Skip to main content Skip to footer

Data Analytics

The Future of HHS: Navigating from Data to Action

Health and human services (HHS) agencies have unprecedented access to data sets, which are growing in number and size. The presumption is that more information is always better. However, that’s true only if organizations can utilize those data sets to take the “next best actions”: decisions that help agencies and case managers provide care at the lowest cost, with improved outcomes, and in a timely manner.

To do this, HHS agencies need to rethink how they collect, analyze and use their growing data collection. That information can include:

  • More than 3,000 structured datasets released as part of the U.S. Health Data Initiative.
  • Population health data from claims, electronic health records, lab reports, prescriptions and health reimbursement arrangements.
  • Data from sources that include censuses, surveys, vital records and customer relationship management systems.
  • Unstructured caseworker and care management notes.
  • Patient health information generated by connected devices.

The Organisation for Economic Co-operation & Development identifies data analytics as an important component for transitioning to a digital government and a way for agencies to gain better insights. A Center for Digital Government report found that the most popular digital tools were dashboards or portals (79% using), enterprise data warehouse (51%) and data mining (34%).

However, not all agencies have been able to realize the desired benefits. Only 14% of respondents in the CDG’s survey said their government’s analytics capabilities are effective. The rest rated the impact as somewhat effective or neutral.

While analytics continues to be an HHS priority, their efforts need to go beyond data mining and intelligence. The goal should be real-time decision-making that can define and deliver the right care interventions and strategies.

Data challenges

Generating and disseminating next best actions requires agencies to operate at the highest level of analytics maturity, not an easy task. The diagram below outlines the standard analytics maturity model, the type of insights that agencies can generate at each level, and the challenges that they face.

How: do you effectively manage the data life cycle, monitor data quality and meet objectives?

Challenge:Aggregation and integration of many types of data, which are in varying formats and sizes.

WHAT hidden relationships must agencies know to create a 360-degree client view?

Challenge: Inconsistent, inaccurate and overwhelming reporting due to lack of a common language to explain the WHAT [above].

WHY are things not happening as they should be? What are causing roadblocks?

Challenge: Inefficient techniques to research and discover previously unknown facts through proper data mining and statistical correlation.

WHO (or WHAT) will be affected most if today's issues are not addressed?

Challenge: Lack of skilled data specialists and inability to generate rapid insights that help organization see the future.

WHICH action should be taken, and WHEN, to mitigate future risk?

Challenge: Inability to quickly build AI models that can analyze insights and generate and recommend actions for stakeholders.

Most analytics projects that fail are generally the result of inflexible analytics systems and inefficient data management. Information about patients, physicians and other stakeholders is generally available in silos. It is difficult for agencies to semantically combine and harmonize different datasets to create an integrated longitudinal record. This lack of data standardization and interoperability turns data lakes into data swamps.

As a result, it is difficult or impossible to obtain real-time, reliable and accurate information to generate relevant insights and recommendations. This affects agencies’ ability to improve their analytics maturity, unlock the power of data and generate next best actions.

Beyond insights to action

Advanced data science, supported by technologies like automation and cloud computing, can help agencies re-imagine analytics initiatives. The diagram shows how a modular data and analytics solution, built using emerging technologies, enables an agency to address challenges at each stage of its analytics maturity and successfully generate next best actions.

Component: Data manager

Capability and insight: As-is integration of data irrespective of format and type. Automated master data management and data mastering.

Component: Data harmonizer

Capability and insight: Automated harmonization of data to create a complete profile of the client.

Component: Data analyzer

Capability and insight: Automated data profiling for in-depth understanding of root causes or uncovering hidden information.

Component: Data predictor

Capability and insight: Machine learning-based modeling to predict future states.

Component: Data insights generator

Capability and next best action: AI to analyze information from previous stages and recommend actions.

This process starts with the creation of a more flexible data store that includes clinical, financial, administrative, social determinant and weather information. Ideally, there will be no need for upfront extract, transform, load tools. And automated master data management will accelerate data harmonization and standardization.

This enables the creation of a dataset with a complete view of every patient, provider and constituent by combining their past with their present. This enables agencies to get a holistic understanding of patients and their relationships to providers, caregivers, friends, family, and social and demographic groups.

Data predictors analyze this information to understand how relationships have influenced patients over time and how they will influence them in the future. This creates a broad profile of each patient. By using artificial intelligence (AI) models, doctors, case managers or agencies can utilize this new information to take further action.

Navigating to better outcomes

The hypothetical case of Mark, a 78-year-old diabetic enrolled in a state Medicaid program, shows how data is used in this new system. Mark’s standard data includes his medical records, doctor visits, claims, care management notes, lab reports and prescription history. The agency also has Mark’s social determinants of health, such as his eating habits, transportation, caregiver support, care program engagement and other information. The data is cleanly aggregated to create a quality patient record.

Higher-level analysis finds that Mark is missing doctor’s appointments, not re-filling his prescriptions on time and isn’t engaging with his case manager. As a result, he visits the emergency room too often. There are no supermarkets near Mark’s home, and public transportation is inadequate. So, this restricts his access to fresh produce and leads him to eat out frequently. Also, Mark does not have a community support system.

A case manager can analyze these factors and develop a more detailed profile of Mark. Through predictive modeling, he will likely be placed in a higher risk category. An AI-based analysis will provide the case manager with the information needed to deliver targeted interventions. Those can include booking an appointment with a nutritionist, connecting Mark with food resources programs to help him manage his diet better, arranging transportation so that he doesn’t miss medical appointments and developing a consistent outreach plan to ensure Mark’s care gaps are tracked and closed.

Harnessing the full power of data

The next generation data science models will allow HHS agencies to unlock the full power of their data through a three-phased transformation approach [see figure below]. It will generate insights and recommendations that amplify and empower healthcare professionals, and ultimately, improve outcomes.

The approach also builds a foundation for HHS agencies to adopt other emerging technologies, including digital avatars, conversational AI, autonomic computing, augmented reality and virtual reality. The results can be improved constituent engagement, which enables stakeholders to become more productive and deliver successful outcomes while optimizing public funds.

Phase 1

Save to Invest

Process simplification: Patients are more easily managed with in-depth health intelligence.

Enhanced decision-making provider and care manager can predict high-risk cases and hidden factor and, then act.

Phase 2

Expand the experience

Intelligent routing patients at risk are better prioritized and routed to correct care support.

Control center Cross agency collaborations are more streamlined for prompt action.

Intelligent dashboard Providers and care manager receive real-time updates featuring next best action

Phase 3

Enable Business

End-to-End outcomes Care management productivity and health outcomeshave greatly improved.

The approach also builds a foundation for HHS agencies to adopt other emerging technologies, including digital avatars, conversational AI, autonomic computing, augmented reality and virtual reality. This drives improved constituent engagement, which enables stakeholders to become more productive and deliver successful outcomes while optimizing public funds.