We're not used to seeing the classic issues of commercial data management discussed in scientific journals, so a recent paper in Nature Biotechnology stands out. The authors look at healthcare data networks and recommend what they call federalist principles. In brief, they say that everyone who needs to use healthcare data should use common data standards, but they can store and control their data locally.
The distinction made here is an important one. It's not always appreciated in business, perhaps because the location of data is seen as a housekeeping issue for the tech guys to worry about. It may also be a by-product of the data warehouse movement and similar fashions for gathering corporate data together into a unified resource. So most data warehouses were and are virtual – but it's always easier to show a slide with one big box in the center.
As the vital role and value of data becomes more obvious across all sectors, the distinction between meaning and location is one that all decision makers must grasp. Without this understanding, bad decisions get made. Maybe the most important examples lie in the area of privacy. For a start, if you don't have uniform data standards, you can't manage personal data well and are more likely to use it inappropriately or allow it to feature in a breach. On the other hand, you can use the best data standards available but if you let personal data sit on an unprotected server that you don't control, you're asking for trouble.
Raise these issues with someone who doesn't get the meaning/location distinction and their reflex response might be to nail down all data in-house with maximum security, and never share it outside the organization. Or they may even see failure to use data standards as a kind of protection. Either way, they'd be retreating from the value chains in which they operate. Going back to the healthcare arena, if we don't share patient data safely, we won't learn anything.
It's worth remembering that a meaning/location error might feature in a manager's thinking about data management issues. We need to help them understand that data standards and data security are not antagonists – they're partners. Nature Biotech
Comments