Resilient Network Systems submits the following comments on the ONC Draft Interperability Roadmap. Our responses follow the questions posed in the introduction to the Roadmap. We omitted sections for which we had no comment. Thank you for this opportunity to respond, Rob Gingell EVP Prooduct Development, Resilient Network Systems GENERAL COMMENTS ---------------- 0. Overall ---------- The Roadmap is an ambitious set of steps to improve interoperability. It has some solid foundations in the work in the C-CDA, some achievable first steps in the inital Common Clinical Data Set, along with a plan for advancement that potentially aligns the goals of all the parties who will have to converge to achieve interoperability. The riskiest technical barrier to effective interoperability is that of establishing clear semantic meaning for the standardized data elements. Although it's a technical barrier the difficulties lie primarily in cultural and training impediments to having shared meaning. The limited scope of the initial Common Clinical Data Set helps this by providing an initial step on which improvements can be incrementally based with time and experience. It is hard to assess whether this is sufficient, but so long as the Roadmap supports interative, converging developments and assessments it should be a workable approach. In general we would conclude that technology, while essential to fulfilling the goals of the roadmap, is not a primary impediment to its goals. Rather it is more the alignment of the human and organizational participants to ensure that what is communicated has sufficiently precise and accurate meaning to enable the real change in healthcare outcomes. There are substantial incentives in the environment to do so, from changing healthcare policies to the economic drivers to trim costs and achieve greater value and results. 1. General ---------- Overall we conclude that the roadmap contains the right actions to achieve interperability and presents a credible and thorough coverage of the issues. The timing of expected events in the roadmap is perhaps ambitious, as much commonality needs to occur early during the roadmap period, but a lot of that commonality depends on cultural and training behaviors of participants which is likely to lag. It may be advisable to push out some of the expectations of having achieved the necessary commonality, perhaps allowing a development cycle with the initial Common Clinical Data Set to "shake out" some of the more deeply embedded impediments to success. Stakeholder alignment appears generally correct but as part of a learning health information system the roadmap should anticipate the appearance of stakeholders not yet well known to the system and the introduction of innovations and approaches which might change expectations for what constitutes successful technical and organizational approaches. RNS experiences this with its Trust Network products in which an applying organization learns to codify policies governing access and release of information externally to the applications which supply it -- and in doing so change their approaches to understanding and managing policies. The roadmap appears to recognize this need by being characterized as a "learning" environment and also in recognition that interoperability achievement may require different changes between dispersed participants. In the past, large efforts have failed by expecting too heavy an influence of centralized definitions which are often too unresponsive to real-world needs, thus becoming one of the impediments to adoption itself. Having a core set of expectations, such as the Common Clinical Data Set, may help to start the process -- but finishing the process will require some branching among more detailed data sets from which commonality is precipitated after the fact rather than required for interoperability to begin. 2. Priority Use Cases --------------------- Of those listed in Appendix H, we would choose: #6, #11, and #27. #6 was chosen because to achieve it requires linkage and closure through an end-result (outcome) that permits assurance that care was delivered and successful. #11 was chosen to ensure the data available is rich enough with relevant data to ensure that it fully replaced alternative paths which are viewed as "more real". This despite the fact that achieving #11 requires a good deal more semantic resolution work than other goals might have. #27 was chosen as a system-wide measure and to support detection threats to the nation's health and well-being. 3. Governance ------------- ONC can support industry-led governance by providing clear metrics, the status of those metrics periodically, and trends among them to allow the industry to have measures transcendent of their individual goals to ensure alignment. ONC has to provide the "nation's interest" for while all of the participants may wish to serve the national interest an industry-led governance would otherwise have too many other masters to serve. 5. Privacy and Security Protections for Health Information ---------------------------------------------------------- Health Information sharing is complex in that elements of a patient's history have varying security and privacy requirements. Further, as information is conveyed, the original policies governing access can become clouded with the transition to subsequent users. Technologies which maintain data privacy and security protection even as the information is communicated amongst several parties would best protect for this -- as the policy can always be tested for each new party gaining access which protects both the information directly as well as providing a protection against inadvertant release by properly authorized parties. The roadmap notes that HIPAA requirements are not always understood are followed and we would maintain that that is partly to avoid the risk of inadvertant release following a legitimate use of the information. 6. Core Technical Standards and Functions ----------------------------------------- Question 6.2 asks whether we believe the approach proposed for Accurate Individual Data Matching will sufficiently address industry needs and current barriers? Our answer is that the approach "sort of" accomplishes this. It specifies steps and criteria all of which are helpful practices for help ensure that consolidation includes the correct target data without conflation with data from incorrect targets. The nature of the problem, however, is one of risk. As practiced in everyday interactions in doctors' offices and in hospitals and clinics, there is a notion of "good enough" for the context in which the assessment is being made. Office administrators routinely determine for themselves that the person standing before them and the records they have in hand and the associated insurance coverage all match up. Pre-operative walk throughs and matching of identities with orders and confirmation of procedures avoid risk while performing a procedure. It's all a question of risk management, and in a given context, the risks are modest while in another they may be quite high. A recognition that the problem is one of risk management rather than of perfect data convergence would enhance the roadmap's goals of providing a "learning" system. 8. Measurement -------------- The Measurement and Evaluation Framework appears to describe a very complete process for arriving at a set of measures. It is necessarily qualitative at this point and we believe the steps planned for the initial period of the roadmap are those appropriate for getting calibrated as to exact measures and their baseline values. SPECIFIC ACTIONS ---------------- Resilient Network Systems (RNS) is working in several areas of direct relevance to the goals of the Roadmap, and in a timeframe that can be supportive of its goals and contribute to its achievement. Specifically, the RNS Trust Network (TN) product set employs and provides for standardized methods for establishing trust between unrelated parties for specific contexts and purposes. That structure can rely upon the semantic commonalities established by the C-CDA to create among parties sharing the goal of interoperability according to the roadmap the means to implement and communicate about it to help entities, associates, and principals to determine what may be shared with whom. Additionally, RNS' technology for communicating policy-protected data supports the Roadmap's critical action of "clarify(ing) privacy and security requirements" by providing granular, policy-protected access to indivudal datums as they transit or are accessed in the course of an exchange of information.