Suitability of Electronic Health Record Data for Computational Phenotyping of Diabetes Mellitus at Nairobi Hospital, Nairobi City County, Kenya

Olwendo, Amos Otieno; Ochieng, George; Rucha, Kenneth

doi:10.37425/eajsti.v2i2.224

Cited by 2 publications

(6 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…All points that are within eps of a core point are termed density-reachable and are considered to be part of a cluster. The rest of the data points that are considered not density-reachable are all considered to be noise (Pekel and Özcan, 2018;Olwendo, Ochieng, and Rucha, 2021;Mehedi, Mollick, and Yasmin, 2022).…”

Section: Unsupervised Learning Modelsmentioning

confidence: 99%

“…As a matter of fact, Silhouette's analysis of the clusters of both the k-means and the fuzzy c-means reported that only about 30% had negative coefficients, meaning that a larger percentage of the dataset had been placed in the clusters they relatively belonged to. Therefore, a comparative analysis of the performances of machine learning methods needs to consider sampling from both the supervised and unsupervised algorithms and compare results within each of the two categories (Olwendo, Ochieng, and Rucha, 2020;Mehedi, Mollick, and Yasmin, 2022).…”

Section: Model Performance For the Fuzzy C-means Algorithmmentioning

confidence: 99%

“…In addition to diet, diabetes is best managed by maintaining an active lifestyle where one could be engaged, for example, in rigorous exercise or work. Diabetes is one of the diseases that require biomarker discovery and translation research to determine the clinical characteristics of their sub-phenotypes right from onset to the manifestation of its complications (Fritsch et al, 2021;Olwendo, Ochieng, and Rucha, 2021).…”

mentioning

confidence: 99%

“…Studies conducted in the United States of America, China, and the United Kingdom, among others, show a common trend in the prevalence of prediabetes between 36 and 50%. As a result, there is a need for improving diagnostic and therapeutic strategies for the effective management of diabetes (Olwendo, Ochieng, and Rucha, 2020).…”

mentioning

confidence: 99%

See 3 more Smart Citations

Comparison of machine learning methods for the prediction of type 2 diabetes in primary care setting using EHR data

Olwendo,

Ochieng,

Rucha

2023

jagst

View full text Add to dashboard Cite

Diabetes remains a major global public health challenge, thus the need for better methods for managing diabetes. Machine learning could provide reliable solutions to the need for early detection and management of diabetes. This study conducted experiments to compare a number of selected machine learning approaches to determine their suitability for early detection of diabetes in the primary care setting. A retrospective study was conducted using EHR dataset of confirmed cases of diabetes collected during routine care at Nairobi Hospital. Institutional ethical approvals were obtained, and data were retrieved from the database through stratified sampling based on gender. Diagnoses were confirmed using the ICD-10 codes. Records with 5% or so of missing values were excluded from this analysis. Data were processed by correction of errors and replacement of missing values using measures of central tendency. The data were transformed through normalization using the decimal-scaling method. Data analysis was conducted using selected supervised and unsupervised learning algorithms. Model performances were validated using metrics for the evaluation of classification and clustering results, respectively. Random Forest had the highest accuracy (0.95) and error rate (0.05), while Gradient Boosting and Multilayer Perceptron (MLP) with 3 hidden layers obtained accuracy (0.94) and error rate (0.06), respectively. The process of selecting machine learning algorithms needs to explore both supervised and unsupervised learning techniques. In addition, an appropriate architectural design of an MLP could present astounding results for classification tasks in primary care settings.

show abstract

Section: Unsupervised Learning Modelsmentioning

confidence: 99%

Section: Model Performance For the Fuzzy C-means Algorithmmentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Comparison of machine learning methods for the prediction of type 2 diabetes in primary care setting using EHR data

Olwendo,

Ochieng,

Rucha

2023

jagst

View full text Add to dashboard Cite

show abstract

“…Poor glycemic control among diabetes patients is associated with the occurrence of diabetes complications, estimated to be approximately 34.6% among adults in Kenya 8 . A retrospective analysis of hospital data collected between 2012 and 2016 showed retinopathy, neuropathy and cardiovascular diseases to be the most common diabetes complications among patients, with a prevalence of 12%, 11%, and 11%, respectively 9 , among which diabetes nephropathy is the leading cause of death 10 .…”

Section: Introductionmentioning

confidence: 99%

The Economic Burden of Type 2 Diabetes on the Public Healthcare System in Kenya: A Cost of Illness Study

Karugu,

Ilboudo,

Boachie

et al. 2023

Preprint

View full text Add to dashboard Cite

Background: The burden of chronic non-communicable diseases (NCDs) is a growing public health concern. The availability of cost-of-illness data, particularly public healthcare costs for NCDs, is limited in Sub-Saharan Africa (SSA), yet such data evidence is needed for policy action. Objective: The objective of this study was to estimate the economic burden of type 2 diabetes (T2D) on Kenya’s public healthcare system in the year 2021 and project costs for the year 2045. Methods: This was a cost-of-illness study using the prevalence-based bottom-up costing approach to estimate the economic burden of T2D in the year 2021. We further conducted projections on the estimated costs for the year 2045. The costs were estimated corresponding to the care, treatment, and management of diabetes and some diabetes complications based on the primary data collected from six healthcare facilities in Nairobi and secondary costing data from previous costing studies in low and middle-income countries (LMICs). The data capture and costing analysis were done in Microsoft Excel 16, and sensitivity analysis was conducted on all the parameters to estimate the cost changes. Results: The total cost of the management of T2D for the healthcare system in Kenya was estimated to be US$ 635 million (KES 74,521 million) in 2021. This was an increase of US$ 2 million (KES 197 million) considering the screening costs of undiagnosed T2D in the country. The major cost driver representing 59% of the overall costs was attributed to T2D complications, with nephropathy having the highest estimated costs of care and management (US$ 332 million (KES 36, 457 million). The total cost for T2D was projected to rise to US$ 1.6 billion (KES 177 billion) in 2045. Conclusion: This study shows that T2D imposes a huge burden on Kenya’s healthcare system. There is a need for government and societal action to develop and implement policies that prevent T2D, and appropriately plan care for those diagnosed with T2D.

show abstract

Suitability of Electronic Health Record Data for Computational Phenotyping of Diabetes Mellitus at Nairobi Hospital, Nairobi City County, Kenya

Cited by 2 publications

References 24 publications

Comparison of machine learning methods for the prediction of type 2 diabetes in primary care setting using EHR data

Comparison of machine learning methods for the prediction of type 2 diabetes in primary care setting using EHR data

The Economic Burden of Type 2 Diabetes on the Public Healthcare System in Kenya: A Cost of Illness Study

Contact Info

Product

Resources

About