Identifying At-risk Populations – Case Manager Insights to Behind the Scenes


Case managers have been using electronic systems that identify clients for outreach for many years. Whether daily, weekly or monthly, a list of clients is presented for the case manager to initiate engagement. This is usually based on a combination of the company’s business needs, program construct, regulatory requirements and others.

Have you ever wondered how these algorithms are developed and who is working to develop them behind the scenes? How did the “system” make the determination to include the member? What are the risk factors driving inclusion in the program? These are all very valid questions to ask, and that should be asked.

This process is commonly known as identification and stratification (ID & Strat). Many times, the reason clients are identified is clear to the case manager, and outreach can begin promptly. Other times, there is a level of investigation that the case manager must undertake before they feel prepared to engage with their client. It’s not uncommon for case managers to question why a client is being identified for engagement.

Historically, the data available for use in ID & Strat has been claims, pharmacy, lab and enrollment data. More recently, clinical (EHR) data has been used in the process and can provide additional insights.

Care management programs can be targeted for clients with certain high-cost, high-prevalence conditions, such as diabetes. Proactive management of these clients can lead to better health outcomes and reduced costs. It is somewhat straightforward to identify a person with diabetes using the data mentioned above. Yet we know some clients can manage their condition well, and others are not well controlled and need assistance. This is where identification and stratification algorithms come into play.

The identification process tends to look at certain clinical indicators that signal higher severity. Some items that could be included in the process are:

  • Targeted conditions and corresponding severity
  • Number of concurrent conditions
  • Open care gaps
  • Number of inpatient and emergency department visits
  • If the client is engaged with a healthcare provider


Stratification typically looks at the entire population and ranks people in order of priority. Many factors go into stratification algorithms such as:

  • Demographics (age, gender, location)
  • Severity of the conditions
  • Co-morbid conditions
  • Historical healthcare costs and utilization
  • Pharmacy utilization (cost, specialty pharmacy, etc.)
  • Social determinant information (geography, education, health literacy, etc.)


Vendors who supply ID & Stratification software should possess a solid understanding of and have experience in:

  • Population health strategies
  • Regulatory requirements
  • Data represented in clinical practice, such as claims data, pharmacy data and lab data
  • How data can be used to create actionable insights/interventions
  • Health plan operations
  • National and regional clinical practice guidelines
  • Staff with clinical background


In conclusion, there is movement to make analytics solutions more transparent. Care management organizations should ask questions from their vendors, so there can be an understanding of the definitions and algorithms. Some vendors may not share all the specifics, but some level of insight should be provided. Case managers should continue to ask questions and understand why clients are being presented to them. Insights into vendor algorithms, even at a high level, can only help ensure positive engagements.

michael johnson

Michael Johnson, BSN, RN, HN-BC,has more than 35 years of extensive healthcare experience in clinical and corporate operations, as well as building programs in population health and analytics to support it. His experience and background within the healthcare industry consists of multiple functional areas including care management, quality assurance, accreditation support, workflow/systems integration and holistic healthcare. He has more than 15 years of experience in healthcare analytics. Michael is currently senior assistant vice president at EXL Health, where he oversees clinical informatics development.

Comments are closed.