Interview with NextGate VP Joerg Schwarz
NextGate handles electronic health records from a unique perspective. Focused on accuracy and efficiency, their MatchMetrix technology platform seeks to weed out duplicate health records using probabilistic algorithms. The goal is simple. Make health data clean, clear, and usable, so the focus can remain on the patient’s health. We recently spoke with NextGate Vice President of Business Development Joerg Schwarz to hear more about the dynamic landscape of healthcare IT solutions.
Zach Urbina: EMRs and EHRs are the big systems today, collecting a variety of enterprise data. Are they adequate or sufficient to accomplish the information technology requirements for an accountable care organization or an HIE? If not, why not?
Joerg Schwarz: Neither EMRs nor EHRs are sufficient for an ACO. Both of these information systems have been designed to coordinate intra-organizational care and are insufficient to coordinate care across a group of providers, analyze data to improve care and support prevention. Most HIEs collect data published by these EMRs into large repositories, but fail to achieve semantic interoperability, which would be required to aggregate, consolidate, and transform data into meaningful and timely information.
ZU: What infrastructure technology is needed to enable the ACO? Are there multiple components?
JS: There are a number of components needed that can be added to existing EMRs and HIEs to enable them for ACO purposes.
- Master Patient Index to reliably identify and map patients across multiple systems, allowing direct data exchange with Laboratories, RIS & PACS systems, Practice Management and Revenue cycle management systems, Payer claims data, and other HIEs
- Provider registries and specific entity and person directories that track credentials, specialties, contact and contract relationships.
- Relationship registries that map relationships between patients and providers, and providers and other providers. This is important to coordinate care, send relevant alerts and notifications to providers, for example for the coordination of post discharge coordination that could prevent early readmissions.
- Transaction and Activity registries to group transactions and map them to episodes of care across organizational boundaries. This is important to determine and improve cost of care delivery and harness incentives for bundled payments.
- Code-set and Terminology registries that map coded data to a canonical format and thus achieve semantic interoperability.
- Portals for care coordinators and patients to access relevant information and track progress.
- Analytical tools that determine risk groups and help identify cost and quality improvement potential.
- Care management tools that detect gaps in care and help implement best-practice care plans. It is ideal if the tools are based on a common platform and standards to increase operational efficiency and coherence, but some of the tasks and contents are very specialized and can be provided by different parties. The important criteria is to follow common data standards for interoperability and coding and implement robust interfaces.
ZU: How does that infrastructure “connect the data dots” or attach knowledge to the data, giving providers and business units relevant and contextualized information to make improved decisions?
JS: After data collected from EMRs and HIEs is transformed into semantically coherent elements, patients can be grouped by various factors, for example by co-morbidities, disease states, and compliance. This can be useful to develop specific programs for these cohorts, i.e. assigning a nurse as a care coordinator for high risk/ high cost cohorts or providing wellness programs to borderline cohorts. Not only does it help to identify these groups and their providers to coordinate these programs, it allows also to measure success rates akin to a quality improvement cycle. So “connecting the dots” requires more than connectivity, even more than interoperability. It requires semantic interoperability and analytics to provide evidence based decision support that helps to provide preventive and continuous quality improvement.
ZU: As separate organizations exchange healthcare data, how do they ensure they are sharing data on the same person?
JS: This is an essential concern. Accurate patient matching is essential for many clinical and financial reasons. Different systems usually assign their own local identifiers for a person, and all of these local identifiers need to be mapped to a system wide identifier. This identifier should allow searches and matching based on many different data elements, including historical data. Relying on simple identifiers such as name and even social security number has proven to be unreliable as names can change or be misspelled, social security numbers can be incomplete, incorrect or simply false. The more factors can be considered to establish an identity, the more likely can clinical quality goals be met, and errors like wrong medication or procedures be avoided. Resolving redundant identities can be costly. Some estimates are ranging from $20 to several hundred dollars to resolve duplicate or suspicious records. It is there important to deploy tools that can achieve a high level of accurate automatic matching and only require manual resolution efforts in about 1% or less of cases.
ZU: How have the roles of data governance, data quality management, and consent management changed as a result?
JS: Data Governance used to be germane to each organization. Certain people would have roles in data governance that could be specific to that organization. In multi-organizational environments, this becomes a more complicated problem and requires careful consideration and workflows. When multiple organizations contribute conflicting data that requires manual intervention, it is useful to allow only one organization at a time to verify and correct data entries. At the same time, it is also useful to notify organizations about changes to the golden record, for example address changes or a deceased status. A full featured EMPI should therefore allow configurable workflow management and rules-based outbound messaging. At the same time, it is important to implement access controls that allow configurable views and edit rights, so that users from institution A might see, but not change data from organization B and C, but might forward a review task to the other organizations. Data governance becomes thereby a team effort, with strong acces controls and audit trails.
ZU: Does this infrastructure require a large investment in time and money? Can it be implemented in phases?
JS: Good news for everybody who invested in recent years in EMRs ad HIEs; rip-and-replace is not necessary, and neither are monolithic solutions that require massive investments and implementation projects. The ACO infrastructure layer outlined here can be implemented in phases and usually provides Return-on-Investment (ROI) in less than one year for each phase. Reducing manual duplicate record handling from 7.5% to 2% requires investment amounts with less than six months ROI. Implementing Provider and Relationship registries can reduce administrative costs and improve care coordination with equally attractive ROIs. Implementing sophisticated care plans requires more time and investment, but in light of the financial pressures from government and financial payers, i.e. with regard to early readmissions and bundled payments, ROI can be excellent even outside a full ACO concept, and in an ACO environment, it is mission critical.
ZU: Who actually is responsible for implementing the infrastructure — all sides of the data exchange, or just a few of its participants? How do messaging standards like IHE profiles, or the DIRECT project aid in the exchange?
JS: These transformational activities can be planned and implemented by small Independent Physician Associations (IPAs), Independent Delivery Networks (IDNs), or ACOs, which usually also include Payer organizations that delegate capitated risk. The goal is usually to reduce the impact on the edges, meaning the individual practices and doctor’s offices, and perform the data transformation without impacting existing workflows. Many physicians and hospitals invested just recently in Meaningful Use (MU) certified infrastructure, and this can be leveraged to collect data and send alerts and notifications. NwHIN Direct is a good example for a non intrusive technology that use to a large extent web based secure messaging clients to transport patient information directly to the targeted provider. Many of the essential tasks of care coordination can actually be performed without the need for an HIE infrastructure by utilizing EMRs and other clinical information sources, such as Laboratory Systems, PACS and RIS systems, ePrescribing interfaces for data input and NwHIN direct secure messaging for notifications (i.e. post discharge care instructions, gaps in care, immunization reminders etc.). More sophisticated and interactive decision support can be provided by portal applications that augment existing EMRs.
ZU: Who can best provide this infrastructure: large vendors, or smaller companies?
JS: Large vendors are usually good in solving big problems. Developing care guidelines and risk stratification requires a lot of domain expertise and resources that are hard to provide by small companies. Many of the aforementioned infrastructure components can provided by innovative, agile companies that are focussed on the topic and are small and agile enough to develop these innovative technologies, but large and experienced enough to provide solid support and services to provide end-to-end solutions with attractive ROI and implementation times. Time is of the essence to gain a competitive edge, therefore it could be fatal for your readers to tie themselves to large vendors that will take years to provide working solutions.
Next Gate joins our Fall 2012 Accountable Care and Health IT Strategies Summit Nov 27-29, in Chicago. Registration information.