Objective A routine health information system (RHIS) enables decision making in the healthcare system. We aimed to analyse data quality at the district and regional level and explore factors and perceptions affecting the quality and use of routine data.Design This was a mixed-methods study. We used the WHO toolkit for analysing data quality and interviewed staff at the point of data generation and along with the flow of data. Data were analysed using the Performance of Routine Information System Management framework.Setting This study was performed in eight districts in four regions of Ethiopia. The study was nested within a 2-year programme of the Operational Research and Coaching for government Analysts.Participants We visited 45 health posts, 1 district hospital, 16 health centres and 8 district offices for analysis of routine RHIS data and interviewed 117 staff members for the qualitative assessment.Outcome measures We assessed availability of source documents, completeness, timeliness and accuracy of reporting of routine data, and explored data quality and use perceptions.Results There was variable quality of both indicator and data element. Data on maternal health and immunisation were of higher quality than data on child nutrition. Issues ranged from simple organisational factors, such as availability of register books, to intricate technical issues, like complexity of indicators and choice of denominators based on population estimates. Respondents showed knowledge of the reporting procedures, but also demonstrated limited skills, lack of supportive supervision and reporting to please the next level. We saw limited examples of the use of data by the staff who were responsible for data reporting.Conclusion We identified important organisational, technical, behavioural and process factors that need further attention to improve the quality and use of RHIS data in Ethiopia.
