Medical Inventory Management Software Problems: The 7 That Cost Real Money, and How to Avoid Them

Why does lot and expiry capture get scoped as a field on the item record?

Because that is how every tool the buying committee has already seen does it. Sortly has an expiry field. Spreadsheets have a column. The requirement gets written as track expiration dates, everyone agrees, and nobody asks who types them or when.

The answer is that a materials coordinator putting away four totes on a Thursday afternoon does not type them, and no amount of training changes that. So the field exists, it is empty on most records, and the reports built on top of it are decorative. Meanwhile the underlying model is still quantity on an item, which means the system cannot represent the fact that you hold eleven of something across three lots with three different dates and one of them cannot be used after next month.

The downstream damage is larger than the write off line. Without lot as a real record there is no genealogy from receiving to shelf to the case where a product was used, so recall response is manual, tissue traceability under 21 CFR Part 1271 cannot be demonstrated from the system, and case costing is an estimate.

The fix is to make the lot a record and the scan the capture mechanism. Device barcodes in the GS1 or HIBC Unique Device Identification formats already encode lot number and expiration date, so a receiving screen that parses them captures both in one movement with no typing. Make that an acceptance test with your own products: hand a bidder a box from your shelf and ask them to show the parse. Then require the three record model in the data design, meaning the item, the lot and the physical unit are different things and expiry lives on the lot. A team that models quantity as a number on an item record will rebuild your spreadsheet problem in a nicer interface.

What goes wrong when you migrate the item master and take opening counts?

The item master is the work item everyone underestimates, and it is the one that gates go live.

The same gown is item 4471 at one center and a free text description at another. Distributor catalogue numbers, manufacturer part numbers and your internal codes exist in a partial and inconsistent mapping. Group purchasing contract prices sit in a PDF from the contracting office. Nobody notices while a human is interpreting a par sheet, because a human knows that blue XL gowns and item 4471 are the same thing. A system does not.

The failure mode is importing purchase history as the item master and going live with duplicates. Two records for one product means two par levels, two reorder points, and a transfer suggestion that never fires because each site holds a different code for the same thing. Staff lose confidence in week one and the private buffer stash in the back cabinet becomes permanent.

Opening counts fail differently. A count taken across five sites on different days, while ordering continues, produces a starting position that is wrong from the first hour, and every variance afterwards is argued rather than investigated.

The fix is to build the item master from distributor purchase history, deduplicate it deliberately, and map every legacy code before cutover. Load contract prices as data with effective dates so orders validate at the moment of purchase rather than at invoice review. Take opening counts site by site with ordering frozen for that site during the count, and go live one site at a time rather than across the group. Expect the item master work to be your staff's time, not the developer's, and put a named owner on it.

Why do the distributor and case feed integrations break after launch?

Electronic data interchange with McKesson, Medline, Cardinal Health or Henry Schein is well trodden and it still drifts, because catalogues change constantly. Items are discontinued and replaced with a successor code, pack sizes change, and a contract renegotiation moves prices mid quarter. If the catalogue sync is a one time load, your ordering slowly becomes wrong: purchase orders go out against retired codes, confirmations come back with substitutions nobody reviewed, and three way matching starts failing for reasons that look like software bugs and are actually stale reference data.

The case feed from a scheduling or clinical system such as HST Pathways or Surgical Information Systems breaks at the edges. A case cancelled after the cart was picked, a procedure changed on the morning, an add on case that never appeared in the overnight extract. Each of those leaves supplies consumed against a case that no longer matches, and case costing drifts without anyone seeing an error.

The fix is treating reference data as a living feed with monitoring. Refresh catalogues and contract prices on a schedule, flag discontinued items into a review queue rather than silently substituting, and reconcile three way match exceptions daily with a named owner. For the case feed, handle cancellations and changes as events rather than as replacements, and reconcile supplies consumed against cases performed weekly so a divergence shows up in days. Ask any bidder to name the specific distributors and the specific scheduling system they have connected before, because each is its own project and general integration experience does not transfer.

What happens when consignment implants and recall traceability are not covered?

These two get deferred together, usually because they sit outside what procurement platforms model, and they are precisely where your highest unit costs and worst documentation live.

