportion of the adult population treated (ga : akin to community-wide treatment), mortality rate of the mature worms in the human host (µ), and the strength of the -dependence of worm egg production (γ). For STH control programmes to effectively reach the endgame (STH elimination in the entire community), these key parameter groupings need to be targeted since together they contribute to a strategic public health intervention.

Determinants of contraceptive use among women 0-23 months postpartum in Kitui County, Kenya: A cross-sectional study.

The risk of unintended pregnancy is high in the postpartum period, especially during the first year of delivery. Yet, short birth intervals are associated with increased risk of adverse maternal and infant outcomes. In Kenya, despite women having multiple contacts with healthcare providers during their pregnancy and postpartum journeys, uptake of contraceptives during the postpartum period remains low. We examine factors that determine contraceptive use among postpartum women in Kitui County, Kenya.A cross-sectional study was conducted in six sub-counties of Kitui County covering a random sample of 768 postpartum women in April 2019. Logistic regression was used to study the association between uptake of contraceptives among women 0-23 months postpartum and several explanatory variables that included socio-demographic characteristics and facility-level factors. Overall, 68% of women in Kitui County reported using contraceptives. The likelihood of contraceptive use increased with the increase in the number of known family planning methods. Women who discussed family planning with a health worker within the last 12 months were 2.58 (95%CI: 1.73, 3.89) times more likely to use contraceptives during the postpartum period compared to those who did not. There was an increased odds of contraceptive uptake among women who received family planning information or service during postnatal care than those who did not (aOR = 2.04, 95%CI: 1.30, 3.24). A positive association was also found between contraceptive use and receipt of family planning information or service during immunization visits or during child well visits. It is evident that facility-level factors such as discussing family planning with women; educating women about different family planning methods; providing family planning information or services during postnatal care, immunization, or well child visits are associated with increased likelihood of contraceptive uptake by women during postpartum period. Programs targeting enhancing women's attendance of postnatal care clinics should be encouraged.

Prevalence and Correlation Analysis of Soil-Transmitted Helminths Infections and Treatment Coverage for Preschool and School Aged Children in Kenya: Secondary Analysis of the National School Based Deworming Program Data.

Background: Soil-transmitted helminths (STH) are among the most common parasitic infections globally, disproportionately affecting children. Treatment of STH in Kenya is often targeted at preschool (PSAC) and school aged (SAC) children delivered through annual mass drug administration (MDA) in primary schools. Understanding group-specific prevalence and dynamics between treatment and coverage is critical for continued treatment success. This study aims to provide detailed information on group-specific infection prevalence and relative reductions (RR), and their relationships with treatment coverage over time. Additionally, it aims to quantify the correlation between the observed school level infection prevalence and treatment coverage. Methods: Secondary analysis of existing data collected between 2012 and 2018 by the monitoring and evaluation (M&E) program of the National School-Based Deworming (NSBD) program was used. The M&E program conducted surveys utilizing cross-sectional study design, at four survey time points, in a nationally-representative sample of schoolchildren across counties in Kenya. In each participating school, the program randomly sampled 108 children per school, of both groups. Infection prevalence was estimated using binomial regression, RR in prevalence using multivariable mixed effects model, statistical correlations using structural equation modeling, and change-point-analysis using the binary segmentation algorithm. Results: Overall, STH prevalence for PSAC was 33.7, 20.2, 19.0, and 17.9% during Year 1 (Y1), Year 3 (Y3), Year 5 (Y5), and Year 6 (Y6) surveys, respectively with an overall RR of 46.9% (p = 0.001) from Y1 to Y6. Similarly, overall STH prevalence for SAC was 33.6, 18.4, 14.7, and 12.5% during Y1, Y3, Y5, and Y6 surveys, respectively with an overall RR of 62.6% (p < 0.001). An overall (all time points) significant but very weak negative correlation was found between treatment coverage and undifferentiated STH prevalence (r = -0.144, p = 0.002) among PSAC but not in SAC. Further, we observed inter-county heterogeneity variation in infection prevalence, RR, as well as correlations. Conclusion: The analysis showed that after six rounds of MDA, prevalence of STH has significantly declined among both groups of children, however not to a point where it is not a public health problem (below 1%). The analysis, additionally established an overall significant but weak negative correlation between treatment coverage and prevalence, indicating that the current treatment coverage might not be sufficient to drive the overall STH prevalence to below 1%. These findings will allow STH control programs in Kenya to make decisions that will accelerate the attainment of STH elimination as a public health problem.

