Situation: A hospital will introduce smart infusion pumps with a drug library on its adult medical-surgical wards next month. A nurse manager joins the team that reviews the infusion process before the pumps go live. At the 22:00 bedside endorsement, the outgoing nurse hands over four clients who are on the new pumps. All four pumps show no alarms. Which client's infusion should the two nurses check together FIRST?

Endorsement starts with the clients at greatest risk, and two rules combine here: high-alert drugs carry the most harm from a pump error, and a newly started infusion has not yet been verified over time. Three clients receive high-alert drugs, but only the potassium chloride infusion is both high-alert and newly started, so its concentration, rate, and site are checked first. The long-running heparin and insulin infusions have settings already confirmed by stable results, and maintenance fluids are low risk.

At endorsement, the first decision is not simply “which pump is alarming,” because all four pumps are silent. The question is which silent pump carries the greatest potential for serious harm if its settings are wrong. Two factors combine to identify the priority: the drug’s risk class and how recently the infusion was started.

High-alert medications are drugs that cause significant patient harm when used in error. Potassium chloride concentrate, insulin, and heparin all belong to this category. However, a newly started high-alert infusion has not yet been observed over time for signs that the rate, concentration, and vascular access site are correct and tolerated. A pump that has run for hours or days with stable clinical parameters has already accumulated indirect evidence that its programming is appropriate.

The potassium chloride infusion began only 30 minutes ago, so it is the only option that is both high-alert and newly initiated. The two nurses should verify the prescribed concentration, the programmed rate, the patency and appearance of the IV site, and the patient’s most recent potassium level and renal function together. Rapid or concentrated potassium administration can cause fatal hyperkalemia and cardiac arrhythmias, and early errors may not yet have produced an alarm or a visible clinical change.

The insulin infusion has a reassuring pattern: glucose in range for 12 hours suggests the current rate is appropriate. The heparin infusion has been at an unchanged rate for 2 days, and stability over that period lowers the likelihood of a newly introduced programming error. Maintenance intravenous fluids are not high-alert and are the lowest priority.

Watch out! Smart pumps with drug libraries reduce errors, but they do not eliminate them. Errors frequently arise from bypassing the drug library, overriding soft limits, or selecting the wrong drug entry ^[2]. A silent pump does not mean a safe pump.

Key point! During endorsement, prioritize the combination of highest-risk drug and shortest time since initiation. A newly started high-alert infusion always warrants a two-nurse bedside verification of the order, pump settings, line, and site before other stable infusions are reviewed.

The rationale also aligns with the purpose of smart pump implementation. Drug libraries standardize dosing limits and guardrails, but their benefit depends on clinicians actually using the library and responding to alerts appropriately . Interoperability with the electronic medical record can further improve programming compliance, yet the bedside check remains essential immediately after a high-alert infusion is started .