Automating indicator data reporting from health facility EMR to a national aggregate data system in Kenya: An Interoperability field-test using OpenMRS and DHIS2

James M Kariuki, Eric-Jan Manders, Janise Richards, Tom Oluoch, Davies Kimanga, Steve Wanyee, James O Kwach, Xenophon Santas

Abstract


Introduction: Developing countries are increasingly strengthening national health information systems (HIS) for evidence-based decision-making. However, the inability to report indicator data automatically from electronic medical record systems (EMR) hinders this process. Data are often printed and manually re-entered into aggregate reporting systems. This affects data completeness, accuracy, reporting timeliness, and burdens staff who support routine indicator reporting from patient-level data.  

Method: After conducting a feasibility test to exchange indicator data from Open Medical Records System (OpenMRS) to District Health Information System version 2 (DHIS2), we conducted a field test at a health facility in Kenya. We configured a field-test DHIS2 instance, similar to the Kenya Ministry of Health (MOH) DHIS2, to receive HIV care and treatment indicator data and the KenyaEMR, a customized version of OpenMRS, to generate and transmit the data from a health facility. After training facility staff how to send data using the module, we compared completeness, accuracy and timeliness of automated indicator reporting with facility monthly reports manually entered into MOH DHIS2.

Results: All 45 data values in the automated reporting process were 100% complete and accurate while in manual entry process, data completeness ranged from 66.7% to 100% and accuracy ranged from 33.3% to 95.5% for seven months (July 2013-January 2014). Manual tally and entry process required at least one person to perform each of the five reporting activities, generating data from EMR and manual entry required at least one person to perform each of the three reporting activities, while automated reporting process had one activity performed by one person. Manual tally and entry observed in October 2013 took 375 minutes. Average time to generate data and manually enter into DHIS2 was over half an hour (M=32.35 mins, SD=0.29) compared to less than a minute for automated submission (M=0.19 mins, SD=0.15).

Discussion and Conclusion: The results indicate that indicator data sent electronically from OpenMRS-based EMR at a health facility to DHIS2 improves data completeness, eliminates transcription errors and delays in reporting, and reduces the reporting burden on human resources. This increases availability of quality indicator data using available resources to facilitate monitoring service delivery and measuring progress towards set goals.


Full Text:

PDF


DOI: https://doi.org/10.5210/ojphi.v8i2.6722



Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org