Every pharmaceutical site runs on a calendar that its quality system does not see, and in India the largest entry on that calendar is Diwali. Employees travel home for the festival, families gather, and many sites run reduced shifts or schedule planned shutdowns around the holiday week. The business plans for this well in advance, so production schedules, maintenance windows and dispatch cutoffs all move to accommodate it.
The deviation and CAPA system, in most cases, does not move with them. That is the gap this post addresses. A deviation carries a clock, because the procedure sets a target closure date that is often 30 days from discovery, and that clock keeps running through the festival even though the people who hold the evidence do not. The sections below describe what that does to open investigations and set out four steps to complete before the festival begins.
What happens to an open investigation during festival week
The evidence ages. The best time to interview the people who witnessed an event is the week it happens, and when the operator, the supervisor or the maintenance technician goes on leave, that interview simply waits. By the time the witness returns the details have softened and the account has shortened, and an investigation built on a shorter account tends to reach a shorter conclusion.
Ownership goes missing. When the investigation owner and the approver are both away, the record sits in a queue where nobody is looking at it. Nothing is wrong with the investigation itself, but nobody is available to move it to the next step.
Extension becomes the easy path. When the due date arrives and the owner is on leave, the simplest available action is to extend the record, and for any single investigation that decision is often reasonable. Across a festival period, however, a cluster of extensions changes what your on-time metric reports, because records that should appear as late are counted as on time once the clock has been reset. I wrote about why this matters in The Backlog Is the Most Diagnostic Number in a 483, where the point was that the overdue backlog tells the truth only when it is measured from the original due date.
Shutdown week produces its own events. A planned shutdown means maintenance, calibration, line clearance and restart, and restart is when problems tend to surface, often while the team is thin or has only just returned. The deviations from that week arrive in a batch at the moment investigation capacity is lowest.
Four steps before the festival starts
Step 1: List what is coming due, not only what is late. Pull every open investigation, CAPA and change control with a target date between now and two weeks after the festival ends. Most sites track overdue items closely, but the items that become overdue while people are on leave are the ones that get missed.
Step 2: Name a backup for every owner. A backup is a specific person who has access to the record, understands the investigation, and holds the authority to move it forward. A department does not qualify, and neither does a manager who is also on leave.
Step 3: Front-load the evidence. Complete the interviews now, and collect the records, photographs, data pulls and retained samples before anyone leaves. For any event that occurs during reduced shifts or a shutdown, capture the evidence on the day, including what happened, which equipment was involved, who was on shift, and what was different about the shift pattern, because the shift pattern itself may turn out to be part of the cause.
Step 4: Set the extension rule in advance. Decide now how many extensions a record may take, who approves them, and what the stated reason must contain. “Owner on leave” should trigger a handover rather than an extension, and on-time closure should be reported against the original due date so that the festival backlog does not disappear into the metric.
What to check when people return
Take the deviations opened in the three weeks around the festival and sort them by equipment and failure mode instead of by date or owner, then compare the result with a normal three-week period. Next, look at how the investigations from the festival period were closed. If the share closing as “human error, retrained” is higher than usual, the pattern is more likely to reflect the calendar than operator behavior.
I described the same sort as a test of inspection readiness in Audit-Ready vs. Inspection-Ready, and it works for a single festival window as well as for a full year.
Conclusion
Festivals are predictable, and so are the deviations that cluster around them. A site that plans production around Diwali without planning its investigations has planned only half of the problem, and the remaining half does not require new tools or additional headcount. It requires a list of what comes due, a named backup for each item, evidence collected early, and an extension rule that was agreed before anyone needed one.
Key Takeaways
- The clock does not observe the festival. Target closure dates keep running while the people who hold the evidence are away.
- Extensions hide the cost. Measure on-time closure from the original due date or the festival backlog will not show up in your metrics.
- Plan investigations the way you plan production. List what comes due, name a backup, and collect the evidence early.
- Check the conclusions afterward. A rise in “human error, retrained” closures around the festival points to the calendar, not the operators.
If you want to know how exposed your quality system is to calendar pressure like this, the Rapid Diagnostic is a fixed-scope review that can be completed in as little as two weeks, depending on the scope. Rapid Diagnostic