Paul Richardson
September 25, 2023
FHIR is now the accepted world standard for exchanging healthcare data. But building FHIR interfaces to existing healthcare systems is not easy. Many organisations have made large investments in building FHIR interfaces, and continue to do so. The costs and timescales continue to be bad surprises.
Why? The tools used to do it are very labour-intensive (thousands of lines of unmaintainable code) or of limited capability (basic mapping tools), or both. These tools are now beyond their limits, and their owners are asking – what can we do next?
The FHIR Transform Engine (FTE) offers a way forward. With its powerful mapping formalism, graphical mapping editor, and comprehensive test tools, it gives a way out of the capability/complexity bind. For complex transforms (i.e. most of them) it is no longer necessary to struggle with inadequate mapping tools, or the opaque code of departed developers.
Yet a move to FTE may appear expensive. Application owners rightly ask: how can we protect the investment in our previous transform tools? Do we need to rebuild the same transforms in FTE, before we can move forward? FTE offers a fast way to cross the river.
One thing you did when developing your current transforms was to build a set of test cases – examples of the source, paired with the target you need the transform to give. Even if the target examples are not easily available, you can use your existing transforms to make them from source examples. Using these example pairs, you can easily migrate your transforms to FTE.
FTE has an AI capability called ‘Transforms By Example’ (TBX) which, given one or more source-target ‘example pairs’, uses them to infer a mapping set and a transform which creates the targets from the sources. TBX is not as powerful as the FTE mappings (after all, it is only AI!) but it gives an excellent start, before refining the mappings for production use. You can use your legacy transforms to make example pairs, which TBX converts to FTE mappings and transforms – giving fast migration to the greater power, quality and adaptability of FTE transforms.
