Collection of Electronic Health Records (EHR) for Validation of Artificial Intelligence Based Tool for Data Quality Assessment
Overview
- Phase
- Not Applicable
- Intervention
- Not specified
- Conditions
- Data Quality
- Sponsor
- University of Portsmouth
- Enrollment
- 60000
- Primary Endpoint
- Validity of AI tool detection
- Last Updated
- 4 years ago
Overview
Brief Summary
Electronic Health Record Systems (EHR) play an integral role in healthcare practice, enabling health organisations to collect, access and manage data more consistently. There is also a great deal of interest in using EHR data to improve decision-making and accelerate medical interventions. However, like all information systems, they are prone to data quality problems such as incomplete records, values outside normal ranges and implausible relationships. These problems are expected to become more prevalent as more organisations adopt electronic health record systems, aggregate, share and explore health data. The investigators believe current efforts to improve health data quality can be made more effective if backed by appropriate technology in the form of a readily accessible intelligent tool. Building on this, the investigators developed an Artificial Intelligence (AI) tool for automating data quality assessment of health data. In this study, the investigators evaluate the AI tool using a real-world dataset.
Detailed Description
The main aim of this study is to assess the reliability and utility of an AI tool in identifying data quality dimensions of interest for secondary use of health data, including completeness, conformance and plausibility. In assessing this tool, this study will retrospectively analyse data captured during routine clinical care and identify records containing listed data quality dimensions. This study will also assess the consistency of the AI tool in generating and executing data quality checks.
Investigators
Obinwa Ozonze
Principal Investigator
University of Portsmouth
Eligibility Criteria
Inclusion Criteria
- •No specific exclusion criteria
Exclusion Criteria
- •No specific exclusion criteria
Outcomes
Primary Outcomes
Validity of AI tool detection
Time Frame: 2 months, through study completion
Validity of data quality dimensions identified by the AI tool
Data quality dimensions prevalence
Time Frame: 12 months, between 01/01/2020 and 31/12/2020
The number of patient records identified by the AI tool with completeness, conformance and plausibility violations
Consistency of AI tool
Time Frame: 2 months, through study completion
Consistency of AI tool in generating measures for detecting data quality dimensions