In my mind, the future of personalized medicine, the emergence of -omics labs, the cloud, and the future of HIT are quickly converging. The pace of this conversion mpw seems to be accelerating based on a recent brief note that I came across in medGadget (see: Dell to Help Develop Personalized Therapy for Pediatric Cancers). Here is the note in its entirety:
At a [recent] press conference ...Dell unveiled a commitment to a project that, with help from massive computational resources, will help fight childhood cancer. Specifically, Dell will install 8.2 teraflops worth of cloud servers at the TGen Translational Genomics Research Institute in Phoenix, Arizona to help with sequencing tumor genomes that can then used to select a personalized therapy for each child. Moreover, the genomes are kept in a database that is used by subsequent patients to compare their cancer DNA to those of others. The institute is currently capable of performing the molecular characterization and DNA comparison for each patient, but the process takes weeks. The new computers should bring this process down to hours, a significant breakthrough when fighting aggressive cancers. The FDA has already given the go ahead for a clinical trial involving 14 kids with neuroblastoma, a rare cancer with difficult to identify treatment options. This will be the first FDA approved trial using molecular-guided individualized therapy and so, in a sense, it will herald a groundbreaking in this exciting field.
In bulleted format, here are my thoughts about this news and the ramifications for HIT:
- We are at an evolutionary dead-end with our current EMR technology. Hospitals will happily pay tens, if not hundreds, of millions of dollars to use old technology designed to mimic the traditional paper medical record.
- All of these EMRs are transaction-based with little "intelligence" and even less ability to integrate and analyze the deluge of patient data generated by the EMRs and their "feeder" systems such as the LIS.
- LISs are highly functional -- pathology and the clinical labs would be unable to function for a day without them. Nevertheless, LISs vendors are not keeping up with the development of modules that can support the exploding -omics science.
- It is highly unlikely that many of the incumbent EMR and LIS vendors will be able to jump to a new generation of analytic healthcare information systems. They are too tied to their current technology, current business models, and their need to recover their sunk development costs.
- In my opinion, most of the progress in cutting-edge healthcare computing will be generated in academic research labs, non-profit biomedical research institutes like TGen, and smaller for-profit reference labs.
- Evidence for this statement is supplied by the note above with these key facts:
- Dell installing 8.2 teraflops worth of cloud servers.
- The sequencing in hours of the genomes of aggressive pediatric tumors.
- Development of massive databases for tracking the genomic changes of individual tumors and making genomic comparisons across patients with the same tumors in the pursuit of better chemotherapy.
Here's what I think that the future holds. Hospitals will continue to install their transaction-based EMRs, LISs, and RISs. They will continue, in the short run, to pay premium prices for products from a company like Epic because of the monopoly it holds on the larger-hospital EMR market. Meanwhile, the "-omics cloud" will continuously evolve on the research and reference lab side, providing the answers pertaining to personalized medicine. There will be no need to install "personalized medicine servers" in hospitals. The most likely scenario will be that hospitals will develop working relationships with the developers of the best research-oriented "-omics clouds", sending them relevant diagnostic and clinical data from the hospital-based EMRs, LISs, and RISs. These research clouds will thus support "personalized medicine" as part of their mission but at a fraction of the cost of current EMRs.
Comments