This study evaluated the reporting of routine health delivery services and prototyped a computer database, district health information software (DHIS), in the health information systems in Tanzania and Mozambique using case study sites in Bagamoyo district and Gaza province in Tanzania and Mozambique, respectively, from March to August 2002. The purpose of the study was to assess and determine the quality and quantity of reporting of health delivery services; and the benefit, challenges, and conditions for implementing a computer database at the district level of the health information systems in both countries. The study falls under the framework of action research, where the phases of action research (diagnosis, action planning, action taking, evaluation, and specifying learning) were executed in sequence and cyclically. The diagnosis phase of the study was carried out through interviews, questionnaire, analysis of texts and documents, and participant observations. Appropriate actions were planned. In the action phase, software was developed in order to migrate data from existing computer databases to the DHIS; data were migrated from paper-based databases to the DHIS; and the DHIS was translated into Swahili while adopted in Tanzania. Evaluation of the actions taken was conducted through group discussions and workshops with health workers. To specify learning, reports of the results were prepared and presented to the Ministries of Health for review. The results were analysed using the DHIS, SPSS computer software, and using content analysis. The findings were presented quantitatively and qualitatively using tables, graphs, figures, photos, and elaboration. The findings indicated that the health data being reported were not sufficient to support informed decision-making and health planning. The causes of the low quality of the data identified include incomplete, inaccurate, and untimely reporting; lack of resources and office space; existence of legacy information systems; and the existence of parallel reporting systems in the health information systems. The findings also indicated the major challenges in introducing computer databases to be the participation of users and the existing computer database systems. The study demonstrated that the DHIS is suitable software for the health information systems, and that data locked in legacy information systems can be safely extracted and migrated to new information systems.