Hypothesis / aims of study
The proper storage and management of data is an essential requirement for clinical services and research. Our department has operated a database for symptomatic, urodynamic and diagnostic data since 1985, beginning with a DOS based system and now under Windows. To date it contains the records of 32,637 urodynamic tests (8,289 male, 24,348 female) made since then, with additional records of over 30,000 free uroflowmetry tests. It is to our knowledge the largest of its kind in the world, and this is the first description in the literature of such a database.
We present here the structure of the database, with an aim to enable departments to set up similar systems. Closer harmonisation of database structures will enable easier collaboration and review of data across centres. We also share systems for screening out anomalies and errors in the data, aiming to assist others in management of data quality.
Study design, materials and methods
The structure of the database is described in terms of data sections and tests recorded. The screening tests are listed with guidelines on plausible data ranges.
The database is a proprietary system developed by our hospital's Information Management and Technology department. Data entry includes demographic, history and symptom details. Summaries of the bladder diary, physical examination, urodynamic test details and results, and the suggested treatment plan are also recorded, followed by text reports of both history and findings. This structure is summarised in Table 1. Data entry of symptom and history is made during patient interview and the test itself, making data entry efficient. Once test data is entered, the database can then generate a test report in the form of a letter for referrer and patient.
Data checks are made in the form of consistency checks (e.g. male data for male patients only), plausibility (e.g. daily micturitions < 40) and field content (e.g. text ‘G3 P2 A1’ is not allowed in the numeric ‘Parity’ field). These checks are summarised in Table 2. In this way, most typographical mistakes and barriers to automatic analysis are removed.
Interpretation of results
Other departments are able to use this structure to ensure all relevant and useful data is recorded. Similar structures for data storage will allow greater ease of data sharing for multicentre reviews.
With or without a database in place, the screening checks presented here can be used to remove anomalous data from records. Remaining outlying data can then be checked and validated against original trace data or patient notes if required.