Escalation Engine

It is 9:15 AM on Thursday at the Rixensart vaccine manufacturing campus. During the TIER 2 weekly review, the Production Director notices that a critical action — "Replace non-OEM vacuum pump gasket on Lyo Chamber 4" — has been sitting at TIER 1 for six days, three days past its due date. But she did not discover this by accident. ProBeya's escalation engine automatically promoted the action to TIER 2 on Tuesday, notified her team, and placed it on today's review agenda. The action is now visible at the right level of the organization, with the right people, at the right time.

This is management by exception in practice. The system does the watching. The leaders do the deciding.

The TIER Hierarchy: Why It Exists

The tiered management system originated in the Toyota Production System and was adapted for pharma manufacturing by companies like Danaher, Johnson & Johnson, and Novartis. The principle is simple: decisions should be made at the lowest competent level. When that level cannot resolve an issue — due to authority, resources, or cross-functional complexity — the issue escalates upward.

TIER 3  ──  Site / Executive Level
  |         Monthly steering committee
  |         Capital decisions, regulatory strategy, cross-site coordination
  |
TIER 2  ──  Department / Value Stream Level
  |         Weekly management review
  |         Resource allocation, cross-team dependencies, budget decisions
  |
TIER 1  ──  Shop Floor / Operational Team Level
            Daily standup (5-15 minutes)
            Frontline problem solving, shift-level issues

ICH Q10 Section 2.3 emphasizes "management responsibility" and "management review" as pillars of the pharmaceutical quality system. The TIER hierarchy operationalizes this: TIER 1 catches the problem, TIER 2 allocates resources, and TIER 3 makes strategic decisions. The escalation engine is the connective tissue that ensures issues flow between these levels without getting lost.

For Site Directors

The escalation engine is your early warning system. If you are consistently surprised by problems at TIER 3 that should have been resolved at TIER 1 or TIER 2, the issue is not the escalation engine — it is the escalation rules. Tighten the trigger days, enable auto-escalation, and review the escalation dashboard monthly.

Configuring Escalation Rules

Escalation rules define when and how actions move between tiers. Rules are configured per board, allowing different SQCDP domains to have different escalation policies. A safety issue should escalate faster than a cost optimization action.

Setting Up Rules

1. Navigate to the board settings.
2. Open the Escalation Rules tab.
3. Click + Add Rule.

Each rule has the following parameters:

Parameter	Description	Example
From Tier	Source tier level (1 or 2)	1
To Tier	Destination tier (must be greater than From Tier)	2
Trigger Days	Days past due date before escalation fires	3
Auto-Escalate	Whether escalation happens automatically or requires manual approval	true
Notify Users	Users notified when this rule triggers	Department manager, shift lead
Active	Whether this rule is currently enabled	true

Recommended Configuration for Pharma Sites

Here is the escalation policy used at several Top 10 pharma manufacturing sites:

Rule	From	To	Trigger Days	Auto	Rationale
Safety actions	TIER 1	TIER 2	1 day	Yes	Safety cannot wait — immediate visibility
Quality actions	TIER 1	TIER 2	2 days	Yes	OOS/deviations have regulatory clock ticking
Delivery/Cost actions	TIER 1	TIER 2	5 days	Yes	Operational but lower urgency
All actions	TIER 2	TIER 3	5 days	No	Executive escalation requires human judgment

The key insight: auto-escalation for TIER 1 to TIER 2 (volume is high, speed matters) and manual escalation for TIER 2 to TIER 3 (executive attention is scarce, context matters).

Rule Evaluation Logic

The escalation engine evaluates rules on a regular schedule (configurable, recommended: every 15 minutes):

1. Find all actions where status is not done or verified and dueDate < now
2. Calculate overdue_days = today - dueDate
3. For each action, check if an active rule matches its current tierLevel and if overdue_days >= triggerDays
4. If autoEscalate = true, automatically escalate the action
5. If autoEscalate = false, flag the action for manual escalation and notify the configured users

Disabling Rules

Rules can be temporarily disabled without deletion by toggling the Active switch. This is useful during:

• Plant shutdowns or holiday periods
• Organizational transitions (new department structure)
• Planned maintenance windows where overdue actions are expected

Auto-Escalation: What Happens

When auto-escalation triggers, ProBeya executes a precise sequence:

1. The action's tierLevel is updated to the destination tier
2. The escalatedAt timestamp is recorded
3. The escalatedFrom field captures the escalation chain for traceability
4. Notifications are sent to all users in the rule's notify list
5. An entry is added to the action's activity timeline: "Auto-escalated from TIER 1 to TIER 2 (3 days overdue)"

The original assignee remains unchanged. The higher-tier manager receiving the notification decides whether to reassign, add resources, or remove blockers.

Audit Trail Integrity

Every auto-escalation event is logged with the actor as system, the source as auto, and the full escalation chain. This creates an immutable record showing that the organization's escalation policy was executed as configured — a powerful evidence point during GMP inspections when auditors ask "How do you ensure overdue CAPAs receive management attention?"

Manual Escalation: When Humans Override the Clock

Sometimes the clock is not the right trigger. A shift supervisor may recognize that an action needs TIER 2 support before it becomes overdue — perhaps because a cross-departmental dependency was discovered, or because a safety observation requires immediate executive visibility.

