Reproduce the number
A qualified analyst should be able to use the documented specification and source data to recreate the reported result. If the calculation lives only in someone's memory or an undocumented spreadsheet, the control environment is weak.
Reconcile competing reports
When two reports disagree, do not average them or choose the one that looks right. Compare populations, date logic, statuses, exclusions, source fields, refresh timing, and calculation versions until the difference is explained.
Test denominator reasonableness
Ask whether the eligible population makes operational sense. Large unexpected shifts often reveal date logic, enrollment status, duplicate, site, or eligibility problems.
Review missingness and impossible values
Missing fields, unexpected nulls, invalid dates, impossible ages, duplicate identifiers, and inconsistent categories can materially change a measure.
Trace the workflow upstream
A data-quality issue may originate in documentation, training, system configuration, interface behavior, or unclear responsibility. Fixing only the report can leave the source problem intact.
Document changes
Measure logic, fields, workflows, and systems change. Keep effective dates and version history so historical results can be interpreted correctly.
Create escalation rules
Define which issues can be corrected routinely and which require clinical, operational, IT, compliance, or leadership review. Data governance prevents unresolved ambiguity from becoming permanent.
CCBHC requirements can vary by program, state, grant, payer, measurement year, and measure steward. Use current SAMHSA, CMS, state, grant, and technical specifications for official reporting decisions.