Data integration, known in the pharmaceutical world as ISS (Integrated Summary of Safety)and ISE (Integrated Summary of Efficacy), is a multi- latitudinal task, and is well recognized for its tediousness and time-consumption. Therefore, it is desirable to automate such a time-consuming task so that a routine process can be developed. Is this wish realistic or achievable to a certain degree?
Research data usually come from multiple sources. An epidemiology study may use historical controls from different time periods and geographic locations in a case-control study. A health economist may want to combine physician data from the AMA database with those of a company’s database.
In a typical clinical study, data may come from different trial sites, especially from international sites, from different labs, or from parallel protocols.Furthermore, for the convenience of analysis, statistical analysts in a multi-protocol study may create their own derived variables or flags for their own individual protocol, which are usually not communicable to each other across protocols
This system tries Data from different sources are rarely the same. In many circumstances, therefore, data integration is an essential step fo r standardized, consistent, and homogeneous reporting. Integration is in fact the process that uniforms different data sets so that they can be pooled together. Because the process is typically done manually, it is tedious, time-consuming, and can easily frustrate or overwhelm the most experienced programmers by the detailed attention and the large amount of home work.
Is there anything that can be done to reduce such a workload? Being aware of a lack of sucha technical tool for data integration in the industry, Kindly share your inputs in regards to this.
