The number is off target. Now what?

A KPI turns red. Performance is below target. The dashboard did its job and got everyone's attention.

Now what?

This is the point where quality improvement can either become useful or become performative. If the immediate response is to assign an action plan before anyone understands the result, the organization may spend the next three months fixing the wrong problem.

First, make sure the result is trustworthy

Before interpreting an off-target measure, validate it.

Was the denominator complete? Did a workflow change during the reporting period? Are encounters missing? Did a new field or EHR configuration affect the extract? Were exclusions applied correctly? Is the current result comparable with the baseline or target?

If there is a material data-quality concern, label it. Do not present an uncertain number as though it were a confirmed performance problem.

Validation is not a delay tactic. It prevents teams from designing interventions around bad information.

Then look beyond the percentage

Suppose a follow-up rate drops from 78 percent to 64 percent. The percentage tells you what happened. It does not tell you why.

Look at the underlying volume and distribution. Did the denominator increase sharply? Is the decline concentrated at one site or during one month? Did staffing change? Are certain populations experiencing a different result? Is there a documentation issue, a scheduling issue, or an actual service-delivery problem?

Talk to the people closest to the workflow. Data can narrow the question, but staff often know where the process is breaking down.

Separate a one-time fluctuation from a pattern

Not every off-target month requires a formal improvement project.

Small denominators can move dramatically. Seasonal patterns can affect demand and staffing. A temporary vacancy may explain a short-lived change. That does not mean the result should be ignored, but the response should match the evidence.

Look at the trend. Compare the current period with prior periods. Determine whether the change is sustained, material, and operationally meaningful.

Choose an intervention that matches the cause

If the problem is incomplete documentation, staff education or EHR workflow changes may help. If the problem is appointment availability, documentation training will not solve it. If clients are difficult to reach after discharge, the team may need to examine contact information, outreach timing, communication preferences, or handoff processes.

This sounds obvious, but organizations often jump from a bad result to a familiar intervention because it is easy to implement.

Do not start with the intervention. Start with the cause you have enough evidence to test.

Define what improvement should look like

A quality-improvement activity needs a measurable expectation. What will change? By how much? Over what period? When will the team review the result again?

SAMHSA's CCBHC criteria require a data-driven CQI approach, including clearly defined improvement projects and evaluation of their effectiveness. The important word there is effectiveness. Completing an action item is not the same as improving performance.

If a team updates a workflow, trains staff, or introduces a new reminder process, the work is not finished when the intervention launches. The measure needs to be reviewed again to see whether the change actually helped.

Use a repeatable data-to-action process

For CCBHC quality review, I like a simple sequence:

1. Collect: Know where the data come from and who owns collection.
2. Validate: Confirm the result is complete and trustworthy.
3. Measure: Compare the result with the target, baseline, and trend.
4. Interpret: Investigate what appears to be driving the result.
5. Act: Test a specific response tied to the likely cause.
6. Monitor: Review the measure again and determine whether the response worked.

The process is intentionally simple. The discipline is in actually completing all six steps.

An off-target number is not the conclusion. It is the beginning of the evaluation question.