Hospital pharmacy inventory control: the methods that actually hold

Most hospital pharmacies do not lack inventory methods. They lack methods that survive a Tuesday. The textbook approach assumes a warehouse with predictable demand, a single point of issue and staff whose whole job is stock. A hospital pharmacy has none of those things.

What follows is the subset of inventory control that holds up in a live hospital, and how each piece changes when the stock in question is a drug rather than a spare part.

Every stock method starts by admitting that you cannot manage every line with equal attention. A mid-sized hospital pharmacy carries a few thousand items. Nobody controls all of them tightly, so the choice is whether the prioritisation is deliberate or accidental.

ABC analysis ranks items by annual spend. A small number of lines will account for most of the money, and those are where tight control pays for itself. This is the classification that protects your budget.

VED classification ranks the same items by clinical consequence instead: vital, essential, desirable. This is the classification that protects your patients, and it routinely disagrees with ABC.

The disagreement is the point. A cheap vital item is invisible to cost analysis and catastrophic to run out of. Score every line on both axes and manage the combination: tight financial control on the expensive, generous buffers on the clinically critical, and light touch par levels on everything that is neither.

Ordering by feel is the single most common stock management method in hospital pharmacy, and it fails in both directions at once. The same pharmacy will hold nine months of one item and run out of another in the same week.

A reorder point is arithmetic, not judgement: average daily consumption multiplied by the supplier's real lead time, plus a safety buffer sized by the VED score. Set it per item, per supplier.

Ward stock is where most pharmacies lose control, because the ward is not a customer placing orders, it is a second store that nobody counts. Requisition-based replenishment means wards ask for what they remember running out of, which is a memory test with a stock consequence.

Par levels invert it. Agree a standing quantity per item per ward, then top up to that level on a fixed round. The ward stops ordering, the pharmacy starts replenishing, and the quantity on the ward becomes a number somebody chose rather than a residue of past panics.

This only works if issue to the ward is recorded properly, which is the same problem covered in the pillar guide to pharmacy inventory management: stock issued against a patient solves inventory and billing in one action.

An annual full count is the least useful form of stock verification. It closes the pharmacy, takes two days, and produces a list of variances that are months old and therefore unexplainable. Nobody learns anything, and the same variances recur.

Cycle counting spreads the same effort across the year and finds errors while their cause is still traceable. Count your A items monthly, your B items quarterly, and the long tail annually. A variance found within thirty days usually has a discoverable cause: a mis-keyed issue, an unrecorded return, a batch substituted at the shelf.

The measure worth tracking is not the size of the variance but how quickly you can explain it.

Every pharmacy intends to issue oldest stock first. Under pressure, staff reach for the nearest box. The intention does not survive a busy evening, so it has to be built into the issue step rather than left to discipline.

That means batch level stock, an expiry date on every batch, and a system that proposes the correct batch at the point of dispensing. The detail of getting that right is covered in expiry and batch tracking.

Narcotics and other controlled items are the exception to every efficiency argument in this article. Here the friction is the feature: named issue, dual verification where policy requires it, a running balance that must reconcile, and an audit trail that no role can edit.

The same logic applies to write-offs and stock adjustments. If any user can adjust a quantity to match the shelf, your stock figure is a reflection of the shelf rather than a check on it. Adjustment should be a permissioned action with a reason code and a name attached.

None of this requires software in principle, and all of it requires software in practice once you are past a few hundred lines and more than one store. The methods above assume you can see consumption per item, stock per batch per location, and issue per patient, which is not something a spreadsheet maintains across three shifts.

BBB HMS carries pharmacy inventory inside the hospital system, so consumption data comes from real dispensing, ward issue is recorded against the patient, and the reorder and expiry views are built from the same records rather than re-entered.

Pick the two or three methods that address your actual failure mode. If you are writing off expired stock, start with batch tracking and first expiry first out. If you are stocking out, start with reorder points and VED. Adopting all of it at once is how inventory projects stall.