NHIN Implementation Experience
We just completed an MPI implementation using Mural for one of our NHIN clients and we were very pleased with how robust and stable Mural was. In the process of loading a few million patient records however we discovered performance problems with the loader packaged with Mural. Loading 200,000 records took over a day, and after waiting several days to load millions of records we simply aborted the process. We dealt with this by using our own batch loading application which we developed in house for other EMPI customers over the past several years. The loader uses its own efficient file based indexing techniques that cuts out all of the latency associated with database operations. We can process 200,000 records in approximately 30 minutes on a laptop, and millions of records in about 12 hours. The loader is now part of our standard MatchMetrix distribution and is a critical part of each of our implementations.
