Quality assurance of data: ensuring that numbers reflect operational definitions and contain real measurements

Any analysis is only as convincing as the quality of the underlying data. In this article, the role of data quality is exemplified by its impact on the interpretation of surveillance data, by operations research projects conducted in the training courses of the International Union Against Tuberculosis and Lung Disease, and the lessons learnt through them. It provides information why double-entry and validation of data are part of 'good clinical practice'. It is suggested how the efficiency of data entry can be maximized to reduce data entry time and data entry errors, so that psychological and physical barriers to double-entry are reduced.