Bayesian Spatial Modeling of Anemia among Children under 5 Years in Guinea.

Anemia is a major public health problem in Africa, affecting an increasing number of children under five years. Guinea is one of the most affected countries. In 2018, the prevalence rate in Guinea was 75% for children under five years. This study sought to identify the factors associated with anemia and to map spatial variation of anemia across the eight (8) regions in Guinea for children under five years, which can provide guidance for control programs for the reduction of the disease. Data from the Guinea Multiple Indicator Cluster Survey (MICS5) 2016 was used for this study. A total of 2609 children under five years who had full covariate information were used in the analysis. Spatial binomial logistic regression methodology was undertaken via Bayesian estimation based on Markov chain Monte Carlo (MCMC) using WinBUGS software version 1.4. The findings in this study revealed that 77% of children under five years in Guinea had anemia, and the prevalences in the regions ranged from 70.32% (Conakry) to 83.60% (NZerekore) across the country. After adjusting for non-spatial and spatial random effects in the model, older children (48-59 months) (OR: 0.47, CI [0.29 0.70]) were less likely to be anemic compared to those who are younger (0-11 months). Children whose mothers had completed secondary school or above had a 33% reduced risk of anemia (OR: 0.67, CI [0.49 0.90]), and children from household heads from the Kissi ethnic group are less likely to have anemia than their counterparts whose leaders are from Soussou (OR: 0.48, CI [0.23 0.92]).

Modeling the Interruption of the Transmission of Soil-Transmitted Helminths Infections in Kenya: Modeling Deworming, Water, and Sanitation Impacts.

Kenya, just like other countries with endemic soil-transmitted helminths (STH), has conducted regular mass drug administration (MDA) program for the last 5 years among school aged children as a way to reduce STH infections burden in the country. However, the point of interruption of transmission of these infections still remains unclear. In this study, we developed and analyzed an age structured mathematical model to predict the elimination period (i.e., time taken to interrupt STH transmission) of these infections in Kenya. The study utilized a deterministic age structured model of the STH population dynamics under a regular treatment program. The model was applied to three main age groups: pre-school age children (2-4 years), school age children (5-14 years), and adult populations (≥15 years) and compared the impact of two interventions on worm burden and elimination period. The model-simulated results were compared with the 5 year field data from the Kenyan deworming program for all the three types of STH (Ascaris lumbricoides, Trichuris trichiura, and hookworm). The model demonstrated that the reduction of worm burden and elimination period depended heavily on four parameter groups; drug efficacy, number of treatment rounds, MDA and water, sanitation and hygiene (WASH) coverage. The analysis showed that for STH infections to be eliminated using MDA alone in a short time period, 3-monthly MDA plan is desired. However, complementation of MDA with WASH at an optimal (95%) coverage level was most effective. These results are important to the Kenyan STH control program as it will guide the recently launched Breaking Transmission Strategy.

Statistical Regression Model of Water, Sanitation, and Hygiene; Treatment Coverage; and Environmental Influences on School-Level Soil-Transmitted Helminths and Schistosome Prevalence in Kenya: Secondary Analysis of the National Deworming Program Data.

