Population Health Examples

Our population health reporting examples illustrate common analytics use cases that use the Bulk Data API. These examples span the development life cycle of a population health initiative, including: orient to a population and describe management opportunities; define an initiative and specify target populations; operationalize process change; and track progress using key indicators.

Describing management opportunities in a population generally follows a two-step process. Each step presents unique data enrichment requirements. First, we compute performance metrics that measure utilization (e.g. hospitalization or ED visit rates) or process compliance (e.g. eCQMs) that can be used to track progress and identify variance within subpopulations.

Second, we compute population features (e.g. enrollment, attribution, demographics, geography, chronic disease) to break down the population into subpopulations in order to help understand which subpopulations are driving the metrics. Variations between subpopulations can identify management opportunities as well as exemplars that can be mined for best practices.

Population health use cases leverage several common data patterns:

Assemble data from multiple data sources to get a complete and up to the moment view of utilization and clinical status. See also: Galileo Encounter Reconciliation.
Aggregate transactional data using value sets. For example, ICD-10 to CCSR and HCC. See also: Condition Profiles and Risk Score.
Reconcile data from multiple sources to identify a single set of features for each individual: provider attribution, program or contract enrollment, demographics, region.

Many of our population health use cases focus on ED visits and unplanned, emergent hospitalizations. Initiatives that can reduce these events can simultaneously lower population costs and improve outcomes.

Reporting Examples and Common Use Cases:

Example	Description
Utilization Metric Analysis	What is driving emergent utilization? Evaluate subpopulations defined by enrollment, attribution, geography, demographics, chronic disease or recent utilization to understand how subpopulations compare with respect to their contribution to an overall metric, rate, or change from a prior period. Example metrics include Hospitalizations, ACSC Hospitalizations, ED Visits, Avoidable ED Visits, and Repeat ED Visits.
Unplanned Hospitalization Analysis	What subpopulations are driving unplanned hospitalizations? CMS utilization measures for readmission (ACO8) and unplanned hospitalization with multiple chronic conditions (ACO38) are analyzed using enrollment, attribution, geography, demographics, and chronic disease. Visualize post-discharge activity such as SNF utilization and PCP followup and how this activity varies in subpopulations.
Analyzing Clinical Patterns	Visualize relationships between chronic conditions, medication therapies, lab results, procedures and acute utilization within subpopulations. Estimate the size of a subpopulation with specified clinical features and observe the rate of comorbidities, medication therapies, and acute utilization within the specified subpopulation.
Data Coverage	Does our data describe all events in a population? With what time lag? Analyze utilization events such as hospitalizations or ED visits by data source to determine how data sources overlap and what run-out period should be considered when using a particular data source.
Population Overview	Orient to a population by understanding how individuals and overall utilization is distributed by enrollment, attribution, geography, demographics, chronic disease or recent utilization.
Care Gap Opportunities	How are care gap opportunities distributed in a population? Overall rates of compliance with clinical best practices are computed using eCQMs and organized by enrollment, attribution, geography, demographics, and chronic disease. Trends highlight subpopulations with effective gap closure initiatives.
Crucial Events	How do crucial events affect utilization patterns? Align key events (e.g. hospitalizations, ED visits, care management initiation, chronic disease detection) on a relative timeline with respect to the event time in order to compare utilization before and after the event and determine the change in utilization that can be attributed directly to the event.
Wellness Opportunities	Describe opportunities to manage extant chronic disease that has not been recently treated or coded. These opportunities can be used to proactively manage lapsed chronic disease and minimize missing components of risk scores.
Risk Stratification	Describe the relationship between modifiable risk factors and prospective avoidable utilization. Historical simulation and AI models are used to prioritize the signals (risk factors such as emergent utilization, abnormal lab results, high risk medication therapies, etc.) that should be used to define a real-time target population.