Pharma Shortage Signal Report — Vol. 1 (Q3 2026)

Every shortage you read about in the news was visible in data weeks or months earlier. This report is our first public look at what that data actually says. It is built on the public shortage notifications in the German BfArM database, the most transparent window into the European pharma supply situation that currently exists.

We plan to publish this report quarterly. Not as a list of who is short of what, the regulators already do that. The interesting layer is the pattern: how shortages start, how late they are reported, and what that reveals about the planning processes behind them.

How we read the data

We track first-time shortage notifications and two signal types in particular. Cluster signals: several ingredients with unrelated owners and supply chains showing first-time notifications in the same short window, which points to systemic causes rather than single-vendor problems. Late-report signals: a notification that arrives months after the shortage demonstrably started, which says something about how early the affected organization itself saw it coming, because a company that discovers its own supply gap at the same moment as the regulator has a planning problem, not a bad luck problem.

Signals that stood out this quarter

A cluster of first-time notifications in a single week.

Amitriptylin, escitalopram, and pantoprazol, three high-volume, long-established ingredients with mature supply chains, entered shortage notification within days of each other in July. Ingredients this generic and this consumed do not fail together by coincidence. The pattern is consistent with shared upstream factors: API sourcing concentration, or capacity being reallocated toward higher-margin products. When three mature molecules wobble in the same week, the cause is usually structural, not operational bad luck.

A shortage reported four months after it started.

Temazepam notifications identified a shortage that had been running since mid-March. Whatever the operational cause was, it existed through an entire quarter of planning cycles. A gap that size is not a forecasting miss. It indicates either no routine supply-risk review for the product, or a review that had no escalation path attached to it.

Both products of a therapeutic line affected simultaneously.

Levothyroxin signals covered the product line at the same time rather than a single presentation. For patients and prescribers, that removes the usual workaround of switching within the therapy. For the manufacturer, it removes the internal workaround too: if one site or one supplier feeds the whole line, there is no slack anywhere in the network to absorb the disruption.

Two further KKH-relevant first reports in early July (prednicarbat, propranolol hydrochloride) rounded out the picture.

What we take from Vol. 1

Three conclusions we will watch over coming quarters:

• 1. Reporting lag is the real story. The gap between a shortage starting and being reported is a proxy for how long the affected organization ran without a decision. That gap is where working capital burns and patients switch therapies.
• 2. Clusters deserve attention as clusters. Single-ingredient shortage response is firefighting. Cluster response requires looking at the category: which suppliers, which sites, which API sources do the affected products share.
• 3. Therapy-line coverage is a network design question. When one failure mode takes out an entire product line, the network has single points of failure by design, and design questions are answerable in advance.

The methodology behind this report, the signal taxonomy and the trigger thresholds we use to read planning quality from public data, comes from the same approach I bring to a client's own supply chain: first a process landscape assessment of the current S&OP/IBP setup, then targeted work on supply planning and network resilience.

This report uses only publicly available regulatory data. It names active ingredients, never companies. Ingredient data as notified; signals reflect our analytical reading, not regulatory findings.