According to the Kenya National School-Based Deworming program launched in 2012 and implemented for the first 5 years (2012-2017), the prevalence of soil-transmitted helminths (STH) and schistosomiasis substantially reduced over the mentioned period among the surveyed schools. However, this reduction is heterogeneous. In this study, we aimed to determine the factors associated with the 5-year school-level infection prevalence and relative reduction (RR) in prevalence in Kenya following the implementation of the program. Multiple variables related to treatment, water, sanitation, and hygiene (WASH) and environmental factors were assembled and included in mixed-effects linear regression models to identify key determinants of the school location STH and schistosomiasis prevalence and RR. Reduced prevalence of Ascaris lumbricoides was associated with low (< 1%) baseline prevalence, seven rounds of treatment, high (50-75%) self-reported coverage of household handwashing facility equipped with water and soap, high (20-25°C) land surface temperature, and community population density of 5-10 people per 100 m2. Reduced hookworm prevalence was associated with low (< 1%) baseline prevalence and the presence of a school feeding program. Reduced Trichuris trichiura prevalence was associated with low (< 1%) baseline prevalence. Reduced Schistosoma mansoni prevalence was associated with low (< 1%) baseline prevalence, three treatment rounds, and high (> 75%) reported coverage of a household improved water source. Reduced Schistosoma haematobium was associated with high aridity index. Analysis indicated that a combination of factors, including the number of treatment rounds, multiple related program interventions, community- and school-level WASH, and several environmental factors had a major influence on the school-level infection transmission and reduction.

A Spatial Survival Model for Risk Factors of Under-Five Child Mortality in Kenya.

Child mortality is high in Sub-Saharan Africa compared to other regions in the world. In Kenya, the risk of mortality is assumed to vary from county to county due to diversity in socio-economic and even climatic factors. Recently, the country was split into 47 different administrative regions called counties, and health care was delegated to those county governments, further aggravating the spatial differences in health care from county to county. The goal of this study is to evaluate the effects of spatial variation in under-five mortality in Kenya. Data from the Kenya Demographic Health Survey (KDHS-2014) consisting the newly introduced counties was used to analyze this risk. Using a spatial Cox Proportional Hazard model, an Intrinsic Conditional Autoregressive Model (ICAR) was fitted to account for the spatial variation among the counties in the country while the Cox model was used to model the risk factors associated with the time to death of a child. Inference regarding the risk factors and the spatial variation was made in a Bayesian setup based on the Markov Chain Monte Carlo (MCMC) technique to provide posterior estimates. The paper indicate the spatial disparities that exist in the country regarding child mortality in Kenya. The specific counties have mortality rates that are county-specific, although neighboring counties have similar hazards for death of a child. Counties in the central Kenya region were shown to have the highest hazard of death, while those from the western region had the lowest hazard of death. Demographic factors such as the sex of the child and sex of the household head, as well as social economic factors, such as the level of education, accounted for the most variation when spatial differences were factored in. The spatial Cox proportional hazard frailty model performed better compared to the non-spatial non-frailty model. These findings can help the country to plan health care interventions at a subnational level and guide social and health policies by ensuring that counties with a higher risk of Under Five Child Mortality (U5CM) are considered differently from counties experiencing a lower risk of death.

Ethnic and Gender Diversity Comparison between Surgical Patients and Caring Surgeons.

BACKGROUND: Diversity within the medical workforce remains a topic of discussion in academia, particularly when it comes to the underrepresentation of certain ethnic groups and gender in the surgical specialties. In this article, we look at how the gender and ethnicity of surgeons at a large academic institution in a rural setting compare with those of the population it serves. METHODS: We looked at demographic data from 2008 to 2018 and compared population trends among surgeons and patients. RESULTS: We found that while whites represent the large majority in both the surgeon and patient populations, absolute number and percentage of whites in the patient population seem to be trending downward from 2008 to 2018, but trending upward among surgeons (attendings and residents). In addition, we found that while Asians make up only 1% of the patient population, they represent the second largest group (17%) among surgeons, with more than twice the proportion percentage of the second largest group within the patient population, composed of Hispanics (6%). Finally, we found a significant gender difference between the 2 populations with almost two-thirds of the surgeons being men, compared with the nearly even split of men and women within the patient population. CONCLUSIONS: Ultimately, understanding how gender and ethnic diversity in the surgical workforce compares with that of the patient population being served may aid in designing training programs to address cultural competency and awareness as well as in impacting administrative decisions and hiring.

Prognostic models for predicting in-hospital paediatric mortality in resource-limited countries: a systematic review.

