Skip to main content

Outcomes and learning

How Should Peer Support Outcomes Be Tracked?

Peer support outcome tracking should begin with a clear learning or accountability question, then use defined measures, participant voice, appropriate baselines and follow-up, stable denominators, context, permissions, and traceable source data. Software should help people understand change without claiming that one service or platform caused every observed outcome.

By PeerakeetPeer support software researchReviewed and sourced 13 min read

What makes an outcome measure useful?

A useful measure answers a named question, matters to participants and the program, has a clear definition and time window, can be collected appropriately, and leads to a responsible decision or learning action.

Peer work is recovery-oriented and person-centered, so outcomes should not reduce progress to service volume or impose a single definition of recovery. The software should distinguish participant-reported change, staff-recorded activity, administrative data, and calculated measures.[1][2]

Which measurement elements must be defined?

Store the meaning and method with the value. A dashboard number without population, denominator, dates, missing-data handling, and source is not enough for a reliable decision.

Decision criteria and evidence
CriterionHow to evaluate it
QuestionName the decision, learning need, contractual requirement, or participant purpose the measure serves.
MeasureDefine construct, instrument or field, response options, calculation, owner, version, and interpretation limits.
PopulationDefine inclusion, exclusion, program, site, cohort, date window, and denominator before viewing results.
CollectionSpecify who asks, when, how, with what consent or notice, and what happens when a participant declines.
ContextKeep service exposure, baseline, follow-up, missingness, changes in population, and external factors visible.
UseLimit access, review interpretation, record decisions, monitor unintended effects, and retire measures that create burden without value.

How should a dashboard be validated?

Recalculate a sample from source records, test every filter, inspect missing and duplicate data, and confirm that labels describe what the measure actually supports.

Outcome dashboard validation
Workflow momentWhat good looks likeEvidence to request
NumeratorExactly which observations count and whyReconcile a sample to source records
DenominatorEligible population and exclusionsCompare displayed count with an independent roster
TimeBaseline, observation, service, and follow-up windowsTest boundary dates and late entries
Missing dataDeclined, not asked, not due, lost follow-up, invalid, and unknown stay distinctNo silent conversion to zero
ChangePaired, cohort, point-in-time, or aggregate method is labeledCalculation is reproducible
InterpretationAssociation, trend, target, and causal language are not confusedReviewer approves external claim

What should outcome software demonstrate?

Ask for one participant-level measure, one program measure, a follow-up window, a declined response, a corrected record, a definition change, and a funder export.

  1. Can the program preserve the exact measure version, response options, scorer, and effective dates?
  2. Can participants decline or skip appropriately without the system turning missing information into failure?
  3. Can the team separate reach, engagement, service delivery, experience, participant-reported outcome, and administrative result?
  4. Can every dashboard total be traced to source records and exported with definitions?
  5. Who reviews external outcome claims for evidence, context, privacy, and participant dignity?

Which outcomes should a program start with?

Start with a small balanced set that participants and staff understand, the program can collect consistently, and leaders will actually use for improvement or required accountability.

Avoid collecting sensitive information because a tool makes it easy. Measure reach and service integrity beside participant-reported progress and experience, disclose limitations, and do not imply causation from a simple before-and-after trend. Review burden and usefulness on a defined schedule.

Frequently asked questions

Are service counts outcomes?

Usually they are outputs or activity measures. They show reach or delivery, not necessarily participant change. A balanced report can include both while labeling them correctly.

Does improvement after peer support prove causation?

Not by itself. Participants and programs change for many reasons. Describe the design and limitations, and use causal language only when the evidence supports it.

Should every participant complete every measure?

No universal rule applies. Use the approved program design, consent or notice, accessibility, timing, and participant choice. Preserve declined, not due, and missing states accurately.

Can a proprietary score be exported?

Ask about measure ownership, licensing, calculation, interpretation, participant-level access, portability, and the right to use historical scores after termination before selecting it.

Sources and product pages

Government sources establish the legal and program requirements covered here. Official vendor pages document the product capabilities and positioning used in this guide.

  1. Core Competencies for Peer Workers in Behavioral Health Services: Substance Abuse and Mental Health Services Administration. Federal framework for recovery-oriented peer work, role clarity, and person-centered practice.
  2. Understand: Dashboards, Outcomes, and Reporting: Peerakeet. Peerakeet dashboards, outcome measures, exports, and funder reporting.
  3. Understanding Confidentiality of Substance Use Disorder Records: U.S. Department of Health and Human Services. Official federal overview of 42 CFR Part 2 applicability, consent, use, disclosure, and breach obligations.