Background. Accurate data are essential for monitoring progress and course correction. Good quality facility-based routine data can be used at the facility, district, national, and global levels to improve quality of care and care policies. However, poor routine data quality has been an ongoing challenge. This thesis aimed to evaluate the quality of routine data for monitoring maternal and newborn care in primary health facilities in Gombe State, Nigeria. Methods. To examine the quality of routine monitoring data, in Study 1 we assessed facility-reported data in the District Health Information Software, version 2 (DHIS 2) according to three routine data quality dimensions: completeness and timeliness, internal consistency, and external consistency. Using direct observations as a gold standard, in Study 2 we assessed the validity of data in facility registers as well as women’s recall of childbirth events. For 21 months (April 2017-December 2018), we implemented a data quality intervention, working with all 11 local government area (district-equivalent) monitoring and evaluation officers and maternal and child health program coordinators of Gombe State which oversee 492 primary health facilities. The intervention included regular self-assessment of data quality, learning workshops, and planning for improvement. In Study 3, we quantified the changes in data quality using before-and-after analyses, comparing the intervention period to the 21-month pre-intervention period (July 2015-March 2017). Results. Twelve of 14 priority facility-based indicators were available in Gombe’s health information system to monitor maternal and newborn care. However, the facility data were incomplete and showed inconsistencies over time, between related indicators, between internal and external data sources. Contact indicators had higher data quality than indicators reflecting the content of care. Though there were challenges with the quality of facility-reported data, the validity study demonstrated that health workers were able to record valid information for some aspects of maternal and newborn care. When compared to childbirth observations, health workers documented accurately in maternity registers for the following indicators: the cadre of main birth attendant; maternal background characteristics, and newborn outcomes. Lastly, the data quality intervention was associated with improved completeness, timeliness, consistency between related data, and accuracy of facility reporting. Conclusion. Facility-based routine data in Gombe State can monitor priority service provision indicators for mothers and newborns. To realize the potential of these data, opportunities to improve data quality include: expanding data quality assessments beyond completeness and accuracy; maximizing the reporting and specificity of existing data; refining supervision feedback on the data quality metrics; and optimizing the digitization of facility data in information systems such as DHIS 2. Further research opportunities include: deepening our understanding of how health workers directly engage with facility documentation to perform clinical care tasks; and developing a composite score to summarize the multi-dimensionality of routine data as a measure for continuous data quality monitoring and as an outcome for data quality interventions.