Semantic Web Return of Investment (ROI) for HCLS
This page is to serve as a discussion on Return of Investment (ROI) for utilizing Semantic Web specifications within Healthcare and Life Sciences.
ROI is defined here in terms of what aspects of healthcare and research informatics are either too costly or intractable by current standards and technologies, thereby presenting the semantic web as a possbile solutions candidate.
Key features of the Semantic Web that resonate well (i.e., have unique potential value) with the industry include:
- identifier universality (URI)
- structural uniformity (triples)
- support for formal semantics (ontologies)
- adaptive/open nature of RDF graphs (incremental additions)
- annotative capabilities (focused statements and data provenance)
- ability to map to existent data
- growing acceptance by IT vendors
Drug Discovery and Development
Within Drug R&D, ROI can be identified at different levels:
- Semantic Web ROI is based on the possible savings as compared to the current cost of data integration, and being limited by closed vendor solutions; this by itself is about ~$5-10M/yr/lg pharma. Here's a quote from a Sanofi-Aventis IT manager that provides a useful data point:
- "Depending upon the scope (e.g. number of users, number of impacted sites, regulatory compliance), replacing stand-alone systems can run in the millions of dollars. A major project (a query platform for sharing all research data) recently completed at a global pharmaceutical company included in its scope over 2000 named users and a number of sites, but did not have regulatory concerns. Nonetheless,the total cost of the four and a half year project was over $20 million. Conversely, a similar project with which I am familiar will focus specifically on one site and cost less than $400,000. In both cases, the nature of the project is to replace multiple, disparate systems with an integrated one. Of course, financial cost is only one aspect of such maintenance; scientists spend an inordinate amount of time converting and distributing data in a way that will fit existing systems. Moreover, these rigid systems lead to noncompliance in data storage, retrieval, and ultimately sacrifice considerable investment and power in analytical tools for simple, yet flexible, decision engines. In total, the cost of multiple siloed systems may in fact be many times greater than maintenance alone, particularly when considering the sub-optimal decision-making and lack of learning that such environments promote. "
- Brian Ellerman, Scientific Computing & Information Program Manager Sanofi-Aventis
- "Depending upon the scope (e.g. number of users, number of impacted sites, regulatory compliance), replacing stand-alone systems can run in the millions of dollars. A major project (a query platform for sharing all research data) recently completed at a global pharmaceutical company included in its scope over 2000 named users and a number of sites, but did not have regulatory concerns. Nonetheless,the total cost of the four and a half year project was over $20 million. Conversely, a similar project with which I am familiar will focus specifically on one site and cost less than $400,000. In both cases, the nature of the project is to replace multiple, disparate systems with an integrated one. Of course, financial cost is only one aspect of such maintenance; scientists spend an inordinate amount of time converting and distributing data in a way that will fit existing systems. Moreover, these rigid systems lead to noncompliance in data storage, retrieval, and ultimately sacrifice considerable investment and power in analytical tools for simple, yet flexible, decision engines. In total, the cost of multiple siloed systems may in fact be many times greater than maintenance alone, particularly when considering the sub-optimal decision-making and lack of learning that such environments promote. "
- A more forward-thinking ROI is the need for drug companies to shift to a "Translational Research" (TR) approach (as proposed by FDA), which would require a re-configuration of their IT infrastructure, and the ways scientists would utilize their data; no good commercial offerings exist yet to handle this, and the cost of not reaching these goals within 2-3 yrs could be in the $billions. This is especially pertinent to improving the testing and validation of drug safety, and may be where most of the early ROI is realized. Most drug companies are quite aware that they need to head in this direction, so HCLS has adopted "Translational Medicine" as a key vision to develop towards, which also applies to healthcare providers. A description of what is desired is cited from the National Human Genome Research Institute:
- "The challenge to researchers and scientists now is to determine how to read the contents of all these pages and then understand how the parts work together and to discover the genetic basis for health and the pathology of human disease ...Clinical opportunities for gene-based pre-symptomatic prediction of illness and adverse drug response are emerging at a rapid pace, and the therapeutic promise of genomics has ushered in an exciting phase of expansion and exploration in the commercial sector."
- "Different cultural approaches to systems, driven by the different silos of the organization, are the most difficult aspects of creating integrated systems in the life sciences. Perhaps more importantly, due to the siloed nature of business processes and management of the business, IT is also, by design, siloed. This leads to optimization of systems in that vertical silo, but does not provide horizontal optimization." - Geoff Odell, CIO, Genomics Institute of the Novartis Research Foundation.
- The ROI for 'participating' in W3C-HCLSIG can be presented in terms of a) better understanding and utilization of SW through HCLS participation, b) early access to critical best practices, c) awareness of vendor implementations upfront in the context of SW activities, and d) helping advance solutions for TR in a more rapid and coordinated way than is possible by only single vendor interactions.
How well the first two points are addressed by SW technologies will determine what the actual ROI is for applying SW within pharmaceuticals.
Examples of ROI
Eli Lilly presentation on R&D Productivity
- [place additional examples here]
Healthcare
The ROI is tied to the need to advance healthcare towards a more effective and personalized approach, while at the same time payer agencies (the Federal Government through Medicare/Medicaid) and insurance agencies (Blue Cross and Blue Shield) move towards a pay for performance model. Currently, the healthcare market is balking to advancements due to healthcare inflation. For example "eight datapoints are now kept for each patient (age, height, weight, etc.), but with new tests and personalized medicine, the number could balloon to 3000. The Semantic Web can reduce the cost and liability of knowledge management" (Dr. Tonya Hongsermeier)