Consignment is vendor owned stock on your shelf where money moves only after an implant goes into a patient. Deferred from scope, the workflow stays as it is: a rep drops off trunk stock, a circulating nurse peels barcode stickers onto a paper implant log, and the business office assembles a bill only purchase order days later, hoping the sticker sheet matches the contract price. Nobody reconciles vendor owned stock against invoices more than occasionally, so price variances and missing charges are found late or not at all.

Recall traceability fails because it depends entirely on the genealogy chain being unbroken. If lot capture was skipped at receiving, or if implant usage is captured on paper, no later feature can reconstruct it. This is one of the few problems in software that genuinely cannot be fixed after the fact.

The fix is to bring both into the first release even in reduced form. A consignment ledger per vendor covering what was checked in, what was implanted and what the rep swapped out, with usage captured by scanning the implant barcode against the case in the operating room and the bill only purchase order generated the same day at your loaded contract price. Mismatches then surface as exceptions rather than as surprises on an invoice. Genealogy follows automatically from the same scans, so a recall becomes a query returning affected lots, their current location or the case they went into, with a printable trail for an accreditation surveyor.

Should you build custom or configure what you already own?

If you run one or two locations, buy standard supplies from a single primary distributor, carry little consignment, and your real pain is ordering workflow, buy. Envi and Hybrent are competent procurement platforms and will improve requisitions, approvals and order confirmations for far less than a build costs. Take that deal and spend the difference elsewhere.

There is also a version of this where the tool you own is underused. Plenty of centers bought a procurement platform, configured requisitions, and never loaded contract prices or turned on catalogue sync, which are the features that would have solved half the complaint. Before commissioning anything, ask your incumbent vendor to show you contract price validation at order time and a par level driven reorder proposal using your own data. If those work, you had an implementation gap. If the honest answer is that the platform does not model lots, consignment or case level usage, that is a genuine fit limitation and no configuration closes it.

Build when three or more are true: three or more locations, implant consignment moving through your operating rooms weekly, expired write offs appearing every quarter, a need for case level supply cost to negotiate with payers or surgeons, and staff maintaining spreadsheets around the edges of a tool you already pay for. That last signal is the clearest one there is, because a shadow spreadsheet is a specification written by the people who know the work.

How do hidden costs get into the quote?

Six items, and two of them are not software.

• Each distributor connection. Purchase orders, confirmations and invoices as separate document flows, plus catalogue and contract price synchronisation, per distributor.
• The case feed. A connection to your scheduling or clinical system is its own project, and cancellations and same day changes are where the work is.
• Item master rationalisation. Your staff's time, and it gates go live at every site.
• Hardware and site setup. Scanners, wall mounted stations, shelf and bin labelling, opening counts and training, multiplied by every location.
• Compliance posture. Once supply usage links to a patient case the system touches protected health information, which means a business associate agreement under HIPAA, role based access and audit logging designed in from the start, plus tissue reporting under 21 CFR Part 1271 if you handle grafts.
• Network reality. Coverage in a concrete walled sterile core, and behaviour when a scanner drops connection mid pick.

What keeps the number down is one distributor, one pilot site and a clone, with consignment added once the receiving and picking chain is proven.

What separates a build that works from one that fails here?

Inventory decrements itself. Supplies picked for a case are scanned against that case, at cart assembly or at a station by the supply room door, so perpetual inventory replaces weekly counting and counting shrinks to a short daily cycle count over a rotating slice of the item master. If the quantity in the system only changes when a human counts, you have bought a digital par sheet with a subscription fee.

Scanning works with a gloved hand, in the room where the work happens, and keeps working when the network does not. Test that specifically during evaluation, in your own sterile core, not in a conference room.

Expiry management produces actions rather than reports: ninety, sixty and thirty day alerts by category, and a transfer suggestion moving product from the site that will not use it to the site that will. The write off line becomes a managed monthly number, and your current write off is the honest baseline against which to judge it.

Recall response is rehearsed. Run a drill in the first month with a real lot number and time it. Finding the gap during a drill costs an afternoon. Finding it during a manufacturer notification costs considerably more.

And ownership is written into the contract as work for hire with full source code and intellectual property assigned to your company, with infrastructure accounts in your name. Digital Heroes delivers complete code ownership on every project. Refuse any arrangement where the developer licenses the platform back to you, because that recreates the dependency you set out to escape.