Your safety levels were set last year. Your patients are here today.
We built a simulation: 500-bed hospital, 3,500 drugs, 15 wards, 24 months of data. ฿600M annual drug budget. What the model reveals: ฿41 million per year lost to a problem hiding in plain sight.
The ICU nurse calls the pharmacy. Vancomycin — stat. The system says zero stock. The patient waits.
It's 2:47 AM in a 500-bed Thai hospital. The ICU attending orders IV Vancomycin for a sepsis patient. The nurse walks to the automated dispensing cabinet. Zero units. She calls the central pharmacy. Also zero. The on-call pharmacist initiates an emergency procurement order. The supplier answers at 3:15 AM. The drug arrives at 4:52 AM — two hours after the order, at 2.8x the contract price.
Meanwhile, three floors down, Ward 5's storeroom holds 142 boxes of Metformin 500mg expiring in 38 days. Nobody will use them in time. Ward 7 has been hoarding Omeprazole 20mg — 6 months of safety stock — while Ward 3 ran out yesterday and borrowed from a neighboring hospital.
This is not a story about a bad hospital. This hospital has JCIA accreditation. Its pharmacists are experienced and dedicated. Its procurement team follows MDS-3 guidelines. The problem isn't people. It's the system they're trapped in.
"The problem isn't shortage or surplus. It's that nobody can see both at the same time."
Static inventory management costs this hospital ฿41 million per year across three categories that most administrators never see combined.
| Cost Category | Annual Amount | % of Drug Budget | Description |
|---|---|---|---|
| Drug Expiry | ฿32M | 5.3% | Drugs that expire before use due to over-ordering and poor rotation |
| Emergency Procurement | ฿9M | 1.5% | Premium pricing on urgent orders (avg 2.8x contract price) |
| Tied-Up Capital | ฿18M | — | Opportunity cost of excess inventory sitting in ward storerooms |
| Total | ฿41M | 6.8% | Combined annual waste from static inventory management |
MDS-3 guidelines assume stable demand, predictable lead times, and rational ordering. Hospital reality delivers none of these.
| Factor | MDS-3 Theory | Hospital Reality |
|---|---|---|
| Demand Pattern | Stable, normally distributed | Intermittent, seasonal, ward-dependent |
| Lead Time | Fixed, known in advance | Variable (3–45 days), supplier-dependent |
| Review Frequency | Regular periodic review | Annual budget cycle, rarely updated |
| Data Quality | Complete consumption records | Dispensing data ≠ actual consumption |
| Ward Behavior | Wards order what they need | Wards hoard to protect against stockouts |
| Safety Stock Formula | Z × σd × √LT | Pharmacist judgment + last year's budget |
"Sick inventory management systems generally feature subjective, ad hoc decisions about what to purchase and how much to buy, often based on budgets rather than actual consumption data."
— MDS-3: Managing Access to Medicines and Health Technologies, Ch. 23Most hospitals set safety stock at a single service level (typically 95%) for all drugs. But a 95% service level for Vancomycin is a clinical risk, while 95% for a common vitamin is wasteful overkill. The right Z-value depends on the drug's criticality, not a blanket policy.
Demand variability amplifies 8.8x as it flows from patient to supplier. This is the hospital bullwhip effect.
The amplification ratio — 72.1% / 8.2% = 8.8x — means a modest fluctuation in patient demand becomes a wild swing in supplier orders. Emergency orders spike. Suppliers respond with longer lead times. The hospital responds by ordering even more. The cycle reinforces itself.
| Metric | Current (Static) | Projected (AI-Driven) |
|---|---|---|
| Bullwhip Amplification | 8.8x | 3.8x |
| Emergency Orders / Month | 23 | 6 |
| Critical Stockouts / Month | 14 | 3 |
| Drug Expiry Rate | 5.3% | 1.8% |
Projected improvements based on simulation modeling. See Chapters 4–5 for methodology.
"The ward nurse who hoards isn't irrational. She's protecting her patients with the only tool she has: excess stock."
6.4 million records across 5 hospital systems over 24 months. Enough to see the patterns hiding in daily operations.
| Data Source | Records | Description |
|---|---|---|
| HIS (Hospital Information System) | 4.2M | Patient encounters, diagnoses, prescriptions |
| Pharmacy Dispensing | 1.8M | Dispensing events, returns, ward transfers |
| Procurement | 340K | Purchase orders, supplier invoices, lead times |
| Patient Admissions | 89K | Admission/discharge events, ward assignments |
| Operating Room | 12K | Surgical drug usage, anesthesia records |
| Total | 6.4M | 24 months of integrated hospital drug data |
Raw hospital data is never analysis-ready. We applied four critical data quality steps:
Identified periods where zero dispensing was caused by stockouts, not zero demand. Cross-referenced with ward requisition denials and emergency orders.
For stockout periods, estimated true demand using Bayesian imputation from historical patterns, ward census, and diagnosis mix.
Mapped batch-level expiry dates against consumption velocity to identify drugs at risk of expiry 30, 60, and 90 days in advance.
Built consumption profiles for each of 15 wards — identifying hoarding patterns, seasonal demand shifts, and inter-ward transfer opportunities.
Data Pipeline Architecture
"Hospital data is abundant but fragmented. The challenge isn't collecting more — it's connecting what already exists."
Quick ROI math, self-assessment questions, and industry benchmarks to see where you stand.
| Metric | Current | Projected | Annual Saving |
|---|---|---|---|
| Drug Expiry | 5.3% (฿32M) | 1.8% (฿11M) | ฿21M |
| Emergency Procurement | ฿9M | ฿2.4M | ฿6.6M |
| Freed Working Capital | ฿18M tied up | ฿7M tied up | ฿11M released |
| Total Potential | — | ฿38.6M/year | |
| Metric | Typical Thai Hospital | Good | Best-in-Class |
|---|---|---|---|
| Drug Expiry Rate | 4–7% | 2–3% | <1% |
| Critical Stockouts / Month | 10–20 | 3–6 | <2 |
| Emergency Order Rate | 3–5% | 1–2% | <0.5% |
| Inventory Turnover (days) | 60–90 | 30–45 | 15–25 |
Get a Free Waste Assessment
We'll analyze your hospital's drug inventory data and identify the top three areas of hidden waste. No cost, no commitment.
Request AssessmentTrusted by hospital pharmacy teams across Southeast Asia. Data handled under NDA.
