Interview with JEN Associates CEO Daniel Gilden

Date: May 15, 2013||   0  Comments

jen220x145Health policy and big data remain hot topics amid the changing landscape of health IT implementation and strategy.

JEN Associates is a voice in that ongoing policy discussion and Daniel Gilden, their president and CEO will be joining the 2013 State Healthcare IT Connect Summit in Baltimore, Maryland next week to lead a Meet The Innovators roundtable discussion.  We caught up with Daniel ahead of our event to discussion the current landscape of state healthcare IT.

Zach Urbina: Medicaid expansion is bringing coverage to millions more Americans in the majority of the 50 states, How important will it be for states to have Medicare data in order to understand the impact of this expansion?

Daniel Gilden: While many of the expansion populations have been completely uninsured, a critical portion will already be covered by Medicare.  To fully understand the needs and service costs for these populations, states will benefit from access to Medicare data.  Beyond analytics and policy planning, these data will also help states to better manage their populations who are dually eligible for Medicare and Medicaid.

ZU: As Medicaid and Medicare systems are being modernized, what are some of the early opportunities to analyze patient data and make it actionable?

DG: One of the most exciting developments has been the availability of more current Medicare and Medicaid claims/encounter data for program self-monitoring and disease management.  In the past the data has been primarily used for either strategic planning or evaluation activities.   The long time lags usually associated with administrative data availability meant that either a policy planner was using data that was several years old to design new programs and benefits or was evaluating programs years after their initial start.  Because of the data lags, clinical applications of administrative data were very limited.

With reasonably current data the states now have the capacity to support much more direct and timely applications.  CMS is now providing monthly feeds of Medicare claims data for approved applications.  Medicaid data is also available on a similar time schedule from the more technologically advanced Medicaid data systems.  Monthly data feeds from these payers are game-changing from the perspective of program analysts.  New programs can now be monitored in real time with analyses of their impact on outpatient emergency room use, hospitalization, pharmacy management, inefficient provision of services, initialization of care management, behavioral health follow-up etc.  In effect, all of the measures linked to program performance can now be compiled in a time frame that permits active program management, and the detail of the data allows for precise forensics that explain the data trends.  So if a program is not making its “numbers” and generating the expected savings it is viable to plan and implement mid-course corrections.

In the past this type of analysis was only possible years after program start-up – too late to fix problems.  Now, in addition to increasing the timeliness of data to states, data can also be made available to providers, e.g. medical homes, can receive regular reports on their performance versus other program providers. As the relevance of the data to program management and clinical practice is increased, the potential community of data-users also scales up.  Additional opportunities can also be realized by bringing together clinical data from electronic medical records, e.g. lab results, to complement the claims data. The use of these data by the ACOs perhaps represents the best case in which program and clinical staff have incentives to receive and use integrated information on patient care and outcomes across providers and payers.

ZU: State Medicaid Agencies are increasingly partnering with providers to support increased care coordination. What advice would you have for State agencies that are embarking on the shift from historical claims data towards supporting provider portals and real time alerts?

DG: The types of data that are being used to support providers is key.  Claims, encounter and enrollment data collected by the states will always include some time lag in record submission and processing.  Typically it may take up to 6 months or more for claims submitted to be fully processed.  The monthly (or more frequent) provision of claims data includes in any receipt-month services that were provided in many prior months.

One of the reasons for incomplete data is delays in processing claims due to complex adjudication cycles. In some cases the processing leads to the same service being represented multiple times in the data stream.  In order to collect as current data as possible the users must be able to tolerate incomplete data and understand how to transform a claim into a unique record of a clinical service.  The challenge is to balance data timeliness with data completeness.  The applications of the data can be designed around these limitations.  In particular, algorithms that are based on risk identification over multiple months can be deployed.  An example of this would be identifying and flagging a patient with a severe mental illness who is chronically missing refills of an atypical anti-psychotic.   In contrast, using claims-based data to notify PCPs of emergency department visits is not practical except in states with highly evolved and inclusive Health Information Exchange systems.  Properly designed alert systems and portals can be deployed by states; however, they need to be purpose-built for realistic end applications and relevance to external data users.

ZU: Has JEN experienced issues in which privacy was impeded the implementation of new technology? How do HIPAA and HITECH regulations inform your strategy around designing new health IT systems?


DG: Designing systems that comply with HIPAA and HITECH is challenging.  We started to invest in rebuilding systems from the ground up to be compliant in 2005.  If the privacy model is part of the original design specifications then the level of difficulty in flexibly serving different user communities with different access privileges is not unreasonable.  In many respects the investment in a compliant privacy model has paid off since many of our clients prefer us to shoulder the burden of working with data sources to establish an approved data management plan and then tightly control any access to PHI.   Typically our clients prefer us to use our controls to prevent any users from building reports that are not publishable because of privacy concerns.  The JEN approach of supporting the user’s capacity to build their own data models based on highly granular data without then being able to see Protected Health Information (PHI) at any point  has proven to be very successful with CMS, state Medicaid agencies and our system users.  Because analysts can create charts, tables and even conduct statistical analyses without accessing PHI,  we have avoided the need to distribute sensitive data regardless of a requestor’s level of authorization.

ZU: In the next five years, what do you envision the most significant changes coming to Medicaid in the data analytics area?

DG: At JEN we would like to see an investment in system intelligence that goes beyond simple surveillance and reporting from the data.  Embedded in the data is a wealth of information on patterns of health services utilization that represents “optimal paths” for beneficiaries to maintain good health, recover from illness quickly and manage disability.  Since health care payments are driven by institutional care–acute hospital for Medicare and nursing homes for Medicaid–the goal of the system should be to maintain people in the community while promoting health.

JEN uses the term “Healthy Days at Home” as an umbrella for this goal.  While this idea encompasses beneficiary well-being it also leads to reduced state and federal payments since hospital days, skilled nursing facility days, and long term nursing facility days are minimized.  The challenge is to establish data models for this outcome and to analyze data to understand what care patterns work best.  For example populations with cognitive impairments are generally not ideal candidates for disease management and health promotion.  However there are providers who are very good at delivering excellent physical medicine care to patients with intellectual disabilities, dementia and other conditions.  An “optimal path” analysis for such high risk populations could entail the identification of service types, care trajectories and individual providers who represent best practices.  This type of micro-analysis that focuses on specific high need populations for intelligent and proactive care management is what we hope future data and systems can support.

Related Post

Leave a Comment

Name (required)
Your Message (required)

  • gal_img



  • gal_img


    View all Sessions in HD Video
    + Download PPTs