ABSTRACT

Background: The problem of short children or stunting is one of the nutritional problem encountered in the world, especially in poor and developing countries. Stunting can provide long-term and short-term impacts. Early stunting detection requiring detailed assessment of various aspects needs to be packaged in a simple, accurate and inexpensive method. Until now no model that can help in the simple stunting screening. For this, a stunting models with Stunting prediction cards can be one of the methods that can be used as an assistive device to detect stunting. Aim: To develop and evaluate scoring models with stunting predictive cards

Method: This research is quantitative research with diagnostic test design and using consecutive sampling. The Total samples in this study were 179 case groups and 178 control groups. The Instrument used in this study is a stunting scoring card and a stunting cause questionnaire. Data analysis used in the processing of questionnaires is a Chi square analysis and a regression analysis of logistic with an accuracy rate of 95%. As for testing the scorers using fit modelling and ROC test with sensitivity and spasticity view. 

Result: Based on the results of logistic regression analysis, it is known that the most dominant factor in stunting is gender, birth weight, maternal knowledge, diet, and the history of maternal anemia. The result of the ROC test are known that the area under the curve is 0816, with 100% sensitivity and 98.9% specificity. This indicates that the scoring model in the predictive card is accurate and good in detecting stunting.

Conclusion: This study shows that there is a meaningful link between factors causing stunting to stunting events, and it can be known that the predictive card scoring model is accurate and can be used to detect stunting early.