Out of Trend (OOT)

What an Out of Trend result is

An Out of Trend (OOT) result is a test outcome that falls within specification limits but shows an unexpected or unfavourable pattern against the historical baseline. The result itself isn't a failure. The pattern is a signal that something is changing — process drift, equipment wear, raw material variability, environmental change, method shift, or an emerging quality issue.

OOT is distinct from Out of Specification. An OOS is a confirmed failure against an acceptance criterion. An OOT is within criteria but trending in a way that warrants attention. Both need investigation; OOT investigation is the proactive equivalent — addressing the signal before it becomes a failure.

OOT appears across stability studies, in-process controls, finished product testing, environmental monitoring, and analytical method performance. The discipline of detecting and investigating OOT is one of the clearest indicators of whether a quality program is genuinely proactive or only reacts to confirmed failures.

Why OOT detection is increasingly inspected

Modern inspections look beyond confirmed OOS failures into the trend data that preceded them. The pattern inspectors find repeatedly: stability data drifted for months, the analyst noted the drift, no action was taken because results stayed within spec, eventually a batch failed, and the inspector retroactively traced the OOT signal that was visible long before. \"Trends ignored\" is now a citation pattern in its own right.

MHRA's GxP DI guidance (March 2018, updated September 2021) and PIC/S PI 041-1 (July 2021) both treat trending as a data integrity expectation, not just a quality-process expectation. The reasoning: if a firm has the data to detect a trend but doesn't act on it, the data integrity controls have failed the firm's quality decisions even if the technical capture was fine.

Inspector perspective: inspectors typically ask for the trend reports the firm runs, the criteria used to flag OOT, and the investigations performed when OOT was flagged. If the trend reports don't exist, that's the finding. If they exist but no investigations were performed even where the data showed drift, that's a worse finding because it shows the system was looking but not acting.

Where OOT obligations come from

OOT isn't named specifically in most predicate rules the way OOS is in §211.192. The obligation comes from a combination of investigation requirements, trending expectations in data integrity guidance, and quality risk management:

• 21 CFR §211.192: investigation of any unexplained discrepancy. An OOT pattern that wasn't investigated and then became an OOS makes the firm liable under this section twice: once for the OOS, once for the missed earlier signal.
• 21 CFR §211.194: laboratory records. The data and the trending of the data are both required record types.
• 21 CFR §211.180(e): quality unit responsibility for evaluation of records. Periodic trend review is the operational form of this.
• FDA Guidance: Investigating OOS Test Results for Pharmaceutical Production (October 2006): the canonical OOS framework also references trending as part of robust laboratory practice.
• FDA Data Integrity Q&A (December 2018): addresses trending of audit trail and laboratory data as a data integrity expectation.
• MHRA "GxP" Data Integrity Definitions and Guidance for Industry (March 2018, updated September 2021): treats trending as a data integrity control; lack of trending is treated as a DI weakness.
• PIC/S PI 041-1 (July 2021): data integrity expectations including trend analysis and review.
• ICH Q9(R1) (effective 2023): quality risk management; trend signals feed risk evaluation.
• ICH Q10 §3.2.1: process performance and product quality monitoring; trending is the operational form.
• USP <1010> — Analytical Data Interpretation and Treatment: provides statistical context for OOT evaluation.
• EU GMP Chapter 6 §6.9: trending of quality control data is required for stability and process data.
• ISO 13485 §8.4: analysis of data; trending obligations for device quality systems.

How OOT detection and investigation actually run

The recurring stages of a defensible OOT program:

