Background Public health agencies require valid, timely and complete health information for early detection of outbreaks. Towards the end of the Ebola Virus Disease (EVD) outbreak in 2015, the Ministry of Health and Sanitation (MoHS), Sierra Leone revitalized the Integrated Disease Surveillance and Response System (IDSR). Data quality assessments were conducted to monitor the accuracy of data generated through the IDSR system.Methods Starting 2016, regular data quality assessments (DQA)were conducted in randomly selected health facilities. A structured electronic checklist was used to interview district health management team (DHMT) members and health facility staff. We used malaria data to assess data accuracy as malaria was endemic in Sierra Leone. Verification factors (VF) calculated as the ratio of verified malaria cases in the health facility register to the number of malaria cases recorded in the national health information database, were used to assess data accuracy. Allowing a 5% margin of error, VF <95% were considered over reporting while a VF >105 was underreporting. Differences in the proportion of accurate reports in the first and fourth assessments were compared using Z-test for two proportions.Results Between 2016 -2018, four DQA were conducted in 444 health facilities where 1,729 IDSR reports were reviewed. Registers and IDSR technical guidelines were widely available in health facilities and health care workers were conversant with reporting requirements. Overall data accuracy improved from VF of 95.3% in 2016 to 100.2% in 2018. Compared to the baseline in 2016, the proportion of accurate IDSR reports in 2018 increased by 19.5% (CI 12.5% -26.5%). Over reporting was more common in private clinics and not for profit facilities while under-reporting was more common in lower level government health facilities. Leading reasons for data discrepancies included counting errors in 358 (80.6%) health facilities, and missing source documents in 47 (10.6%) health facilities.Conclusion This is the first attempt to institutionalize routine monitoring of IDSR data quality in Sierra Leone. Regular data quality assessments may have contributed to improved data accuracy over time. Data compilation errors accounted for most discrepancies and should be minimized to improve accuracy of IDSR data.

Abstract Background Public health agencies require valid, timely and complete health information for early detection of outbreaks. Towards the end of the Ebola Virus Disease (EVD) outbreak in 2015, the Ministry of Health and Sanitation (MoHS), Sierra Leone revitalized the Integrated Disease Surveillance and Response System (IDSR). Data quality assessments were conducted to monitor accuracy of IDSR data. Methods Starting 2016, data quality assessments (DQA) were conducted in randomly selected health facilities. Structured electronic checklist was used to interview district health management teams (DHMT) and health facility staff. We used malaria data, to assess data accuracy as malaria was endemic in Sierra Leone. Verification factors (VF) calculated as the ratio of verified malaria cases in health facility registers to the number of malaria cases in the national health information database, were used to assess data accuracy. Allowing a 5% margin of error, VF <95% were considered over reporting while VF >105 was underreporting. Differences in the proportion of accurate reports at baseline and subsequent assessments were compared using Z-test for two proportions. Results Between 2016 -2018, four DQA were conducted in 444 health facilities where 1,729 IDSR reports were reviewed. Registers and IDSR technical guidelines were available in health facilities and health care workers were conversant with reporting requirements. Overall data accuracy improved from over- reporting of 4.7% (VF 95.3%) in 2016 to under-reporting of 0.2% (VF 100.2%) in 2018. Compared to 2016, proportion of accurate IDSR reports increased by 14.8 % (95% CI 7.2%, 22.3%) in May 2017 and 19.5% (95% CI 12.5% -26.5%) by 2018. Over reporting was more common in private clinics and not- for profit facilities while under-reporting was more common in lower level government health facilities. Leading reasons for data discrepancies included counting errors in 358 (80.6%) health facilities and missing source documents in 47 (10.6%) health facilities. Conclusion This is the first attempt to institutionalize routine monitoring of IDSR data quality in Sierra Leone. Regular data quality assessments may have contributed to improved data accuracy over time. Data compilation errors accounted for most discrepancies and should be minimized to improve accuracy of IDSR data.

Background Public health agencies require valid, timely and complete health information for early detection of outbreaks. Towards the end of the Ebola Virus Disease (EVD) outbreak in 2015, the Ministry of Health and Sanitation (MoHS), Sierra Leone revitalized the Integrated Disease Surveillance and Response System (IDSR). Data quality assessments were conducted to monitor accuracy of IDSR data. Methods Starting 2016, data quality assessments (DQA) were conducted in randomly selected health facilities. Structured electronic checklist was used to interview district health management teams (DHMT) and health facility staff. We used malaria data, to assess data accuracy, as malaria was endemic in Sierra Leone. Verification factors (VF) calculated as the ratio of confirmed malaria cases recorded in health facility registers to the number of malaria cases in the national health information database, were used to assess data accuracy. Allowing a 5% margin of error, VF <95% were considered over reporting while VF >105 was underreporting. Differences in the proportion of accurate reports at baseline and subsequent assessments were compared using Z-test for two proportions. Results: Between 2016 -2018, four DQA were conducted in 444 health facilities where 1,729 IDSR reports were reviewed. Registers and IDSR technical guidelines were available in health facilities and health care workers were conversant with reporting requirements. Overall data accuracy improved from over- reporting of 4.7% (VF 95.3%) in 2016 to under-reporting of 0.2% (VF 100.2%) in 2018. Compared to 2016, proportion of accurate IDSR reports increased by 14.8 % (95% CI 7.2%, 22.3%) in May 2017 and 19.5% (95% CI 12.5% -26.5%) by 2018. Over reporting was more common in private clinics and not- for profit facilities while under-reporting was more common in lower level government health facilities. Leading reasons for data discrepancies included counting errors in 358 (80.6%) health facilities and missing source documents in 47 (10.6%) health facilities. Conclusion This is the first attempt to institutionalize routine monitoring of IDSR data quality in Sierra Leone. Regular data quality assessments may have contributed to improved data accuracy over time. Data compilation errors accounted for most discrepancies and should be minimized to improve accuracy of IDSR data.