Objective Indicator Selection:Data-Driven Approach for Transitional Data Collection Tools at NACSCOP

Abstract

Abstract: The transition of data collection tools at the National Center for Communicable Disease Prevention (NACSCOP) necessitates a systematic approach to select indicators unbiasedly. With two concurrent sets of tools, the need arises for an objective model leveraging historical data and facility characteristics to guide indicator selection. Our project aimed to develop a classification model for advising on indicator choice amid this transitional phase.The Challenge: During the tool transition phase at NACSCOP, the coexistence of two sets of data collection tools introduces complexity in indicator selection. Human bias and the absence of an objective methodology might lead to arbitrary indicator choices. There's a critical need for a data-driven approach that considers historical data, facility characteristics, and indicator fluctuations. Solution Approach: Our team embarked on developing a classification model that harnesses historical data values, facility characteristics, and fluctuations in indicators. Leveraging machine learning and data analysis techniques, this model objectively advises on the selection of the most appropriate indicator to use, regardless of biases or arbitrary decision-making.Outcome: The classification model developed provides a systematic and objective approach to recommend the indicator whether new or old for data collection and reporting. By analyzing historical trends, facility-specific attributes, and changes in indicators, this model assists decision-makers in choosing the most relevant indicator for accurate and consistent data reporting.Next Steps: Further refinement and validation of the model using real-time transitional data, collaboration with NACSCOP to integrate the model into their decision-making process, and expanding the model's capabilities to accommodate evolving indicators and data collection tools.Keywords: Data Collection Tools, Indicator Selection, Transitional Phase, Data-Driven Model, Machine Learning, Historical Data Analysis, Facility Characteristics, Objective Decision-making.