Mobile-Based Early Detection of Stunting Using Risk Factors in Gunungkidul Regency, Special Region of Yogyakarta

Abstract

Background: In 2017, out of 11 countries in Southeast Asia, Indonesia ranked second for the country with the highest number of stunting events, 36.4%. The need for screening / early detection of stunting based on risk factors to prevent stunting in children. Purpose: To detect the risk of stunting based on mobile-based risk factors in Gunungkidul Regency, Yogyakarta. Methods: This research is an analytic survey with cross sectional approach. The sample was 98 mothers and their children aged 6-12 months with a quota sampling. In determining stunting risk categories based on risk factors, a data mining process is carried out using the C4.5 algorithm in predicting stunting in children with the results of categories in the form of risk and no risk of stunting. Data analysis was performed using univariate analysis, D2S application validity test (Deteksi Dini Stunting), and D2S application system feasibility test. Results: The results of the D2S application validity test found a sensitivity value of 20% (95% CI 5.67% -50.98%), specificity 97.73% (95% CI 92.09% - 99.37%) and an accuracy of 89.79%. Based on the results of the feasibility test application using the user acceptence questionnaire on 20 users it was found that as many as 19 people (95%) stated that the D2S application was very feasible and as many as 1 person (5%) stated it was fit to use. Conclusion: The D2S application is less sensitive in detecting the risk of stunting in children, but is very precise in detecting no risk of stunting in children.