Situation: A 59-year-old woman on post-operative day 1 after a bowel resection develops fever, back pain, and hypotension 15 minutes into a transfusion of packed red blood cells, and she dies the next day. The unit had been crossmatched for another client in the same four-bed room. The hospital's patient safety committee reviews the event. Three months after the action plan starts, the committee wants to know whether nurses actually carry out each step of the new bedside identification procedure. Which method gives the MOST direct measure of this?

Whether a procedure is carried out is a process measure, and the most direct way to measure practice is a concurrent audit: watching transfusions as they happen against each required step. A retrospective audit of discharged charts shows only what was documented, and a signed form does not prove that the bedside check took place. Reaction rates are a rare outcome influenced by many factors, and a written test measures knowledge rather than practice.

Core concept: measuring practice, not knowledge or outcomes

The committee is asking whether nurses actually perform each step of the new bedside identification procedure. In quality improvement terms, this is a process measure—it evaluates what staff do in real time, not what they know, what they wrote later, or what eventually happened to the patient. The most direct way to capture actual practice is to watch the procedure as it occurs and compare each observed action against a predefined list of required steps. That is precisely what a concurrent audit using a checklist does.

A written test measures knowledge, not performance; a chart review measures documentation, not bedside behavior; and reaction rates are an outcome that is too rare and multifactorial to tell you whether each step was performed.

Why the other options fall short

Option	What it actually measures	Why it is less direct
1. Written test	Knowledge of the procedure	Knowing the steps does not guarantee performing them under clinical pressure
3. Monthly transfusion reaction count	Outcome (rare adverse event)	Reactions are influenced by patient factors, product factors, and detection; a low count cannot confirm each identification step was done
4. Review of completed forms	Documentation after the fact	A signed or checked form shows what was recorded, not what actually happened at the bedside
2. Direct observation with checklist	Actual practice in real time	Captures each step as it is performed, allowing immediate comparison with the required procedure

Watch out! Retrospective chart audits are tempting because they are easy, but they only verify that a form was completed. In transfusion safety, the critical failure is often at the bedside identification step—and that step can be skipped even when the paperwork looks perfect.

How the evidence supports direct observation

The Sudan clinical audit ^[1] used bedside observation against WHO-aligned criteria to measure compliance with identification, monitoring, and documentation. The authors specifically chose to audit the bedside steps because those are the actions that prevent wrong-patient transfusion and allow early recognition of reactions. That study design—watching transfusions as they happen and checking each required element—is the same logic as the correct answer here.

The College of American Pathologists Q-Probes study ^[2] also used prospective audit of nonemergent red blood cell transfusions to measure how often health care workers completed required identification and monitoring procedures. The prospective design is important: it captures practice at the moment of care rather than relying on memory or records. The study linked hemolytic transfusion reactions to failure to follow established identification procedures, reinforcing that the bedside check is the step that matters.

The Johns Hopkins sentinel event report describes a wrong-patient transfusion that occurred despite existing manual verification processes. The hospital’s quality improvement response included direct observation and technology-assisted verification because paper-based documentation of dual-signature verification had not prevented the error. This illustrates why reviewing forms or tracking outcomes alone would not answer the committee’s question: the gap is between what is documented and what is actually done.

A concurrent audit—observing the transfusion with a checklist of each required step—is the only option that directly measures whether nurses carry out the new bedside identification procedure.

Key point! When a quality improvement question asks about carrying out a procedure, choose the method that observes the procedure in real time. Knowledge tests, chart reviews, and outcome tracking are indirect and can all look acceptable even when the critical bedside step is missed.