Closing The CAPA Is Not Closing The Loop: Why Completing Corrective Actions Is Not Enough To Show Effectiveness—and What FDA Expects To See In The Record

With the Quality Management System Regulation (QMSR) now in effect, FDA scrutiny of corrective and preventive action is on the rise—both during inspections and when the agency reviews related remediation. One recurring issue is closing a CAPA as soon as the listed tasks are complete, without asking whether the underlying problem was reduced or eliminated. A CAPA file may contain a revised procedure, a training record, or a note such as “SOP revised, effective 3/15.” Those records show that an action was implemented. They do not show that it worked. If the original problem later returns, an investigator will look closely at the effectiveness assessment and ask a practical question: Did the manufacturer define, in advance, how success would be measured? This article draws the line between completing a CAPA and proving that it worked. We explain what FDA investigators look for, the judgment involved in designing a defensible verification of effectiveness, and how complaint and MDR trending fit into the record. Those data matter—but they should confirm the primary effectiveness measures, not serve as the sole basis for closing the CAPA.

This is the third article in our new thought leadership series, “Making Medicine: Legal insights at the cutting edge of Life Sciences Manufacturing,” which will provide novel analyses addressing the array of legal and regulatory issues affecting companies operating at the intersection of the manufacturing and life sciences & health care sectors.

Completion and effectiveness are separate inquiries

When reviewing a CAPA, we start with two questions that companies should keep separate in the record:

• Verification of implementation asks whether the manufacturer completed the actions identified in the CAPA plan. The evidence may include an approved procedure revision, completed training, installation records, supplier correspondence, or other documentation showing that the planned work was completed.
• Verification of effectiveness (VoE) asks whether those actions addressed the identified root cause, reduced the chance of recurrence, and avoided an adverse effect on the device or quality system. The answer requires objective evidence collected under a predefined plan and measured against stated criteria.

A manufacturer can have complete implementation evidence and still lack evidence of effectiveness. Companies should preserve that distinction in the CAPA record—including any follow-up needed before closure.

The regulatory requirement continues under the QMSR

Under the former Quality System Regulation, 21 CFR 820.100(a)(4) required manufacturers to verify or validate corrective and preventive action to ensure that the action was effective and did not adversely affect the finished device. The QMSR, effective February 2, 2026, incorporates ISO 13485:2016 by reference. ISO 13485 clauses 8.5.2 and 8.5.3 address corrective action and preventive action, respectively, and require review of the effectiveness of action taken, together with verification that the action does not adversely affect the ability to meet applicable regulatory requirements or the safety and performance of the device.

The framework still requires evidence that corrective and preventive actions work. Under the QMSR, the assessment should match the risk presented by the failure mode and stay connected to the manufacturer's risk-management and production processes. A conclusory statement that a CAPA was effective carries little weight if the record does not show how the company reached that conclusion.

CAPA remains a recurring focus during FDA inspections—and when the agency reviews a company's inspection response to decide the inspection classification or whether further enforcement is warranted. The question is increasingly pointed: How does FDA know that the company's commitments and corrective actions will work? An inspection response, revised procedures, and completed training may show what the company intends to do or has implemented. FDA may also expect objective evidence that the new process is working in practice. The response should identify the planned VoE and, where available, provide examples that show the revised process is producing reliable results—to name a few, properly investigated and closed CAPAs, well-documented complaint files, effective supplier corrective actions, accurately evaluated and timely submitted MDRs, or other records showing sustained execution. A response limited to procedure revisions, training records, or promises of future action may not support a favorable classification or persuade FDA that additional enforcement is unnecessary, even when the underlying corrective action is reasonable. Companies should make clear both how effectiveness will be evaluated and what evidence shows that the corrective action is working.

Recurring problems in VoE records

The following deficiencies show up regularly in CAPA files we review:

• A VoE entry simply says “training completed.” A training record shows that instruction occurred; it does not show that the relevant error stopped occurring during the work performed.
• VoE is documented as “SOP revised” or “document updated.” Those entries show document control and implementation—but they do not show that the revised process produced the intended result.
• Acceptance criteria are written only after the data are available and appear tailored to the observed result. Investigators may ask whether the manufacturer had a real decision rule—or simply documented a favorable outcome.
• The selected metric has no logical connection to the identified root cause. In many CAPA records, complaint review or complaint trending is the only stated effectiveness measure. A downstream complaint count, however, may not show whether the process condition that caused the failure remains controlled. Standing alone, it does not ordinarily provide a sufficiently direct or timely basis for concluding that the corrective action was effective.
• The VoE period is too short to detect the failure mode with reasonable confidence—for example, monitoring one production lot when the defect historically occurs once in fifty lots.
• The record states “no recurrence” without identifying the denominator, the monitoring method, or why the observation period was sufficient.
• The company closes the CAPA when implementation is complete, without performing or scheduling an effectiveness assessment.

