This page is part of the Situational Awareness for Novel Epidemic Response (v1.0.0: STU 1) based on FHIR R4. This is the current published version. For a full list of available versions, see the Directory of published versions

Aggregatating Measure Reports

There are four different ways to consolidate a collection of MeasureReport resources referencing a common Measure into an aggregated MeasureReport:

1. Consolidating data within a facility from multiple reporting agents.
2. Over time, e.g., consolidating daily reports into a weekly report,
3. Over a larger geographic location, e.g., consolidating reports for all hospitals in a city or county, or
4. Consolidating data over both time and geography, e.g., consolidating all daily reports for all hospitals in a county for a week.

All MeasureReport resources being consolidated must reference a common Measure in MeasureReport.measure.

The process to perform this aggregation is the same for all of the above cases, and is defined in more detail in the Aggregate operation. That operation can be used by an Intermediary to consolidate MeasureReport resources into an aggregated MeasureReport for further processing.

The MeasureReport resource can be used to report on situational awareness not just for individual facilities, but also on geographic regions such as cities, towns, counties, states, provinces, territories, and other ad hoc geographic regions. They can also be used for reporting of non-geographically identified groups (e.g., all facilities managed by the same organization).

Group reports by aggregation region. This implementation guide does not specify how these groupings are created, but there are several mechanisms that might be used in an implementation:

1. Aggregate within a Facility to support consolidated reporting from multiple agents (e.g., information systems) within a facility for external reporting.
2. Aggregate by geolocation - When an implementation provides geographic coordinates, aggregates can be defined by polygons defining the set of geographic regions.
3. Aggregate by address - Some geographies can be aggregated by postal address. Postal addresses identify some geopolitical boundaries, including:
   • Cities,
   • Counties or parishes (known as the district in the FHIR Address data type),
   • States,
   • Country, and
   • Zip Code or Postal Code.
   • Provider or Hospital Service Areas or Referral Regions (see Dartmouth Atlas in Supporting Vocabulary).
4. Aggregate by a List - A list of Location resources can be created by enumerating every Location resource within the List resource.

While the most common use case is aggregation by geopolitical boundary (e.g., city, county, state), other useful geographic regions exist. Neighborhood or (in the US, a Census Tract) can define regions that are smaller than a city or county. Smaller geographic regions can support additional analysis, e.g., in regard to social determinants of health, or geographic regions addressing patterns of referral or service areas (e.g., hospital referral regions and service areas).

The Aggregate operation defines the algorithm for aggregating measure reports.

Aggregation of Population Counts

The aggregation of population.count requires special attention.

For a given facility and time period, measures can be aggregated in different ways depending on the scoring of the measure. The scoring of measures implies the way that which rate aggregation can be performed:

• Capacity
  Measures point in time capacity or utilization
• Event Growth
  Measures cumulative growth by counting current and cumulative events over time
• Queue Length
  Measures the number of activities awaiting completion at a point in time.
• Service Time
  Measures the aggregate total time to complete activities / total number of activities
• Availability Measures the point in time availability of resources

Measure population values can be aggregated across multiple measures to report across a wider group. The aggregation method is indicated in the rateAggregation field found in the MeasureGroupAttributes extension on the Measure.group element. These values come from the Measure Rate Aggregation Values value set. Each of these values describes the method of aggregation.

• count
  A count of events that happened, or changes in status, or of things consumed or performed over that period in time (e.g., admissions, deaths, tests performed). These are aggregated within and across subjects by summing all values for each MeasureReport.
• point-in-time
  Point in time measurements (e.g., bed occupancy, ventilators in use) represent the current state at a point in time. Group the MeasureReport resources by MeasureReport.subject, and aggregate within each group by taking the most receent point-in-time value. Then aggregate across groups by summing the most recent values.
• cumulative
  Cumulative measures are a combination of count and point-in-time measurements. These represent a count of the current quantity (e.g., tests performed) in the numerator over the total performed over “all time” (e.g., cumulative total tests performed) in the denominator. The numerator is aggregated as for a count. The denominator is aggregated as for point-in-time.

Across subjects (locations or facilities), counts are always summed. This assumes that subjects are non-overlapping (one is not trying to aggregate county data with the data for an entire state that contains the county).

Coding Geographic Locations

See Recommended Coding Systems and Value Sets for Location.identifer in Supporting Vocabulary for recommendations on how to identify geographic locations.