Computerization of Patient records

GR Sridhar, Endocrine and Diabetes Centre, Visakhapatnam - 530002.
as published in BMJ 1998; 317: 1594
Information systems introduced carefully can aid clinicians

 To Bottom  Article Index  Previous Article  Next Article

Clinical computerisation from India

We wish to share our thirteen years of experience, which led to computerizing patient database in Endocrine and Diabetes Centre at Visakhapatnam, south India.

We started with free text case records on paper. Later the clinical features, depending on the frequency were codified into a structured data entry paper form. Four years into this, we shifted data collection into a computer, beginning with FoxPro, a DOS database software.

A series of separate flat data files were designed on diabetes, thyroid, obesity, short stature, infertility and miscellaneous endocrine disorders. Separate follow up files were created for each of these categories. Every new patient was assigned a unique identification (‘EDC’) number, in a separate record.

Later we migrated to a RDBMS system, viz Paradox ver 4.5 (Borland) and later upgraded to ver 7, for Windows 95. As the number of records began to increase, problems showed up: eg duplication of EDC numbers, mis -typing of values etc.

As the latest upgrade, we exported the data to Access relational database software (Microsoft), which had built in facilities to verify and fix data duplication and missing data. Its built in queries '‘Find duplicates'’ and '‘Find unmatched’ helped in identifying and correcting errors. Several records had to be renumbered after verification from the paper register containing EDC number, name and diagnosis.

The new design took considerable planning. We shifted the patient demographic details into a common file, identified by the incremental unique EDC number, generated by the software. One could then go into a particular file (diabetes, thyroid, etc) with follow up data shown at the bottom of the screen. Thus three files ; patient demographic date, case record and follow up were simultaneously displayed. In addition, we automated diagnostic labels and calculations based on previous entries (ie, age of onset, current age, duration of diabetes at follow up, body mass index, hypertension, ischemic heart disease, peripheral neuropathy, hyperlipidemia, hypertriglyceridemia, hypercholesterolemia).

The main obstacle in developing the system was to design an interface that was intuitive, flowing according to the thought process of the physician. It should not interfere with patient care -- data entry into the computer being a distraction. This took time and effort. Once the framework was set, communication with a software professional helped in designing an easy to use interface. The difficult part for physicians is usually in codifying the data entry, deciding what and how much detail to include. Once there is cohesion in this one can derive a lot more power from information technology, than mere generation of bills or laboratory investigation reports.

 To Bottom  Article Index  Previous Article  Next Article