In a recent blog note about "big tech" knocking on hospital doors related to commercialization of diagnostic and predictive algorithms, I ended with the following statement (see: Big Tech Is Knocking on Hospital Doors; It's All About the Data).
Regarding healthcare analytics in general and diagnostic analytics in particular, the development of these fields must take place in collaboration with the IT companies that have the special expertise to analyze the data using techniques such as deep learning, big data, and neural networks. Such a collaboration between these two parties will not be easy because of what I envision as "points of contention" that I intend to discuss in upcoming blog notes.
In today's note, I want to list a few specific points of contention that I think will arise in the future development of diagnostic and predictive analytics. These differences will color the relationship between hospitals that act as honest stewards of patient data and the for-profit IT companies that will bring analytics products to the market.
(1) Clinical data ownership and the monetary value of algorithms derived from such data
Providers (i.e., hospitals and physicians) own/control the data in health computer systems (e.g., EHRs, LISs, RISs) (see: Who "Owns" PACS: Radiology or Central IT in Hospitals?; Who Owns EHR Data? What constitutes proper data stewardship?). There have been relatively few conversations in the past about the question of hospital data ownership because the data had little practical value beyond its use to manage individual patient treatment and research. However, such ownership now looms as an important issue and discussions need to start quickly. As only one example, we are on the verge of the availability of an algorithm based on EHR and LIS data that can predict fairly accurately which male patients will develop aggressive cancer of the prostate (see: Genetic Risk Scores: Ready for Inclusion in the Medical Record?). If early intervention for such patients could save lives and reduce medical expenditures, what would be the value of such an algorithm and who would benefit financially from its use in hospitals?
(2) Cultural differences between healthcare and IT cultures regarding entrepreneurship
Many of the largest and prestigious providers are non-profit institutions although some have created business arms that are involved in venture capital investments (see: Hospitals discover their inner venture capitalist). In upcoming years, analytics products based on hospital patient data will be coming to market and many hospitals are being approached regarding access to their EHR, LIS, and RIS data. The recent controversy at Memorial Sloan Kettering Cancer Center (MSK) regarding access to the surgical pathology image archive is a case in point (see: Controversy at MSK Cancer Center Regarding the Pathology Archive and Database). The problem there may have been resolved by revoking the exclusive access to the surgical pathology database by a hospital-funded start-up but additional disputes will undoubtedly arise.
(3) Different perspectives on wellness and disease states
Despite protestations to the contrary, most physicians and hospitals have relatively little interest in actively promoting the wellness of patients. Even the notion of health monitoring/analysis outside of the walls of a clinic or hospital is questioned by fears of faulty data. Many IT initiatives for healthcare consumers are focused on "wearables" whereby consumers can be empowered as first-line diagnosticians. With a wearable, they can be alerted at home to their atrial fibrillation and then notify their cardiologist that they are on their way to the office (see: Smartwatch Technology Detects Afib Prior to Cardioversion).
Comments