Care Management Analytics

Advanced Analytics In Healthcare

Health insurers have several business imperatives heading into 2013, including preparations for Health Insurance Exchanges (HIX), ICD-10 preparations, 2014 Patient Protection and Affordable Care Act (ACA) requirements, value-based network development programs (e.g. ACOs, PCMH), care management and clinical decision support, B2C model enhancements, consolidation and perhaps global expansion.

All of these imperatives have dependencies on enterprise analytic capabilities, which can positively impact the resulting business performance post-implementation.

With increasing expectations and requirements for Payers, Providers and Pharmacy Benefit Managers (PBM) to optimize clinical outcomes and lower overall aggregate costs through more proactive care management and care coordination, Knowledgent believes there are growing needs to enhance clinical analytics.  Care management and care coordination can be increased by integrating claims-based and EMR-based data sources at the member/patient level.  Additionally, information from physician and care manager notes is now accessible by the newest analytical tools which can mine this text and semantically extract useful information to deepen analytics for predictive modeling, clinical decision support, care management program placement, and care planning

CARE MANAGEMENT ANALYTICS

Armed with information from claims, Personal Health Records (PHR) and Health Risk Assessments (HRA) — and potentially information from EMRs, care managers and devices — health insurers can play an active and increasingly personalized role to help engage patients in their care and well-being to maximize health outcomes and minimize aggregate overall costs.  Rich information regarding the network of healthcare providers also allows health insurers to help coordinate optimal care delivery.  With the increase in Accountable Care Organization (ACO) delivery models, health insurers can also play a role coordinating care more seamlessly with and amongst ACOs, Patient-Centered Medical Homes (PCMH) and traditional providers.  To do this effectively and efficiently, the care management/care coordination operations need timely and accurate information regarding which members to engage in these programs, which programs would most effectively serve their needs, which care managers would best handle a particular case, and which providers are leading which aspects of care delivery.

In order to create a complete member profile, an information management architecture is required that supports historical and real-time data, as well as both structured and unstructured information. Each piece of member data can provide a more complete picture, enable better decision-making and optimize member outreach.  Traditional data sources can also be enriched with third party data, member service interactions, emails, texts, clickstreams and social media to build a more complete and holistic view of the member for improved analytics.

PREDICTIVE MODELLING

To determine which members to engage in Care Management programs, health insurers have traditionally relied on statistical models to identify signals in claims history correlated with future high-cost claims.  More recently, health insurers have developed clinical models to identify patterns of care which diverge from the most current evidence-based medical practices.  With the latest technologies, health insurers can now also mine the textual notes from care managers and physicians and find additional signals which help calculate which conditions the member has, and the extent to which the member has those conditions.  Additionally, unstructured external “Big Data” sources can help identify which members with certain chronic conditions may be most responsive to care management programs.  For example, a diabetic may be found to have shown recent interest in fitness clubs, diet programs or healthy cooking.  By fine tuning and adding these additional capabilities to predictive models, health insurers and ACOs can increase the effectiveness and efficiency of care management programs.  Operations will identify more of the most impactable members, engage more of those individuals through more timely and relevant information, improve staffing and resource planning.

CARE MANAGEMENT PROGRAM PLACEMENT

With more timely and accurate information, health insurers and ACOs can also more readily determine which care management programs would be most effective for helping a member get optimal results:

  • Utilization Management
  • Complex Case Management
  • Disease Management
  • Behavioral Health
  • Wellness Programs
  • Lifestyle Coaching

Having personalized information about an individual’s longitudinal health history and persona, including rich data from unstructured notes and external “Big Data”, a health insurer or ACO can be most effective in coordinating outreach from the most appropriate care managers.  This improves the member’s experience, saves wasted effort on the part of care management operations, and produces optimized results.

There are a growing set of technologies well suited for delivering and servicing the data needs of any care management solution, including Customer Relationship Management, Master Data Management, and Entity Identification Management. These technologies would typically be supported by peripheral systems that absorb and distill both structured and unstructured data by means of ETL/ELT, traditional data warehouses and data marts, Enterprise Service Buses (ESB), SOA Web Services, as well as Big Data ecosystems to perform analysis utilizing Hadoop, NoSql databases, semantic analysis and cloud infrastructure. Also, these capabilities would be coupled with email, chat, voice channel systems, and semantic social data, to provide a truly holistic longitudinal health history for personalized member interaction. Optimizing the data management infrastructure helps leverage information as an asset.
Semantic analysis may be performed based on relationships and affiliations identified through Facebook, Google, LinkedIn and other party affiliated web sites for universities, employers or other membership organizations. Building domain ontologies of entities and relationships enriches the member profile. Advanced semantic analysis of unstructured data through natural language processing and data mining platforms help distill unstructured internal or public domain data into a smaller set of attributes and indicators predefined as relevant to enriching member data. Disparate modules should come together to create a holistic profile that is then open to iterative stewardship.

The basic notion that “the more you know about someone, the most productive the interaction will be”, holds true.  Knowledgent has found across all industries that quantity of data improves the quality of the analytical outcome and looking forward in Healthcare there is no greater requirement.

Never miss an update:

Subscribe to our newsletter!

Newsletter Sign Up Form

  • This field is for validation purposes and should be left unchanged.
 

TwitterFacebookYouTubeGoogle+Instagram
New York, NY • Warren, NJ • Boston, MA • Toronto, Canada
©2017 Knowledgent Group Inc. All rights reserved.