Several years ago, I led a large-scale data integration project to build a data warehouse that compiled relevant data from multiple CROs (i.e., product, study, site, and subject data). The sponsor realized a number of benefits, including saving money (they could retire many IT systems), better and faster access to their data, and better metrics across their CROs. Overall, the project provided a great ROI and gave the sponsor a competitive advantage.
The visionary data owner knows the power to be gained by providing better tools and access to valuable study information, but it can still be a very complex and intimidating project. To get the most out of the investment, keep these things in mind.
1. It’s all about planning.
Spending the time upfront to develop a clear process, identify risks, and align stakeholders will go a long way.
2. Understand your business process.
Ensure everyone is aligned on the new business process being implemented to support the larger initiative; and, if the integrations will be rolled out in phases, ensure the phased business process changes are clearly communicated.
3. Definitions matter!
It was a dark and stormy night (literally) when 100 people sat down to review, debate, and align on attribute definitions. Don’t take for granted that common attributes (like First Subject First Visit, Last Subject Last Visit, and Site Status) are commonly understood.
4. Spend time on your integration process design.
Build out an exchange framework first (such as the hierarchy of XML files) as well as the keys that hold the data together; then prioritize the critical attributes and include them in the data transfer. Failing to do so could extend your project timeline 4-fold. Note: Following an agile process can be beneficial here.
Data integrations and data warehouse implementations are complex projects, but don’t let that intimidate you. Plan carefully, take the time to align processes and definitions, and design a robust process for the integration, and you can realize the benefits and opportunities that come from stronger ownership of your data.