Authors:
Juliano Gaspar
1
;
Fernando Lopes
2
and
Alberto Freitas
3
Affiliations:
1
CIDES – Department of Health Information and Decision Sciences, Portugal
;
2
CINTESIS – Center for Research in Health Technologies and Information Systems, Portugal
;
3
University of Porto, Portugal
Keyword(s):
Inconsistencies, Errors, Data quality problems, Hospital databases.
Related
Ontology
Subjects/Areas/Topics:
Biomedical Engineering
;
Databases and Datawarehousing
;
Health Information Systems
;
Software Systems in Medicine
Abstract:
Introduction: Health professionals need data, in sufficient quantity and quality, and tools that can manage the vast amount of available data. They need help for data management and appropriate support for decision making.
Introduction: Health professionals need data, in sufficient quantity and quality, and tools that can manage the vast amount of available data. They need help for data management and appropriate support for decision making.
Aim: The focus of this study is to develop a prototype that can contribute to the identification of data quality problems in clinical and administrative data.
Methods: Methods involve the definition of requisites and business rules, the prototype development and testing, and the realization of two studies using the prototype.
Results: Studies performed using the prototype resulted in the detection of many data problems and inconsistencies. Amongst those we can point out, for instance, that 82,000 (15%) episodes had ‘diagnostic code does not exi
st in ICD-9-CM table’ and that 783 (0,2%) episodes within ‘female breast cancer’ had the variable gender equal to ‘male’.
Discussion: This prototype, besides contributing to the detection of data quality problems, is also expected to be an incentive to the improvement of information system architectures. It shows the importance of the development of mechanisms to detect and validate data in health environments.
(More)