Reading provider outcome claims without inventing a conclusion

A success percentage is a compact claim with a large amount of missing context unless the measure is explained. Ask what was measured, in whom, at which time and through which method. Keep the actual source before comparing it with any other figure.

Behavioral Health Ratings does not produce outcome scores. This examination helps readers identify the information a qualified professional would need to assess a provider's claim. It does not determine the effectiveness of a service from a headline alone.

Define the outcome before reading the percentage

Ask what “success,” “improvement” or “completion” means in the source. Is the measure attendance, completion of a program, a reported change, a clinical assessment or another stated outcome? Record the definition in the provider's terms and ask whether it was decided before the data were collected.

Completion and later wellbeing are different measures. Satisfaction and symptom change are different again. They can all supply information, but they cannot be compared as though each represents the same result. If the provider has not defined the term, preserve the percentage as a claim awaiting a definition.

Find the denominator and the time point

Ask how many people the percentage describes and which people were eligible to be counted. Record when the outcome was measured: during a program, at discharge, or at a stated follow-up point. Identify whether the denominator includes all enrolled participants, completers or only people reached for follow-up.

Consider a fictional arithmetic example, unrelated to any provider: 24 favorable responses among 30 people reached equals 80% of respondents. If 50 people were eligible, that is not an established 80% outcome for all 50. The 20 people not reached have an unknown recorded result in this example. Do not label them successful or unsuccessful without data.

Ask how the information was collected

Identify the source of the measure: participant report, staff assessment, records review or another method. Ask who collected it and whether the question, instrument or definition is available. Record how incomplete responses or missing follow-up were handled.

NCCIH's health-research checklist highlights study size, participant relevance, duration and limitations. These are useful reading questions. Its checklist concerns health research, so do not imply that following it independently validates a provider's own unpublished data.

Separate provider data from general research

A source showing that a treatment approach has a research basis is not automatically evidence of this provider's own outcome rate. Ask whether the claim concerns the method in published research or the organization's measured results. Preserve the reference and the exact statement it supports.

NIMH's psychotherapy guidance recommends discussing goals and progress assessment with the treating professional. Bring the provider's definition and measurement method to that conversation. A research citation with a related title should not be treated as proof of a different population, setting or result.

Record what the evidence supports

Write a narrow result: “provider supplied its completion definition and denominator” or “follow-up population not yet described.” If a full report is available, identify the relevant section. If only a marketing summary is available, record that limitation as a source condition, not a verdict about the provider's clinical work.

Avoid comparing providers until the measures and populations are sufficiently aligned for the question being asked. The population examination helps identify whether a figure describes people and services relevant to the proposed care. A different measure may be informative without being comparable.

Printable outcome claim card

• Exact claim and dated source:
• Defined outcome and measurement instrument or question:
• Numerator, denominator and eligibility rule:
• Measurement or follow-up time point:
• Data collection method and responsible party:
• Missing-response treatment explained:
• Provider data or general research reference:
• Relevant population and service setting:
• Evidence actually supplied and remaining question:
• Qualified professional reviewing relevance:

Keep the original numbers intact. An honest evidence record can explain a result's scope without manufacturing either an endorsement or an adverse conclusion.