Data flow

From ECHSA Congenital Database
Revision as of 23:32, 23 September 2005 by MaciejBlizinski (talk | contribs)
Jump to navigationJump to search

The Data Flow until the end of 2004

There were two basic ways of submitting data to the database.

echdd6.exe

This is the piece of software used to collect and submit data. echdd6 is written in C++, and is available to download in binary form from ECDB's website. This application is avaliable only for Windows platform, and it's intended to be installed locally in the hospital, on the PC. Some Windows systems require additional database libraries installation. They are also available to download.

The echdd6 application makes use of the Windows database libraries and stores its data in MDB file, which can be read with Access97 application. Hence the need for database libraries.

This application allows users to enter the data and export it. Exporting means sending the data to the data manager, who imports the it into main database. Not all of the data is exported. The patients' names for instance, remain in the local PC.

Web interface

Web Client Interface, also referred as the Data Entry Form is intended to be substitute to the client application. It is written in Python language and allows users to enter the data directly into the main database, without need of exporting. However, the data is not visible in the reports immediately, because it must first pass validation procedure.

Data Validation Procedure

The DataValidation Procedure was designed to improve the quality of data by finding inconsistencies in the data. For example, the patient's birth date can't be later than the date of the operation. There is a number of validation rules that the record has to meet in order to be published. Some of the data may be labeled "incorrect" by the computer by may in fact be correct which the ownewrs of the data can confirm, for example unusuall age and weight combinations. In such case, the hospital can mark the record as correct, despite not meeting the validation rule (see ValidationRules). The last word belongs to the data manager. He is the one to finally accept or reject the record.

{i} Please note that data validation does not ensure that the data reflects the reality of the facts. It only prevents the data from containing obvious inconsistencies. Another method of improving the quality of data is called the DataVerification.

DataFlow.png

The data-flow graph. -- This graph shows the data flow, concentrating on the data validation procedure.

The validation rules

The validation rules are stored in the file ValidationRules.xml, which is a XML file of the simple format. Validation rules apply to the "operation" table only. If one or more rules are not matched, the record appears in "data validation" part of the web interface, and the hospital is asked to either correct the data, or to check that the data is correct.

The Validation Rules file is in XML format for flexibility and easy editing. You can read the rules here: ValidationRules.xml. The validation mechanism was re-written from old CSV-like config file in August 2004.