Saturday, November 12, 2011

Health Information Exchange: Infrastructures and Market Dynamics

Kuperman, G.J. 2011, “Health-information exchange: why are we doing it, and what are we doing?”  Journal of the American Medical Informatics Association, 18(5), 678-682.

This important article offers a very useful conceptual view of health information exchange – set in the context of  US health care sector market dynamics.  The author summarizes the early history of the Nationwide Health Information Network (NHIN) and the HITECH Act of 2009 promoting the introduction of health information technology  (HIT) on a national scale.  The article begins with the vision of health information exchange (HIE) as a key enabler of high quality and efficient health care.  According to the author, early demonstration projects conducted from 2005-2007 have shown that the interconnection of RHIOs for health information exchange in the “network of networks” requires neither a centralized national infrastructure nor a national patient identifier.  Unfortunately, these conclusions are more ideological than scientific, as there is little corroborating evidence in policy or organizational research. 

Since 2005, the eHealth Initiative has reported on the development and sustainability of RHIOs and State Designated Entities (SDEs) across the United States.   Kuperman cites the 2010 eHealth Initiative HIE Survey  findings (available at )  to substantiate the existence of 73 “operational” initiatives, but he does not mention that among those, the report finds that only 18 can be described as “sustainable” – sustained on operational revenue alone and not dependent on federal funding. (See page 2 of the Key Findings.  One of the findings listed on page 1, states that “Sustainability is an attainable goal for health information exchange organizations. There is a small but critical mass of sustainable organizations.  This finding is without adequate foundation in the eHealth Initiative data analysis or other studies of health care organization.  Such data interpretation threatens the formulation of credible policy on health information technology in US system reform.)  The terms RHIO, SDE and HIE refer to organizations that address the” business issues of interoperability”, but critical review of  eHealth Initiative research as well as other published scholarly articles suggests that sustainable business models have not been identified. 

My reviews of these reports reveal some methodological deficiencies that tend to weaken published study conclusions. (See  .  My blog review was completed before the eHealth Initiative redefined their HIE reports as proprietary – despite Federal funding supporting the research.  Only the “key findings” are available for public review.  The validity of such findings cannot be evaluated without access to the research methodology.)  From year to year, the eHealth  Initiative reports that the number of HIE entities has increased - without accounting for sample mortality or changes in their definition of HIEs. 

Another publication (See  Julia Adler-Milstein, David W. Bates and Ashish K. Jha, 2009, U.S. Regional Health Information Organizations: Progress And Challenges, Health Aff , 28( 2) 483-492 available at  .) based in part on data from the  eHealth Initiative (See  )  includes a measure of time spent in HIE planning.  This indicator seems to be negatively associated with operational status of the entities in the sample. While Adler-Milstein et al. conclude that a lengthy planning process may challenge HIE viability, it is also possible that this result reflects the short life expectancy of HIE entities – as time spent in planning may serve as a partial surrogate for longevity. 

These authors have also published another study (apparently based on some of the same data). (See Adler-Milstein et al. 2011,  A Survey of Health Information Exchange Organizations in the United States: Implications for Meaningful Use, Annals of Internal Medicine, 154(10) 666-671 available at )
They elaborated the definition of a “comprehensive RHIO” in light of the HIE requirements for meaningful use of electronic health records (EHRs).  This definition was developed by a panel of 9 national health policy experts using a Delphi methodology to arrive at consensus.  Analysis revealed that none of the RHIOs included in the sample satisfied the criteria of this definition.   This finding suggests the possible failure of the market driven “network of networks” approach to development of the NHIN.     

Such failure may also be attributed to the short time frame (2-4 years) for public funding in support of RHIOs as well as the requirement that they develop business models based on revenue streams from private stakeholders and system users.  Such business models are all the more difficult to identifiy given that the significant benefits of health information exchange often accrue at the system level rather than the individual provider or payer level of analysis.  Adler-Milstein et al. (2011) conclude that their findings “…call into question whether RHIOs in their current form can be self-sustaining and effective in helping U.S. physicians and hospitals engage in robust HIE to improve the quality and efficiency of care.” (See abstract.)  Questions raised in this study suggest that the “network of networks” strategy based on the sustainability of RHIOs cannot be assumed as in Kuperman’s analyses of other projects, such as Direct and Connect  for health information exchange in the context of health sector market dynamics.

The second issue related to health information infrastructures required for national system reform is the creation of a unique patient identifier.  As mentioned above, Kuperman suggests that such an identifier was shown to be unnecessary in early demonstration projects for the NHIN and the “network of networks” approach.  This is another ideologically convenient finding, without an evidence base in policy experience or organizational research.  Kuperman argues that the advantage of a PUSH model such as Direct is to avoid the necessity of linking patient identifiers across systems before data transfer between health care organizations.  While directories of authorized organizations (and their identifiers) would have to be established – individual patients would be identified by the authorized senders and receivers using internal matching algorithms– or even manual procedures.  This approach would probably be effective in small-scale systems,  but may be impracticable at the regional and national levels (not to mention the global level) as the volume of data increases with mobility and diversity of patient populations served.

According to Kuperman, clinicians will expect both PUSH and PULL service dynamics for health information exchange, including transmission among providers as well as retrieval of individual patient data across the entire health care system.  The responsibility for designing and managing these services apparently resides with RHIOs under the assumption of their sustainability: “As RHIOs (grapple) struggle to support interoperability-based services that improve the quality and efficiency of care, they will have the opportunity to understand how best to combine pull- and push-oriented capabilities.” (page 681) Given the ongoing failure of federal investments in RHIOs and the “network of networks” strategy to develop infrastructure, this policy direction lacks credibility and remains unfounded in research evidence or policy experience.

No comments: