Lesson 08 — The 30-percent discount is not a deal

Swiss generics list at 30–50% below brand by ordinance, not market force

Run a price-comparison signal on a hospital’s drug catalogue and a clean pattern emerges: across most generic-brand drug pairs, the generic sits 30–50% below the brand. The first read is “good procurement discipline — the hospital is favouring generics”. The second read is “these spreads must indicate variability worth investigating”. Both reads are partial.

The actual reason is regulatory: Swiss pharmaceutical pricing is heavily governed, and generic discounts are set by BAG ordinance, not negotiated. The discount is a regulatory floor, not a vendor concession.

Why the 30–50% range is what it is

The BAG (Bundesamt für Gesundheit) sets generic pricing under the SL system (Spezialitätenliste). The discount tier depends on the molecule’s market size and the patent timeline:

• Higher-volume molecules → larger required generic discount (toward the 50% end).
• Lower-volume molecules → smaller required discount (toward the 30% end).
• The discount is set as a percentage of the brand’s regulated price.

The hospital pays the regulated generic price. The catalogue records that price. The brand’s catalogue price is also regulated. The “30–50% spread” is the regulator’s design, mechanically reproduced in every hospital catalogue in Switzerland that participates in SL pricing.

What changed in the analysis

Two corrections to early signal logic:

1. signal_price_outlier retired the brand-vs-generic comparison. The signal still detects price outliers, but the brand-vs-generic spread is treated as expected baseline, not as variance.
2. signal_cross_site_price_consistency got a regulator-aware filter. Two hospitals showing the same generic-brand spread within the BAG range is consistency. Two hospitals showing different spreads outside the BAG range is the signal — that’s where procurement or data-quality is varying.

A third move: the Wisdom anchored a per-tenant 50cents listing which molecules in the catalogue are governed by SL pricing. Materials outside SL pricing (medical devices, non-pharmaceutical supplies, off-label drugs) follow market dynamics and remain on the standard outlier-detection track.

What stays interesting

• A generic-brand spread outside the BAG range — either tighter or looser than 30–50% — is real signal. It usually means either a catalogue data-entry error or a non-SL drug that someone classified as SL.
• A generic-brand spread that drifts over a billing year can indicate that one side of the comparison (typically the brand) had a regulatory price update that the catalogue didn’t pick up. Stale catalogue, real downstream effect.

Why this is in a Lessons journal

Because the line between signal and noise often runs through external regulation. The pricing data looks variable until you know the rule that fixes most of the variability; once you know the rule, the actionable signal becomes the deviation from the rule, not the variability itself. The Wisdom is the rule — committed in YAML, attributed to the provider who knows it, applied at the silver layer so the signals downstream read the right baseline.

Sources

• Wisdom: smebit_swiss_generic_pricing.yaml
• Provider: a hospital pharmacist with knowledge of BAG/SL pricing structure
• Anchor signals: signal_price_outlier, signal_cross_site_price_consistency
• Date Wisdom captured: 2026-05-08