6 Quality Improvement in Medication Safety

Purpose of medication-safety improvement

Medication-safety improvement uses information about errors, near misses, and hazardous conditions to make medication-use processes safer. The aim is not simply to count errors or remind individuals to be more careful. Teams identify how the system made an error possible, test changes that address those conditions, and check whether safety improves without creating new problems.

Gather and interpret safety data

Safety information can come from incident and near-miss reports, pharmacy interventions, direct observation, chart review, barcode and smart-pump records, medication reconciliation audits, and staff or patient feedback. Include events that caused harm, reached a patient without harm, were intercepted before reaching a patient, or involved unsafe conditions that could lead to an error.

For each event, capture consistent details where possible:

• Medication and risk: drug, strength, dosage form, route, and whether it is a high-alert medication or a look-alike/sound-alike (LASA) medication.

• Process stage: prescribing, transcription, dispensing, preparation, administration, or monitoring.

• Event and outcome: what happened, whether the event reached the patient, and any harm.

• Context: location, shift, workload, interruptions, staffing, technology, packaging, storage, and safeguards that were present or absent.

• Detection: who or what detected the problem and how it was intercepted.

Look for combinations of conditions, not just repeated drug names. For example, several selection errors involving similar-looking containers on the same shift may indicate a storage or labeling hazard. High-alert medicines deserve particular attention because errors involving them can have especially serious consequences, although such errors are not necessarily more frequent. LASA medicines may be confused because of similar names, packaging, strengths, or dosage forms.

Interpret counts and find patterns

A rise in reports does not automatically mean safety has worsened; staff may be reporting near misses more consistently. Fewer reports do not prove that errors have declined either. Voluntary reports can help detect hazards, but may capture only a fraction of events and usually lack a reliable denominator.

Combine reports with other data sources. When comparing periods, use a meaningful denominator, such as errors per $1{,}000$ doses or per $100$ administrations, rather than raw counts alone. Stratify results by medication, process stage, unit, shift, route, patient group, or event severity to identify where risks cluster. Protect confidentiality and use findings to learn about the process, not to rank staff by report counts.

Analyze how events occur

A systems analysis reconstructs an event in sequence: what was expected, what actually happened, how the error was detected, and what conditions shaped the actions of people and technology. Ask what made the unsafe action possible and what made it difficult to notice or stop.

Contributing conditions may include confusing labels, similar storage locations, poorly designed order screens, interruptions, unclear procedures, workload, or missing monitoring. Do not stop at an individual’s final action or treat “human error” or “re-educate staff” as a complete explanation or solution.

For an event that has already occurred, root cause analysis is a structured retrospective approach. A timeline, fishbone diagram, and Five Whys can help organize the inquiry.

Look ahead for process vulnerabilities

For a process that has not yet caused harm, failure mode and effects analysis (FMEA) offers a prospective way to examine risk. Map the process steps, identify ways each step could fail, consider the possible consequences and the chance that a failure would be detected, then prioritize vulnerabilities for action. FMEA ratings help with prioritization but do not replace judgment or local evidence.

Choose safeguards that address the cause

Choose changes that make the safe action easier and the unsafe action harder. A layered approach is often more dependable than relying on memory, training, or a single warning. The safeguard should fit the identified risk:

• Standardize: use approved concentrations, order sets, labels, and preparation processes.

• Separate and distinguish: store LASA medicines apart, improve visual differentiation, and use clear labeling or appropriate Tall Man lettering where the organization has adopted it.

• Add reliable barriers: use barcode verification, suitable clinical decision support, or smart-pump safeguards. Check whether alerts are meaningful rather than routinely overridden.

• Limit risk: restrict access to selected high-alert medicines and standardize where and how they are stored, prepared, and administered.

• Support the workflow: reduce avoidable interruptions during medication preparation and make current drug information easy to access.

• Strengthen monitoring: define the required observations and response steps for medicines with serious potential consequences.

A manual independent double check may help in selected situations, but it should not be assumed to be the best safeguard for every high-alert medication. Define what must be checked independently and ensure the check fits the workflow. Education can support a system change, but education alone is unlikely to correct a confusing or unsafe process.

Test changes and measure their effects

The Model for Improvement begins with three questions: What are we trying to accomplish? How will we know a change is an improvement? What change can we make? Its Plan–Do–Study–Act (PDSA) cycles let teams test changes on a small scale, learn from the results, and adapt before wider implementation.

Write a specific, time-bound aim and agree on definitions and a baseline before testing. For example: reduce wrong-strength selection errors involving selected high-alert medicines on an adult medical unit by $50\%$ within six months.

Use a small set of linked measures:

• Outcome measure: Did patient safety improve? For example, wrong-strength errors per $1{,}000$ administrations.

• Process measure: Was the safety step used? For example, the percentage of sampled administrations with the intended barcode scan completed.

• Balancing measure: Did the change cause a problem elsewhere? For example, medication-delivery delays or workarounds after a new verification step.

Plot data over time, note when changes were introduced, and examine results alongside frontline feedback. A run chart can help teams see patterns over time. Use the same definitions and denominators when comparing rates. A successful small test is a reason to test under additional conditions, not proof that the change will work everywhere. When evidence supports a change, build it into standard work, training, and system design, then continue monitoring for drift or unintended effects.

Apply the improvement process to a selection error

A unit reports several near misses in which one injectable medicine was selected instead of a similarly packaged medicine. Every event was intercepted before administration. The reports cluster around one storage cabinet, and observations show that the two products are in adjacent bins with similar labels.

1. Identify the pattern: Classify each event by medication, selection stage, location, shift, and detection point. Review cabinet observations and pharmacy interventions as well as reports.

2. Analyze causes: Map the process from selection through administration. Ask whether proximity, label design, cabinet search results, replenishment practices, or time pressure contributed. Do not assume that a staff reminder addresses the hazard.

3. Select a change: Test separating the products, improving bin differentiation, and checking whether electronic selection displays can better distinguish them. Train affected staff on the redesigned process.

4. Evaluate: Track selection near misses per relevant administrations, compliance with the new storage arrangement, and dispensing or administration delays. An initial increase in reports could reflect better detection or reporting; interpret it alongside other measures.

5. Adapt and sustain: Use PDSA cycles across different shifts. If the change reduces the hazard without unacceptable delays or workarounds, standardize it and continue auditing the process.

Sustain learning and improvement

Reliable medication-safety improvement combines reports with other safety data, uses appropriate denominators to find patterns, and examines system conditions rather than stopping at individual actions. Teams select safeguards that address the identified risk, test them in practice, and evaluate outcome, process, and balancing measures over time. Near misses and unsafe conditions offer opportunities to prevent harm before it reaches a patient.

• near miss

• LASA medication

• Plan–Do–Study–Act (PDSA) cycle