Abstract
OBJECTIVESThe aim of effective data quality control and management is to minimize the impact of errors on study results by identifying and correcting them. This study presents the results of a data quality control system for the Korea HIV/AIDS Cohort Study that took into account the characteristics of the data.METHODSThe HIV/AIDS Cohort Study in Korea conducts repeated measurements every 6 months using an electronic survey administered to voluntarily consenting participants and collects data from 21 hospitals. In total, 5,795 sets of data from 1,442 participants were collected from the first investigation in 2006 to 2016. The data refining results of 2015 and 2019 were converted into the data refining rate and compared.RESULTSThe quality control system involved 3 steps at different points in the process, and each step contributed to data quality management and results. By improving data quality control in the pre-phase and the data collection phase, the estimated error value in 2019 was 1,803, reflecting a 53.9% reduction from 2015. Due to improvements in the stage after data collection, the data refining rate was 92.7% in 2019, a 24.21%p increase from 2015.CONCLUSIONSDespite this quality management strategy, errors may still exist at each stage. Logically possible errors for the post-review refining of downloaded data should be actively identified with appropriate consideration of the purpose and epidemiological characteristics of the study data. To improve data quality and reliability, data management strategies should be systematically implemented.
Highlights
Longitudinal data are obtained from repeated investigation of specific factors through a long-term follow-up of the same subject
Bias in cohort studies includes a volunteer bias caused by differences in characteristics between those who voluntarily agreed to participate in the study and those who did not, a follow-up loss bias resulted by death or dropout during the study, an ascertainment bias led by different investigation process of disease information between the exposed and non-exposed group, Hawthorne effect caused by changes in the subject’s behavior by repeated measurement of risk factors, and a time bias affected by changes in diagnostic criteria or subject’s personal factors depending on the follow-up period
The entire process of the Korea human immunodeficiency virus (HIV)/acquired immune deficiency syndrome (AIDS) cohort research data quality management strategy program is largely divided into prephase of data collection, phase of data collection, and post-phase of data collection
Summary
Longitudinal data are obtained from repeated investigation of specific factors through a long-term follow-up of the same subject. A type of longitudinal data, are an epidemiological study design that compare the incidence or mortality in two groups through long-term follow-ups of a population exposed to a specific risk factor and another that was not exposed which have the advantage that yields a clear temporal precedence relationship between cause and effect. Bias in cohort studies includes a volunteer bias caused by differences in characteristics between those who voluntarily agreed to participate in the study and those who did not, a follow-up loss bias resulted by death or dropout during the study, an ascertainment bias led by different investigation process of disease information between the exposed and non-exposed group, Hawthorne effect caused by changes in the subject’s behavior by repeated measurement of risk factors, and a time bias affected by changes in diagnostic criteria or subject’s personal factors depending on the follow-up period. It is necessary to perform appropriate quality management in the entire process from pre-phase of the data collection to the time when data are collected
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
