Assessing the Quality of Immunization Data from Administrative Data in Enugu State, South-East Nigeria: A Cross-Sectional Study

Abstract

Background: Studies evaluating immunization data quality and its determinants in low- and middle-income countries are scarce. Aim: Therefore, this  study assessed the accuracy of routine immunization data and its associated factors in Enugu State, South-East Nigeria.

Materials and Methods: This was  a descriptive cross-sectional, mixed-method study. A multi-stage sampling technique was used to select 60 out of 180 primary health facilities in six local  government areas (LGAs). Routine data on Bacillus Calmette Guerin (BCG), pentavalent vaccine dose3 (Penta3), and Measles2 vaccinations in tallied  registers, facility reports, and District Health Information System II (DHIS2) in 2020 were abstracted using a questionnaire. The dependent variables were  the accuracy ratios of the facility and DHIS2 reports. We collected data on the independent variables including LGAs, type of location, facility type, tool  availability, tool use, frequency of supervision, defaulter tracing, and vaccine tracking. Additionally, semi-structured, in-depth interviews were conducted  with 35 purposively selected actors on factors affecting the data quality. The proportions of facilities with normal, under-, or over-reporting were  summarized in percentages. The median accuracy ratios and interquartile range (IQR) were also reported. Differences in median accuracy ratios were tested using the independent sample median test. We tested the predictive model using a generalized linear model. Statistical significance was set at P <  0.05. The qualitative data were analyzed using content analysis.

Results: The accuracy of facility reports ranged from 38% to 68% (normal), 17–30% (under-reporting), and 15–35% (over-reporting) of health facilities. In  DHIS2, the accuracy ranged from 5% to 10% (normal), 37–42% (under-reporting), and 53–63% (over-reporting) of health facilities. The median (IQR)  accuracy ratios of facility reports were 100.0% (98.3–103.2%), 100.0% (98.5–103.6%), and 100.0% (81.6–110.2%) for BCG, Penta3, and Measles2,  respectively. The median (IQR) accuracy ratios of DHIS2 reports were 83.6% (43.3–192.7%), 88.4% (37.8–200.8%), and 46.2% (10.7–202.7%) for BCG,  Penta3, and Measles2, correspondingly. No facility characteristic predicted the accuracy of routine immunization reports. Reasons for inaccurate data  include untimely recording, wrong counting, delayed reporting, infrequent supervision, lack of data audits, resource constraints, and high workload.   

Conclusions: Routine immunization data are not always accurate in Enugu state. Continuous efforts to improve the data monitoring system, supervision,  data audits, funding, and staffing are warranted.