EVD Demand Forecasting: Using ICU Census and Procedure Data

EVD demand forecasting works best from the bottom up: take twelve to twenty-four months of actual usage and procedure data, split it by catheter type and drainage route, adjust it for known changes in neuro-ICU census and case mix, and then add safety stock that reflects both variability and replenishment lead time. The result is a par level.

Why is EVD demand harder to forecast than elective implants?

Elective implants follow the OR schedule. External ventricular drains follow emergencies: subarachnoid and intraventricular hemorrhage, traumatic brain injury, acute hydrocephalus and perioperative needs in tumor surgery. Demand arrives irregularly, sometimes several drains in a few days and then none for a week.

Volumes are also small enough that random variation matters. A hospital that places a modest number of drains per month will see months that look very different from the average without anything having changed. On top of that, a single account may use two or three variants: standard and antibiotic-impregnated catheters, ventricular and lumbar drainage, and sometimes different system configurations for the OR and the ICU.

A good forecast accepts this irregularity instead of trying to remove it. Its purpose is to make sure a drain is always on the shelf when a patient needs one, without filling the storeroom with stock that expires.

Which data sources should a distributor ask for?

Hospitals hold most of the data needed, but it sits in different systems. Ask for it early, and explain that the purpose is to set reliable stock levels.

No single source is complete. Usage from materials management combined with procedure records usually gives the most realistic picture of how many systems are consumed.

How do you build a baseline forecast?

Build the baseline in a few simple steps:

• Collect monthly usage for at least twelve months, ideally twenty-four, for each item number.
• Separate the series: standard ventricular drainage, antibiotic-impregnated ventricular drainage, lumbar drainage, and any system used only in the OR.
• Remove one-off events that will not repeat, such as stock transferred to another hospital or a bulk return.
• Calculate the average monthly usage and how much it varies from month to month.
• Adjust for known changes: a new neurosurgeon joining, a protocol change in catheter choice, or a new service line.

For a DESU account, the series would map to the products in use: the EVD-010 Series with standard or antibiotic-impregnated catheters, the EVD-020 Series for teams that use one drainage system with ventricular or lumbar catheters, and the lumbar drainage catheter kit. Each needs its own line in the forecast, because the split between them is what usually goes wrong.

How do ICU census and case mix refine the forecast?

Usage history tells you what happened. Census and case mix tell you what is likely to change. The neuro-ICU's admissions by diagnosis are a useful leading indicator, because drains are placed for a recognizable set of conditions.

Watch for changes such as:

• a hospital gaining or upgrading its trauma center or stroke center designation;
• new neurosurgeons or neurointensivists joining the group;
• additional neuro-ICU beds, or a step-down unit that keeps drained patients longer;
• transfer agreements that bring more patients with hemorrhage from community hospitals;
• a protocol change, such as broader use of antibiotic-impregnated catheters or more lumbar drainage after skull base surgery.

A simple internal ratio helps: the number of drains used per neuro-ICU admission with a relevant diagnosis, calculated from the hospital's own history. If admissions rise, the ratio gives a first estimate of how usage will follow. Treat it as a cross-check rather than a precise model, and recalculate it as new data arrives.

How should a forecast turn into par levels and orders?

The forecast describes demand. Par levels and orders describe how supply meets it. Two numbers drive the par level: the average demand during the replenishment lead time, and the safety stock that protects against variation during that time.

A widely used approach sets safety stock in proportion to the variability of demand and the square root of the lead time. In plain terms, the more irregular the usage and the longer it takes to replenish, the more buffer is needed. For low-volume, emergent products like EVD, many hospitals also set a practical floor: a minimum number of systems at each location, regardless of the calculation, so that a cluster of emergencies never empties the shelf.

Hospitals usually keep stock in more than one place: the neuro-ICU, the OR and sometimes the emergency department. Each location needs its own par level, and the distributor's own warehouse needs a buffer that covers the time it takes to receive goods from the manufacturer. For a manufacturer outside the US, that lead time includes production planning, international shipping and customs processing, so the distributor's buffer matters as much as the hospital's.

Setting par levels by location

Each storage location behaves differently, so each deserves its own reasoning:

• Neuro-ICU: most bedside placements and all ongoing management happen here, so it usually holds the largest share of drainage systems and both catheter types.
• Operating room: drains placed during or before surgery, including posterior fossa and tumor cases; often a smaller stock, with a clear process for replacement after use.
• Emergency department: some hospitals keep one or two systems for urgent bedside placement; others bring them from the ICU. Agree which model the hospital uses.
• Distributor warehouse: the buffer behind all locations, sized for the replenishment lead time from the manufacturer.

Review each location separately. A par level that is right for the hospital as a whole can still leave one location empty while another holds stock that is close to expiry.

How do expiry and rotation affect the forecast?

Every sterile kit has a shelf life. Stock that sits too long expires and becomes waste, which damages the account relationship as much as a stockout. Rotation rules keep this under control: first expiry, first out in the warehouse and at each hospital location, with periodic checks for items approaching expiry.

Low-volume items, often the less used catheter type, are where expiry risk concentrates. Consider consolidating them into one location in the hospital, or holding them at the distributor's warehouse with fast delivery, rather than spreading small quantities across several units.

How often should the forecast be reviewed?

Review usage against the forecast every month, and review the assumptions with the hospital every quarter. Share the rolling forecast with the manufacturer as well. A manufacturer that sees demand trends early can plan production and shipments, which shortens effective lead times and reduces the buffer everyone has to carry.

DESU manufactures in its own cleanrooms, with 5,000 m² of production in Ankara, Türkiye, and works under an ISO 13485 quality management system. Partners who share forecasts give the planning team the visibility it needs to keep supply predictable.

Talk to a DESU product specialist

If you are planning stock for an EVD line or preparing par levels with a hospital, talk to DESU about forecasting, order cadence and buffer stock for your market. Write to sales@desu.tr or request a quote with your expected volumes.