• Defined trending criteria. Before any OOT can be detected, the firm has defined what constitutes a trend — control limits, alert limits, action limits inside the specification, statistical thresholds (Shewhart, CUSUM, EWMA), or rule-based pattern detection (three consecutive points moving in one direction, six consecutive points on one side of the mean, and so on). Without defined criteria, OOT detection is improvised.
• Periodic trend analysis. Stability data, in-process data, environmental monitoring data, finished-product data all run through scheduled trend analysis — daily, weekly, or monthly per the data type and risk.
• OOT signal capture. When a trend crosses the defined alert or action criterion, the OOT is captured as a controlled event with timestamp, data set, the criterion that flagged it, and an initial assessment.
• Investigation. Scope depends on severity and source. A single OOT in stability data may warrant a lab-level investigation (method, instrument, sample handling). A recurring OOT pattern or one in a critical process parameter expands to manufacturing, raw materials, environment, and the wider quality system.
• Risk evaluation. Under ICH Q9(R1), the OOT is evaluated for impact on product quality, patient safety, and validated state. Outcome drives whether routine investigation closure is enough or whether CAPA, deviation, or batch impact assessment is triggered.
• Linkage to deviation or CAPA when warranted. Significant or recurring OOT routes into deviation investigation or CAPA for systemic correction. Single, non-systemic OOT may close with documented investigation only.
• Trending of the trends. Aggregate OOT activity is itself reviewed: counts by product, parameter, method, period. Increasing OOT activity in any area is a leading indicator that triggers wider review.
• Standing input to Management Review. OOT data, investigation closure rates, and recurring patterns are standing inputs to Management Review under ICH Q10 §3.2.4 and ISO 13485 §5.6.2.

What strong OOT programs share

OOT programs that hold up under inspection have a recognisable shape:

• Defined criteria, documented in SOP. Not improvised per analyst. The SOP names the statistical method, the alert and action limits, and the rules that trigger OOT capture.
• Risk-based criteria depth. Critical parameters (stability-indicating assays, sterility, potency) get tighter criteria. Lower-risk parameters get appropriate sensitivity.
• Automated detection where data volume permits. Manual periodic trending of high-volume data is error-prone. Automated detection with human review of flagged signals is more reliable.
• Investigation depth scaled to signal. A single isolated OOT may close after lab review. Recurring or critical OOT triggers a full multi-disciplinary investigation.
• Risk evaluation under ICH Q9(R1). Each significant OOT goes through documented risk evaluation against product quality and patient safety.
• Defined CAPA triggers. Documented criteria for when OOT requires CAPA: severity, recurrence pattern, predicate impact, regulatory commitment risk.
• Data integrity baseline. The data behind the trending is ALCOA+-compliant; otherwise the trending itself is unreliable.
• OOT aggregate review. Trending of the trends; counts and patterns reviewed periodically.
• Management Review input. OOT activity is a standing MR input, not pulled together at MR time.
• Linkage to stability program. Stability OOT signals route into stability program review, not handled separately.

How Complere supports OOT detection and investigation

An OOT signal is your early warning. Catching it depends first on the statistical trending your lab or process information system performs against the underlying data — that's where the numbers live, and Complere is honest about not being that system. What Complere does is what happens after: once your team has detected the signal, the investigation, risk evaluation, and follow-up run on the same controlled workflow as your deviations and OOS results.

When your analyst or process owner flags an OOT, the event becomes a regulated record in Complere. It carries the same disciplines as any other quality event: a named person, a clear action, a trustworthy timestamp, a permanent audit trail, and access scoped to who actually needs to act on it. The investigation moves through the lifecycle your team has configured — capture, classification, investigation, risk evaluation, conclusion, closure — with the right signatures at each stage and the meaning behind each signature preserved on the record.

Where the OOT shows a systemic pattern or warrants corrective action, the workflow routes it into CAPA. The investigation and the CAPA cross-reference each other, so your auditor can walk the line in either direction. Your evidence stays in place for the period your regulations require, in your own space, with the same separation as the rest of your regulated records.

You can see the activity on the OOTs themselves — counts by product, parameter, period; recurring patterns; how quickly you're closing them. Your Management Review pulls this directly under the standing input expectation regulators look for.

What stays with your team: the statistical method you trust (Shewhart, CUSUM, EWMA, rule-based pattern detection), the alert and action limits, the periodic trending cadence, the risk-based criteria depth, and the discipline of acting on the signals you see. Complere supports the investigation infrastructure; the trending discipline is your quality work.

Related laboratory and quality investigation terms

OOT sits in the same family as OOS, deviation, and the broader data integrity controls:

• Out of Specification (OOS) — the confirmed failure OOT detection is designed to prevent
• Deviation — recurring or significant OOT may route into deviation investigation
• Root Cause Analysis — applied inside OOT investigation
• Data Integrity — OOT detection depends on the integrity of the underlying trend data
• ALCOA+ — the principles applied to trend data evidence
• Audit Trail — every OOT capture and investigation action goes to the audit trail
• CAPA — significant or recurring OOT routes into CAPA
• Risk Management — ICH Q9(R1) risk evaluation applied to OOT impact
• Management Review — OOT activity is a standing MR input