The IP dispensing machine
The OneDose IP dispensing machine releases patient-specific doses at the point of care in an inpatient setting. It checks each release against the active, pharmacist-verified order, records the release against the patient, the nurse and the time, and writes the event back to the medication record — so the dispense stage of the medication loop is confirmed rather than assumed.
The design goal is not to lock medication away. It is to make the fast path and the recorded path the same path, because when they differ, staff correctly choose the fast one.
How does it work?
It loads against verified orders, not against a stock list
The machine holds doses assigned to patients on the unit, populated from orders a pharmacist has already verified. That is the difference between patient-specific dispensing and floor stock: the dose in the drawer is already associated with a person and an order before anyone reaches for it.
The nurse authenticates, and the machine resolves the patient
Access is per-user rather than per-unit, so every release has a name attached. The nurse selects or scans the patient and the machine presents what is actually due for that patient at that time — rather than presenting the whole formulary and trusting the selection.
It releases only what matches the active order
The check runs at the moment of release against the current order, so a dose discontinued an hour ago is not in the list. This is the stage where paper systems and simple cabinets are weakest: they reflect the order set as of the last restock.
An override is permitted, and permanently logged
Clinical reality includes doses needed before pharmacy verification. Blocking that outright produces a workaround, and the workaround has no record at all. So the override exists, requires a reason, and produces a logged exception a pharmacy director can review as a list rather than discover during an audit.
The release is written back in real time
Stock movement and the patient record update from the same event rather than being reconciled afterwards. Reconciliation performed later is reconciliation performed from memory, which is where controlled-substance counts begin to drift.
Early and late releases are flagged rather than silently allowed
A dose released well outside its window is a real clinical event — it may be entirely appropriate, and it should not be invisible. The machine records the deviation with the reason, which turns a timing question from an audit archaeology exercise into a report.
Controlled substances are counted at the event, not at the shift
Counts update at the moment of release and waste is documented against the same event. A discrepancy then surfaces as a specific transaction with a name and a timestamp rather than as a number that failed to reconcile eight hours later, which is the difference between an investigation with a starting point and one without.
Where does it sit in the medication loop?
Closed-loop medication management runs in seven stages. This machine operates on the ones below, and the pillar page describes all seven including the two that belong to a pharmacist and a nurse rather than to any equipment.
Stage 4 · Dispensing to the point of care
The OneDose IP dispensing machine releases patient-specific doses at the point of care and records the release against the patient, the nurse and the time, so an override becomes a logged exception rather than an invisible one.
Stage 5 · Administration at the bedside
OneDose supports barcode-verified administration against the packaged dose and writes the result back. The clinical override decision stays with the nurse; the record of it does not stay with the nurse.
Stage 6 · Documentation and reconciliation
Administration events are written back in real time with an immutable audit trail. Where a facility runs its own eMAR, OneDose writes to it rather than replacing it.
How is this different from an automated dispensing cabinet?
An automated dispensing cabinet is primarily a secured stock location with an access log. It knows what left the drawer and who opened it; whether that dose was the right one for the patient it was intended for is a question the cabinet largely trusts the nurse to have answered.
Patient-specific dispensing inverts that. The unit of organisation is the patient and the order rather than the drug and the pocket, so the machine can check the release against what is actually due. That removes a class of selection error the cabinet is structurally unable to catch.
It is worth being fair about the trade, because it is real. Cabinets are excellent at immediate availability across a broad formulary, which is exactly what an emergency department or a theatre needs. A unit whose medication use is unpredictable and time-critical is a unit where stock-based access is the right answer, and a patient-specific machine is the wrong tool. Most hospitals end up running both, and the sensible question is which units get which.
What happens when the pharmacy is closed?
This is the question that decides whether a dispensing deployment survives its first month, and it is the one most vendor material skips.
Out-of-hours access has to work, because the alternative is a nurse finding another route to the medication and that route is invisible. So the machine supports out-of-hours release with a recorded reason and a pharmacist review queue that is worked the following morning — the dose happens on time and the verification happens late, rather than the dose happening off-record and the verification never happening at all.
The honest framing: this does not close the loop at that moment. It converts an untracked event into a tracked one with a known gap, which is a smaller claim than "closed loop" and is the true one.
Dispensing, packing, or a cabinet — which do you actually need?
These three get sold interchangeably and they solve different problems, so it is worth being blunt about the distinction.
