Abstract
Background: In dentistry, the use of electronic patient records for research is underexplored. The aim of this paper is to describe a case study process of obtaining research data (sociodemographic, clinical and workforce) from electronic primary care dental records, and outlining data cleaning and validation strategies. This study was undertaken at the University of Portsmouth Dental Academy (UPDA), which is a centre of education, training and provision of state funded services (National Health Services). UPDA’s electronic patient management system is R4/Clinical +. This is a widely used system in general dental practices in the UK.
Method: A two-phase process, involving first Pilot and second Main data extraction were undertaken. Using System Query Language (SQL), data extracts containing variables related to patients’ demography, socio-economic status and dental care received were generated. A data cleaning and validation exercise followed, using a combination of techniques including Maletic and Marcus’s (2000) general framework for data cleaning and Rahm and Haido’s (2010) principles of data cleaning.
Results: The findings of the case study support the use of a two-phase data extraction process. The data validation processes highlighted the need for both manual and analytical strategies when cleaning these data. Finally, the process demonstrated that electronic dental records can be validated and used for epidemiological and heath service research. The potential to generalise findings is great due to the large number of records. There are, however, limitations to the data which need to be considered, relating to quality (data input), database structure and interpretation of data codes.
Conclusion: Electronic dental records are useful in health service research, epidemiological studies and skill mix research. Researchers should work closely with clinicians, managers and software developers to ensure that the data generated are accurate, valid and generalisable. Following data extraction the researchers need to adapt stringent validation and data cleaning strategies to guarantee that the extracted electronic data are accurate.
Method: A two-phase process, involving first Pilot and second Main data extraction were undertaken. Using System Query Language (SQL), data extracts containing variables related to patients’ demography, socio-economic status and dental care received were generated. A data cleaning and validation exercise followed, using a combination of techniques including Maletic and Marcus’s (2000) general framework for data cleaning and Rahm and Haido’s (2010) principles of data cleaning.
Results: The findings of the case study support the use of a two-phase data extraction process. The data validation processes highlighted the need for both manual and analytical strategies when cleaning these data. Finally, the process demonstrated that electronic dental records can be validated and used for epidemiological and heath service research. The potential to generalise findings is great due to the large number of records. There are, however, limitations to the data which need to be considered, relating to quality (data input), database structure and interpretation of data codes.
Conclusion: Electronic dental records are useful in health service research, epidemiological studies and skill mix research. Researchers should work closely with clinicians, managers and software developers to ensure that the data generated are accurate, valid and generalisable. Following data extraction the researchers need to adapt stringent validation and data cleaning strategies to guarantee that the extracted electronic data are accurate.
Original language | English |
---|---|
Pages (from-to) | 88-94 |
Number of pages | 7 |
Journal | International Journal of Medical Informatics |
Volume | 127 |
Early online date | 12 Apr 2019 |
DOIs | |
Publication status | Published - 1 Jul 2019 |