Planning HIV case surveillance: A data quality assessment of paper and electronic medical records in Addis Ababa, Ethiopia

7^th International Conference on HIV/AIDS, STDs and STIs

March 18-19, 2019 | New York, USA

Wudinesh B Belihu

Organizationa Name, Country

Posters & Accepted Abstracts: J AIDS CLIN RES

Abstract :

Background: With an estimated 613,823 HIV-positive Ethiopians who have ever enrolled in HIV care, there is a need to better understand how well health services diagnose, link, and retain HIV-positive people in care. Utilization of existing health service data may be an effective way to monitor patient clinical outcomes and target resources at the population level. HIV Case Surveillance use these service data to generate information on patient clinical outcome. Hence, assessment of health service data was conducted to explore its utility for HIV case surveillance.

Methods: Using convenience sampling, 24 HIV treatment facilities in Addis Ababa, Ethiopia, was selected to participate. In facilities with an electronic medical record (EMR), we reviewed HIV patientsâ?? paper health records from patients attending the facility during October 1st to December 31st, 2014and extracted electronic health data from facilitiesâ?? EMR. In facilities with no EMR, paper records from patients attending the facility during January 1st to December 31st, 2014, were reviewed. A standard assessment tool quantified data completeness and validity as quality indicators. Data were analyzed using SAS 9.4 and Microsoft Excel.

Results: 1,500 paper health records were reviewed from 21(87.5%) health facilities with an EMR and 3(12.5%) with no EMR. Of 53 paper-based variables assessed, 16(30.2%) variables, including patient cell phone number, sex, and age were high quality (>90% completeness and validity), 24(45.3%) variables, including patient first name and address, were medium-high quality (51-89%), 10(18.9%) variables, including patient last name and year of HIV diagnosis, were medium-low quality (26-50%), and 3(5.7%) variables were of low quality (<25%). A total of 52,817 electronic records were available from 21 health facilitiesâ?? EMR of 82 electronic variables assessed, 15(18.3%) variables, including patient age, sex, and address were high quality, 26(31.7%) variables were medium-high quality, 12(14.6%) variables were medium-low quality, and 29(35.4%) variables, including patient date of birth, were low quality. Quality of paper and electronic data varied by health facility and service unit.

Conclusions: Existing paper and electronic patient health data in Ethiopia could allow for HIV case surveillance. Opportunities exist to improve data quality, particularly patient demographic data that could facilitate the identification of unique patients across health records.

Biography :

E-mail: wudineshbelete@yahoo.com