Measuring care coordination begins with a specific question about how care connects across people, teams or settings. Counting calls or referrals can describe workload, but it does not by itself show whether the patient received a coordinated plan.
AHRQ's Care Coordination Measures Atlas organizes measures around activities and perspectives. Its framework includes communication, responsibility, transitions, needs assessment and follow-up. It distinguishes the experiences of patients and families, healthcare professionals and health systems.
Define the question before choosing a number
Describe the process you want to understand. “Improve coordination” is too broad for a single measure. “Confirm whether referred patients obtained an appointment and whether the referring team received a response” is more specific.
Choose a population, setting and time period. Decide whether the unit being counted is a patient, referral, discharge or encounter. One person can have several referrals, so these counts should not be used interchangeably.
The Atlas is a measurement resource, with its update published in 2014. It should not be treated as a current list of mandatory reporting measures.
Specify numerator and denominator
Write down exactly what qualifies for inclusion and what counts as completion. Define exclusions, duplicate handling and the source of the data before reviewing results.
For a fictional local project, suppose eighty eligible referrals are due for review during a month and fifty have a documented response from the receiving service. The response rate is fifty divided by eighty, or 62.5%. This example is an invented calculation, not a published benchmark.
That rate does not establish that fifty patients attended appointments. It answers only the documented-response question under the stated definition.
Combine process, experience and outcome information
AHRQ's framework explains that the same interaction can look different from the patient, clinician and system perspectives. A chart may show a completed communication while the patient still cannot explain the next step.
Pair a process measure with an appropriate patient-experience question or record review. Use an established instrument according to its specifications when one fits the purpose.
An outcome such as hospital readmission can be relevant, but a change after a coordination project is not, on its own, proof that the project caused it. Examine population changes, other interventions and the study design before making a causal claim.
Examine missing information and unintended effects
Keep “unknown” distinct from “not completed.” Missing responses may reflect a documentation problem, an inaccessible service or another gap that needs investigation.
Consider whether improved results conceal additional workload or exclude people who are harder to reach. Review relevant differences across groups using appropriate privacy protections and sufficient data.
Use findings to select a specific change, then measure the same defined process again. Preserve the definitions and limitations so that a later comparison is meaningful.
For a measure of closed-loop referrals, specify whether the endpoint is acceptance, a completed visit or review of the result. For care plans across organizations, ask whether the receiving team can act on the plan. A measure of social-needs screening answers a different question again: identifying a need does not establish that it was met.
Sources
- Care Coordination Measurement Framework (accessed 2026-09-17) — AHRQ, Care Coordination Measures Atlas Update.
- Atlas Background, Purpose and Scope (accessed 2026-09-17) — AHRQ, 2014.