A packing machine addresses preparation: the right contents assembled into an identifiable package, in advance. A dispensing machine addresses release: the right dose leaving at the right time, attributed to a person and a patient. A cabinet addresses availability and security: the drug is present, locked, and access is logged.
A hospital with a preparation problem — hand-sorted rounds, unlabelled split tablets — does not fix it by buying a cabinet. A hospital with a controlled-substance reconciliation problem does not fix it by buying a packing machine. The most common expensive mistake in this category is solving a well-understood problem with the adjacent product.
Does automated dispensing reduce medication errors?
It removes specific error types and introduces others, and any vendor answering this with a single percentage should be asked which error types they counted.
What it genuinely removes: selecting the wrong product from a shelf of look-alike packaging, giving a dose that was discontinued, and losing the record of a controlled substance movement. These are real, common, and mechanically preventable by a machine that checks the order at release.
What it can introduce: automation complacency, where a check that used to be performed by a person is now assumed to have been performed by the machine — and loading errors, where a stocking mistake is faithfully propagated to every subsequent patient rather than being caught by the next person to look at the shelf. A machine turns a one-off human error into a systematic one, which is a genuinely different risk profile and belongs in the implementation plan rather than in the objections file.
It does not perform clinical verification. A pharmacist verifies the order; the machine enforces what was verified.
It does not decide whether an override is clinically appropriate. It requires a reason, records it, and puts it in front of a human who can judge it.
It does not replace an emergency stock supply. Time-critical, unpredictable medication use is a legitimate case for cabinet or floor stock, and pretending otherwise produces workarounds.
It does not administer. A nurse administers; the machine records what was released and what was scanned.
It does not judge whether an out-of-window release was appropriate. It records the deviation and the reason for a human to review.
It does not remove the need for a bedside identity check. Releasing the right dose and giving it to the right patient are two different guarantees.
It does not eliminate stocking error. A machine loaded incorrectly propagates that error systematically, which is why loading verification is part of the deployment rather than an afterthought.
This page also publishes no throughput, capacity or cycle-time figures. Those are measured values that depend on product mix and changeover frequency, and a number printed without the conditions it was measured under is a marketing number. Ask, and what exists will be shown with its conditions attached — and against a product mix resembling yours rather than a favourable one.
Frequently asked
- What is an IP dispensing machine?
- An IP (inpatient) dispensing machine releases patient-specific medication doses at the point of care on a hospital ward. Unlike a stock-based cabinet, it is organised around the patient and the active order rather than around the drug, so it can check each release against what is actually due for that patient at that time, record it against the nurse and the timestamp, and write the event back to the medication record.
- How is an IP dispensing machine different from an automated dispensing cabinet?
- A cabinet is a secured stock location with an access log — it records what left the drawer and who opened it, and largely trusts the nurse to have selected correctly for the patient. A patient-specific IP dispensing machine is organised around the patient and the verified order, so it can check the release itself. Cabinets remain better where medication use is unpredictable and time-critical, such as an emergency department or theatre, and most hospitals run both.
- Can nurses override an IP dispensing machine?
- Yes, and a system that cannot be overridden is a system that will be worked around. Clinical reality includes doses needed before a pharmacist can verify them. The OneDose machine permits the override, requires a reason, and produces a logged exception a pharmacy director can review as a list — which is materially better than blocking the release and losing the event entirely.
- What happens if the network or the machine goes down?
- Medication access cannot depend on a network being up, so there has to be a defined downtime path — physical access with a recorded reason, and reconciliation when the system returns. The question to ask any dispensing vendor, including OneDose, is to walk you through their downtime procedure and show you the reconciliation report it produces. A vendor without a rehearsed answer here has not been deployed in a hospital.
- What is the difference between an IP dispensing machine and a dose dispensing machine?
- For OneDose, none — they are the same machine under two names, which is why this page covers both. In general industry usage a "dose dispensing machine" describes any equipment that releases an individual dose at the point of care, and an "IP dispensing machine" specifies that it does so for inpatients against a named patient and an active order. If you are comparing two vendors, establish which stage of the medication process each is describing — preparing a dose and releasing one are different jobs.
- Does automated dose dispensing reduce medication errors?
- It removes specific error types — selecting the wrong product from look-alike packaging, administering a discontinued dose, and losing the record of a controlled substance movement. It can also introduce automation complacency, where a check formerly performed by a person is assumed to have been performed by the machine, and it converts a stocking error from a one-off into a systematic one propagated to every subsequent patient. Both belong in an implementation plan rather than being argued away.