FDA Enforcement Part 3: What Happens When FDA Comes Back

TL;DR

When FDA came back to a compounding pharmacy, it usually cited the same kinds of problems as the first time.

For 87 pharmacies in the enforcement set, I compared the first Form 483 with the last one. When the first 483 cited a kind of problem, the last one cited it again 69% of the time. Other kinds of problems showed up on the last 483 only 34% of the time.

Even when FDA came back 3 years or more later, the last 483 cited the same kind of problem again 71% of the time. That matches the rate when FDA came back within 2 years.

A 483 does not end with the response. Someone still has to keep each correction working until FDA comes back.

Pick one observation from your pharmacy's last inspection or internal audit. Can you show what was corrected, who owns it now, and how you know it is still being done?

How Pharmacies Appear More Than Once

The 680 documents cover 322 pharmacies. 213 of them (66%) have more than one document.

A second document is often about the same inspection, because FDA issues a Form 483 at the end of an inspection and later sends the state board a hand-off letter, called an FMD-145, about it. 49 pharmacies in the set have both.

Counting inspections instead of documents, FDA came back to 92 pharmacies at least once.

Pharmacies With Two Inspections

When FDA came back to a pharmacy, it was about twice as likely to cite a kind of problem that was on the first Form 483.

92 pharmacies in the set have two or more dated Form 483s. For 87 of them, both the first and the last 483 have readable observations. For 49 of those 87, at least 3 years passed between the first 483 and the last.

As in Part 2, I tagged each 483 by theme. A theme is a kind of problem, such as cleaning or batch records.

Across all 18 themes, a theme on the first 483 was back on the last one 69% of the time (529 of 765 times). A theme that was not on the first 483 showed up on the last one 34% of the time (270 of 801 times).

A cited theme came back 71% of the time whether FDA returned within 2 years or after 3 years or more.

For each theme in the table, I split the 87 pharmacies into two groups by whether their first 483 cited it. Each bar shows the percent of pharmacies in that group whose last 483 cited the theme.

Insanitary conditions was on most last 483s whether or not the first 483 cited it. Quality unit and SOPs, batch records, and complaint handling were each about twice as likely to be on the last 483 if the first 483 had cited them.

Part 2 found these three themes rising fastest, and each one depends on a person doing a review on schedule. A monthly batch record review stops the first month nobody owns it.

The Pharmacies That Escalated

24 pharmacies in the set progressed to a Warning Letter or a state referral. 5 received Warning Letters and 22 were referred to a state board, with 3 pharmacies in both groups.

Only 4 of the 24 had two or more Form 483s in the set before they escalated. 13 had exactly one, and the other 7 have no 483 in the set dated before their escalation.

The escalation came a median of about 7 months after the last 483. The shortest gap was 4 months and the longest just over 3 years.

The response to the first 483, and the records behind it, may be a pharmacy's only chance to show FDA the fix.

24 pharmacies is a small group, and the public record does not say why each one escalated.

What a Repeat Finding Costs

A repeated observation becomes part of a pharmacy's compliance history, which FDA's own procedures name as a factor in deciding whether to issue a Warning Letter.

In the Michigan Medicine letter from Part 1, FDA compared the medical center's written response against the observations and asked for evidence that the corrections existed.

A second inspection asks a harder version of that question: is the correction still working three years later?

This is an important lesson: a repeated theme does not mean the pharmacy ignored the first 483. Some of these pharmacies almost certainly fixed the cited item and moved on.

Cleaning the hood the investigator cited is a fix. A cleaning schedule with an owner, a log, and a review step is a control, and only the control is still running when FDA comes back.

What to Do With This

To turn each past fix into a control, start with the last inspection or internal audit your pharmacy completed and list each observation.

For each one, answer the five questions from Part 1: what was expected, who owned it, when it occurred, what supports completion, and who reviewed it.

Then ask how you know the correction is still being done. The answer should be a recurring task with an owner, a due date, and a record, not a person's memory.

If it isn't documented, it didn't happen.

Next in the Series

Part 4 looks at what FDA says it evaluates in a Form 483 response, and what a response needs to contain to be judged on evidence.

Sources and Scope

• FDA Form 483 Frequently Asked Questions
• FDA Regulatory Procedures Manual, Chapter 4: Advisory Actions
• FDA Compounding: Information for States
• University of Michigan Warning Letter, September 28, 2017

Counts come from Pomio's parsed set of 680 FDA compounding enforcement documents (2016 through 2026) covering 322 pharmacies and outsourcing facilities. Names were matched as written on the documents, so a pharmacy with a renamed or relocated site may be counted twice. Themes were assigned by keyword matching; the theme comparison uses the first and last dated 483 for each pharmacy with readable observations on both. A repeated theme does not establish that a pharmacy failed to correct an earlier observation. This article is educational and is not legal advice.