Managing missing values in routinely reported data: One approach from the Democratic Republic of the Congo
Data for Impact (D4I) held the second webinar in a series sharing lessons learned and practical suggestions on using routine and other existing data for evaluation and research. This one-hour webinar focused on managing missingness in routinely reported data and explored this through use of District Health Information System 2 (DHIS2) data from the Democratic Republic of the Congo (DRC). The webinar was led by Matt Worges, MS, PhD candidate at Tulane University and monitoring and evaluation specialist with D4I.
D4I is conducting an impact evaluation of health systems strengthening activities being implemented in the DRC by the USAID Integrated Health Project (IHP). One evaluation question seeks to determine the impact of the IHP on the utilization of health services over the course of the study period. To answer this question, D4I will use DHIS2 data in a difference-in-differences with propensity score matching model.
Use of DHIS2 data has several advantages, including access to a wide breadth of data elements that cover an array of health service areas, ability to be analyzed at different levels of the health system, and collection using standardized reporting tools at regular intervals. However, not all data elements are well-reported, and it is typically necessary to clean DHIS2 data before use. This webinar described the process the D4I team used to prepare time series data from the DRC DHIS2 in the lead up to the impact evaluation. Emphasis was placed on managing missing values within the DHIS2.
Matt Worges is a monitoring and evaluation specialist with Data for Impact and a PhD candidate in the School of Public Health and Tropical Medicine at Tulane University. His current research includes assessing the impact of a DRC-based health systems strengthening project designed to increase access to and use of health services.