What a regulator expects to see in a VoE

When an investigator or compliance officer reviews a CAPA, the record should support a reasoned review—from the problem statement and root-cause analysis through the selected action and the evidence supporting closure. In our experience, the following features make that review easier and make the effectiveness conclusion more defensible:

• Predefined, objective acceptance criteria. State the criteria in the CAPA plan before implementation. They should be measurable and specific enough that a later reviewer can determine whether the result passed—without relying on judgment supplied after the fact.
• A direct connection from root cause to action to metric. Measure the condition at which the root cause operates. If the root cause is an uncontrolled process parameter, the primary effectiveness measure should address that parameter or the immediate process outcome—not only a downstream field event.
• A justified sample size and monitoring period. The CAPA plan should explain why the number of units, lots, transactions, or events observed is adequate—and why the monitoring period can detect the failure mode. The rationale should reflect the risk, production volume, lot frequency, and known event rate.
• Baseline data for comparison. Identify the relevant pre-action performance and explain how the post-action result will be compared with it. Without a credible baseline, a statement that performance improved is difficult to evaluate.
• Evidence that the action did not create a new risk. Update the applicable risk analysis and perform design or process verification or validation when the nature of the change warrants it. The record should address effects on the device, manufacturing process, and applicable regulatory requirements.
• A defined response to a failed criterion. The CAPA plan should state what will happen if the criteria are not met—including reopening the CAPA, reassessing the root cause, extending monitoring, or taking additional action as appropriate. The result and the response should be visible in the CAPA record and, where applicable, management review.

Complaint and MDR data cannot carry the VoE alone

Complaint records and Medical Device Reports (MDRs) are important sources of information about field performance. They are also imperfect primary measures of CAPA effectiveness. Field data may reflect devices manufactured before the corrective action, may arrive long after the relevant event, and may not show whether the root cause remains present in the manufacturing or design process. Companies should keep the following practical limitations in mind:

• Timing. Complaint and MDR data generally arrive after the device has been manufactured, distributed, and used. They can provide useful confirmation over time—but they are not a timely measure of whether a newly implemented process control is operating as intended.
• Underreporting. Some device problems are never reported by patients or clinicians. MDR reporting also covers a defined subset of events, including deaths, serious injuries, and certain malfunctions. The absence of a complaint does not establish that the failure mode is absent.
• Small numbers. For an infrequent failure mode, a short period without a complaint may reflect limited exposure rather than improvement. The manufacturer should explain why the observation period and event count support the conclusion reached.
• Denominator and confounding. The installed base may include both pre-action and post-action units, with differences in use, exposure, complaint handling, seasonality, or market distribution. Unless the data are segmented by build date, lot, serial number, or another appropriate characteristic, the trend may not answer the CAPA question.
• Distance from the root cause. Field data show that a user experienced or reported a problem. They may not show whether the specific process condition, design feature, supplier characteristic, or human-performance issue identified in the CAPA has been controlled.
• Reliance on the user to detect the failure. A primary VoE measure should provide earlier and more direct information where reasonably available. Waiting for a patient or clinician to encounter and report the problem leaves the quality system dependent on the very failure the corrective action was intended to prevent.

How field data should be used

Complaint and MDR analysis remains valuable. We advise clients to use it as confirmatory evidence, with the method of analysis defined in the VoE plan. The field trend should be consistent with the primary process, design, supplier, or performance measure. If it is not, investigate further.

• Confirm the primary measure. If in-process data, design verification, supplier data, or task-performance data show that the root cause has been controlled, review the trend for the same failure mode over an appropriate period. Where possible, normalize the trend to units distributed and segment it by manufacturing date, lot, or serial number after the corrective action effective date.
• Investigate disagreement. If complaints for the failure mode do not improve even though the internal metric appears acceptable, investigate the discrepancy. It may indicate an incomplete root-cause analysis, an action that did not address the mechanism of failure, or a metric that is not measuring the relevant condition. Reopening the CAPA may be appropriate.
• Continue longer-term monitoring. Field data may serve as a secondary VoE criterion with a defined monitoring period and decision rule. The primary internal measure should address the root cause directly; field data can then provide longer-term confirmation that the corrective action is producing the intended result in use.

Practical checklist for VoE

We use the following questions when reviewing a CAPA plan or closure package. Companies should support the answers with the record, with the level of detail proportionate to the risk and failure mode:

• The CAPA plan states the VoE criteria before implementation begins.
• Each effectiveness metric is tied to the identified root cause and the action taken.
• Relevant pre-action baseline data are documented for comparison.
• The sample size, statistical rationale where applicable, and monitoring period are defined and supported.
• The pass/fail criteria are objective, measurable, and stated before the results are reviewed.
• Implementation verification is documented separately from effectiveness verification.
• A leading internal metric is the primary VoE measure; complaint and MDR data are confirmatory where appropriate.
• Field data are segmented by build date, lot, serial number, or another characteristic that isolates post-action units.
• The plan states what will occur if a criterion is not met, including whether the CAPA will be reopened or the root cause reassessed.
• The results and any required response are escalated through management review.
• The risk file reflects the corrective action and any residual or newly identified risks.
• The CAPA record permits an investigator to follow the problem, root cause, action, evidence, and closure decision.

Conclusion: Build the evidence into the CAPA record

An FDA investigator reviewing a CAPA file will look for a coherent record: a defined problem, a root-cause analysis supported by evidence, an action that addresses that cause, and objective results collected under a planned method. The record should also address whether the action affected the device, the process, or applicable regulatory requirements. An investigator should not have to reconstruct that reasoning from scattered attachments or assumptions.

The distinction between completing an action and demonstrating effectiveness is straightforward—but it requires attention when the CAPA is planned. Define the success criteria before implementation. Select a measure that addresses the root cause. Provide enough data and time to evaluate the failure mode. Use complaint and MDR trends as supporting evidence where they add value.

In our experience, a well-supported CAPA closure does not depend on a lengthy explanation after an inspection begins. The explanation should already be in the plan, the data, the analysis, and the decision to close or continue the CAPA. In sum, medical device manufacturers should stay apprised of FDA expectations and be mindful of the evidence needed to show that corrective actions are working. That is the record an investigator can evaluate—and the standard companies should aim to meet.