OBJECTIVES: To identify and appraise the methodological rigour of multivariable prognostic models predicting in-hospital paediatric mortality in low-income and middle-income countries (LMICs). DESIGN: Systematic review of peer-reviewed journals. DATA SOURCES: MEDLINE, CINAHL, Google Scholar and Web of Science electronic databases since inception to August 2019. ELIGIBILITY CRITERIA: We included model development studies predicting in-hospital paediatric mortality in LMIC. DATA EXTRACTION AND SYNTHESIS: This systematic review followed the Checklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies framework. The risk of bias assessment was conducted using Prediction model Risk of Bias Assessment Tool (PROBAST). No quantitative summary was conducted due to substantial heterogeneity that was observed after assessing the studies included. RESULTS: Our search strategy identified a total of 4054 unique articles. Among these, 3545 articles were excluded after review of titles and abstracts as they covered non-relevant topics. Full texts of 509 articles were screened for eligibility, of which 15 studies reporting 21 models met the eligibility criteria. Based on the PROBAST tool, risk of bias was assessed in four domains; participant, predictors, outcome and analyses. The domain of statistical analyses was the main area of concern where none of the included models was judged to be of low risk of bias. CONCLUSION: This review identified 21 models predicting in-hospital paediatric mortality in LMIC. However, most reports characterising these models are of poor quality when judged against recent reporting standards due to a high risk of bias. Future studies should adhere to standardised methodological criteria and progress from identifying new risk scores to validating or adapting existing scores. PROSPERO REGISTRATION NUMBER: CRD42018088599.

Developing excellence in biostatistics leadership, training and science in Africa: How the Sub-Saharan Africa Consortium for Advanced Biostatistics (SSACAB) training unites expertise to deliver excellence.

The increase in health research in sub-Saharan Africa (SSA) has generated large amounts of data and led to a high demand for biostatisticians to analyse these data locally and quickly.  Donor-funded initiatives exist to address the dearth in statistical capacity, but few initiatives have been led by African institutions. The Sub-Saharan African Consortium for Advanced Biostatistics (SSACAB) aims to improve biostatistical capacity in Africa according to the needs identified by African institutions, through (collaborative) masters and doctoral training in biostatistics. We describe the SSACAB Consortium, which comprises 11 universities and four research institutions- supported by four European universities. SSACAB builds on existing resources to strengthen biostatistics for health research with a focus on supporting biostatisticians to become research leaders; building a critical mass of biostatisticians, and networking institutions and biostatisticians across SSA.  In 2015 only four institutions had established Masters programmes in biostatistics and SSACAB supported the remaining institutions to develop Masters programmes. In 2019 the University of the Witwatersrand became the first African institution to gain Royal Statistical Society accreditation for a Biostatistics MSc programme. A total of 150 fellows have been awarded scholarships to date of which 123 are Masters fellowships (41 female) of which with 58 have already graduated. Graduates have been employed in African academic (19) and research (15) institutions and 10 have enrolled for PhD studies. A total of 27 (10 female) PhD fellowships have been awarded; 4 of them are due to graduate by 2020. To date, SSACAB Masters and PhD students have published 17 and 31 peer-reviewed articles, respectively. SSACAB has also facilitated well-attended conferences, face-to-face and online short courses. Pooling the limited biostatistics resources in SSA, and combining with co-funding from external partners is an effective strategy for the development and teaching of advanced biostatistics methods, supervision and mentoring of PhD candidates.

Determination of optimal vaccination strategies using an orbital stability threshold from periodically driven systems.

We analyse a periodically driven SIR epidemic model for childhood related diseases, where the contact rate and vaccination rate parameters are considered periodic. The aim is to define optimal vaccination strategies for control of childhood related infections. Stability analysis of the uninfected solution is the tool for setting up the control function. The optimal solutions are sought within a set of susceptible population profiles. Our analysis reveals that periodic vaccination strategy hardly contributes to the stability of the uninfected solution if the human residence time (life span) is much larger than the contact rate period. However, if the human residence time and the contact rate periods match, we observe some positive effect of periodic vaccination. Such a vaccination strategy would be useful in the developing world, where human life spans are shorter, or basically in the case of vaccination of livestock or small animals whose life-spans are relatively shorter.