Objective: To determine if there are more accurate, faster methods of data-entry from paper forms to electronic databases using modern, easy-to-learn, and customizable programming languages.
Methods: Case subjects (n=10, dental students) perform data-entry for patient records (n=50) with ten fields per patient using single-entry, double-entry, and a prompt-based data entry program written in the Python programming language. A script also written in Python parses the text data by looking for misspellings, repeated entries, etc., alerts the case subject of potential errors and, gives the user the ability to fix or ignore the potential problem. Measurable metrics will include time, error rates, and initial accuracy. Error rates per field, type of data error (e.g., string, date, medication), and total time to enter data will be analyzed. Data will be analyzed using chi-square tests to determine if there are statistically significant differences between the control double-entry standard, and the prompt-based entry system in dental record data abstraction.
Results: Initial results run on one test case subject were indicative that a prompt-based data entry program saved time and improved accuracy for electronic abstraction from patient record forms. This study, over the next couple months, will expand the data set to fulfill a statistically analyzable result.
Conclusion: A prompt-based data-entry system can decrease data entry error rates and time entering data by permitting a “single-input-at-a-time” method to allow the user to focus on data entry and not on formatting issues. A programming-script parsing the entered data, looking for common data-entry errors, will further reduce entry-error. Modern programming languages can allow researchers to accurately and more efficiently enter large amounts of data from a patient population for case-based analysis in the clinical setting.