1. Open the action detail panel
2. Click the Escalate button (upward arrow icon)
3. Select the target tier level
4. Add an escalation note explaining why
5. Confirm the escalation

Manual escalation follows the same process as auto-escalation but is tagged as source = "manual" in the escalation history. The note field is critical: it provides context that the receiving manager needs to make a decision.

Decision Framework: When to Manually Escalate

Situation	Escalate?	Rationale
Action blocked by another department	Yes	Only TIER 2 can coordinate cross-department resources
Safety concern requiring capital investment	Yes	TIER 3 budget authority needed
Action on track but complex	No	Let TIER 1 work the problem — escalation is not delegation
Repeated recurrence of a resolved action	Yes	Pattern indicates systemic issue beyond TIER 1 scope
Action owner on leave, no backup assigned	Yes	TIER 2 can reassign and reset the timeline

De-Escalation: Closing the Loop

Once a higher-tier manager has made a decision — approved budget, allocated headcount, removed a cross-departmental blocker — the action returns to the appropriate level for execution.

1. Open the action detail panel
2. Click the De-escalate button (downward arrow icon)
3. Select the target tier level (must be lower than current)
4. Add a de-escalation note describing the resolution or decision
5. Confirm the de-escalation

warning

De-escalation does not reset the due date. If the action is still overdue after de-escalation, it may re-trigger escalation based on active rules. Update the due date to reflect the new timeline agreed at the higher tier.

Escalation Dashboard

The Escalation Dashboard provides the consolidated view that TIER 2 and TIER 3 leaders need during their review meetings.

Active Escalations Table

Shows all actions escalated at least once that are not yet resolved:

Column	Description
Action	Title and linked board
Original Tier	Where the action started
Current Tier	Where it sits now
Days Overdue	Calendar days past due date
Escalated On	When the last escalation occurred
Assignee	Current responsible person

Escalation Funnel

A visual summary showing action distribution across tiers:

TIER 1:  42 actions (35 on track, 7 overdue)
TIER 2:   8 actions (5 on track, 3 overdue)
TIER 3:   2 actions (1 on track, 1 overdue)

In a healthy organization, the funnel narrows sharply. If TIER 2 and TIER 3 are holding a disproportionate number of actions, it signals either that TIER 1 teams lack resources or that the organization is over-escalating.

Trend Chart

A time-series chart showing escalation volume per week/month, broken down by direction (escalation vs. de-escalation) and source (auto vs. manual). A rising trend in auto-escalations is a leading indicator of operational stress.

For OpEx Coordinators

Track the ratio of auto-escalations to manual escalations. A healthy ratio is roughly 60/40. If auto-escalations dominate, teams may be ignoring due dates. If manual escalations dominate, due dates may be set unrealistically and the auto-escalation rules need recalibration.

Escalation History Timeline

Every action maintains a complete chain of custody:

Mar 15, 09:00  Created at TIER 1 by Alice (source: tier_meeting)
Mar 15, 09:05  Assigned to Bob
Mar 18, 10:00  Auto-escalated to TIER 2 (3 days overdue)
               Notified: Carol (dept manager), Dave (shift lead)
Mar 19, 14:30  Reassigned to Eve by Carol
Mar 21, 11:00  De-escalated to TIER 1 by Carol
               Note: "Budget approved, proceed with OEM gasket order"
Mar 23, 16:00  Status changed to Done by Eve
Mar 24, 09:00  Verified by Carol

This timeline satisfies the "complete, consistent, accurate" data requirements of EU GMP Annex 11 and provides the evidence chain that auditors expect when reviewing CAPA effectiveness.

Cron Job Setup

The auto-escalation engine runs via a protected HTTP cron endpoint.

Endpoint: POST /api/cron/escalation

Authentication: Requires the x-cron-secret header matching the CRON_SECRET environment variable.

Recommended schedule: Every 15 minutes (*/15 * * * *)

# Example: crontab entry
*/15 * * * * curl -s -X POST https://your-instance.probeya.com/api/cron/escalation \
  -H "x-cron-secret: $CRON_SECRET" \
  -H "Content-Type: application/json"

Each run:

1. Fetches all active escalation rules across all organizations
2. Queries overdue actions matching fromTier + triggerDays
3. Auto-escalates matching actions to toTier (if autoEscalate = true)
4. Sends 24-hour pre-warning notifications to action owners approaching threshold
5. Creates notifications for users in the rule's notifyUserIds list
6. Logs all escalation events to the activity audit trail

attenzione

Never expose the CRON_SECRET in client-side code or public URLs. The endpoint returns 401 if the secret is missing or incorrect.

Permissions

Action	Required Role
View escalation dashboard	Any board member
View escalation history on actions	Any board member
Configure escalation rules	Department manager or above (escalate permission)
Auto-escalation (system)	Automatic — no user role required
Manual escalation	Board admin or workspace admin
Manual de-escalation	Board admin or workspace admin

Related Features

• Action Log — The actions that flow through the TIER hierarchy
• KPI Boards — Red KPIs trigger actions that may later escalate
• TIER Meeting Mode — Review escalated items during meeting agendas
• Problem Solving — Investigate root causes of escalated issues
• Templates — Escalation rules are included in board templates