Work

Usability of Cholera detection system in Tigray

Year: 2026

Type: article

Abstract: Cholera remains one of the most significant public health challenges in low-income countries, particularly in regions affected by limited healthcare infrastructure, inadequate sanitation systems, and recurrent conflicts. In Tigray, Ethiopia, recurrent Cholera outbreaks have caused substantial morbidity and mortality, creating a need for innovative technological solutions that support early decision and timely intervention. This study develops and evaluates a mobile-based decision support tool designed to predict cholera risk from symptoms and risk factors, and outputs a risk category. The study employed qualitative and quantitative research involving healthcare professionals from six selected institutions to evaluate the system usability. The usability assessment was conducted using the System Usability Scale (SUS) along with additional measures of effectiveness, efficiency, learnability, user satisfaction, and reliability. Data were collected from anonymized patient records, questionnaires, direct observations, and system testing sessions, and subsequently analyzed using statistical methods (%, Mean ± SD, time, CI) to evaluate the overall usability performance of the developed system. The developed Cholera Detection System (CDS) is an android-based triage tool with a Tigrigna and English language interface designed for usability evaluations by professionals. The prediction engine employs a Random Forest machine learning model trained on a balanced dataset of 3,250 Cholera-related records, comprising 1,625 Cholera-positive and 1,625 Cholera-negative cases. The data were collected from local healthcare facilities in the Tigray region using standardized data attributes derived from the World Health Organization and Ethiopian Public Health Institute Cholera reporting guidelines. The dataset was divided into 2,600 training records and 650 testing records, and model validation was performed using 10-fold cross-validation. Among the evaluated machine learning algorithms, Random Forest achieved the highest classification performance with an accuracy of 95.8%, demonstrating robustness against local healthcare data, reduced overfitting, and reliable for Cholera risk prediction. The usability evaluation results showed that the CDS achieved an overall SUS score of 84.5, which is categorized as excellent usability, and significantly exceeds the accepted benchmark value of 68. The findings indicate that healthcare professionals perceived the triage tool as easy to learn, efficient, reliable, and effective for decision support activities. The tool also improved consistency and reduced the time required for predicting cholera risks. The findings of this study demonstrate that the proposed CDS is a usable and effective tool for predicting cholera risk based on patients’ presenting symptoms, particularly in resource-limited and war-affected areas. Therefore, the proposed system is intended to support early screening and clinical triage. It does not provide a definitive diagnosis of cholera, which requires laboratory confirmation.

Source: BMC Health Services Research

DOI: 10.1186/s12913-026-15389-y

Authors: Dawit Teklu Weldeslasie, Shishay Teweldebirhan Tekle, Berhane Tsegay Assefa

Institution: Aksum University

Language: English

Topic: Artificial Intelligence in Healthcare

Subfield: Health Information Management

Field: Health Professions

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $3,290