• Gbenga A Kayode, Mary Amoakoh-Coleman, Charles Brown-Davies, Diederick E Grobbee, Irene Akua Agyepong, Evelyn Ansah, and Kerstin Klipstein-Grobusch.
• Julius Global Health, Julius Center for Health Sciences and Primary Care, University Medical Centre Utrecht, Utrecht, The Netherlands.
• Plos One. 2014 Jan 1; 9 (8): e104053.

ObjectivesThe District Health Information Management System-2 (DHIMS-2) is the database for storing health service data in Ghana, and similar to other low and middle income countries, paper-based data collection is being used by the Ghana Health Service. As the DHIMS-2 database has not been validated before this study aimed to evaluate its validity.MethodsSeven out of ten districts in the Greater Accra Region were randomly sampled; the district hospital and a polyclinic in each district were recruited for validation. Seven pre-specified neonatal health indicators were considered for validation: antenatal registrants, deliveries, total births, live birth, stillbirth, low birthweight, and neonatal death. Data were extracted on these health indicators from the primary data (hospital paper-registers) recorded from January to March 2012. We examined all the data captured during this period as these data have been uploaded to the DHIMS-2 database. The differences between the values of the health indicators obtained from the primary data and that of the facility and DHIMS-2 database were used to assess the accuracy of the database while its completeness was estimated by the percentage of missing data in the primary data.ResultsAbout 41,000 data were assessed and in almost all the districts, the error rates of the DHIMS-2 data were less than 2.1% while the percentages of missing data were below 2%. At the regional level, almost all the health indicators had an error rate below 1% while the overall error rate of the DHIMS-2 database was 0.68% (95% C I = 0.61-0.75) and the percentage of missing data was 3.1% (95% C I = 2.96-3.24).ConclusionThis study demonstrated that the percentage of missing data in the DHIMS-2 database was negligible while its accuracy was close to the acceptable range for high quality data.

11 keywords...

hide…