I had a conversation with a journalist recently about whether hospital labs and reference labs are selling anonymized lab data to pharmaceutical companies, the most logical buyer of such data. This idea has been percolating for decades in clinical lab and healthcare circles without much (apparent) success. However, even if it were occurring, it would be in the shadows. Hospital and reference lab executives would not relish the blowback from patients if such sales were to become public. Moreover, lab data, in and of itself, would probably not be that useful without the accompanying patient demographic and clinical data.
Having made this point, our conversation turned to the very valuable and recently developed "research databases" in hospitals. The data is extracted from EHRs and contain lab test and imaging results, drug history, clinical data, and patient demographics. EHRs, despite their high cost, are not designed to respond to even relatively simple research questions (see: Hospital EHRs Inadequate for Big Data; Need for Specialized -Omics Systems). Hence the need in order to create research databases by the transfer of data from EHRs to powerful, dedicated computers (see: Mount Sinai's Icahn School of Medicine Pioneers the Use of Big Data to Study Diabetes).
I blogged in 2014 about the development of a research database at Memorial Sloan Kettering Cancer Center (MSK) in collaboration wiht Quest Diagnostics (see: Memorial Sloan Kettering and Quest Form Strategic Alliance). Here's an excerpt from that note pertaining to the value of cancer genomic data as a component of a hospital research database:
Cancer genomics differs radically different than the other tests that Quest offers to its mainly office-based clinician clients and these differences offers clues to the benefits of the alliance for both parties. First of all, MSK controls a treasure trove of a wide variety of biopsied and resected malignant tumors that can be genetically analyzed. This biorepository is immensely valuable and much different than the standard serum samples that Quest acquires in a huge volume on a daily basis. Moreover, molecular and genomic testing is quickly changing and Quest wants to share some of the new knowledge acquired by the MSL bench and clinical scientists.
Amother very prestigious cancer hospital in the U.S., M.D. Anderson, also set out to develop a research database with IBM as a partner and quickly became immersed in a scandal which is still playing out (see: Scandal at M.D. Anderson -- Operating Loss and Then Watson Deep-Sixed). The reference to IBM's AI initiative, Watson, emphasizes a very important point with relation to clinical research databases. Extracting usable information from these systems requires participation by highly skilled IT professionals who understand big data and deep learning (see: Big Data Coming In Faster Than Biomedical Researchers Can Process It). IBM Watson is also a partner with MSK in their cancer research initiatives (see: IBM Watson, Quest Diagnostics, Memorial Sloan Kettering partner for cancer research).
So what does all of this have to do with "monetizing" hospital clinical data. These new integrated clinical database have great value for both basic research and for pharmaceutical companies. Just imagine some of the questions that could be potentially answered by such systems. They include the following:
- To which drug(s) are patients with Cancer A responding the best?
- Considering image analysis of biopsies and imaging studies of patients with Cancer B, which drugs are proving to be the most effective?
- For patients with Cancer C, which basic laboratory tests are most effective in diagnosing the lesion and predicting survival.