Report Errors and Improve the System

Introduction

When something goes wrong with a medicine, two jobs follow. The first is to describe the event accurately and send the information to the people who can learn from it. The second is to find out why it happened and change the work so it is less likely to happen again. This lesson teaches both.

Reporting is not the same as blaming, and a report is not proof. A report says "this happened, and here is what we know." Other people, sometimes many reports later, work out what it means.

What you will be able to do

• Tell a near miss, a medication error and an adverse event apart, including cases where more than one label applies.
• Explain what MedWatch and VAERS are for and which one a problem belongs to.
• Sort a reported outcome into FDA's serious-outcome categories without deciding whether the medicine caused it.
• Identify who takes part in a report, what kind of product concern is being reported, and how a VAERS report is followed up.
• Reconstruct the conditions that contributed to an event, propose a focused improvement, and judge whether a small test of that improvement worked.

Before you start

You need everyday reading and simple counting. Every technical term is defined before it is used.

This lesson covers federal U.S. reporting programs and general patient-safety methods. It does not tell you any state's rules or any employer's procedures. Those can add their own requirements, and a federal program being voluntary does not cancel them. In real work, deciding whether a medicine caused a reaction, and what to do clinically, belongs to the pharmacist and prescriber. Your job as a technician is to notice the problem, preserve accurate information and bring it to the right person.

Three words for what went wrong

Medication error: a preventable mistake in the medication process

A medication error is a preventable event that may cause or lead to wrong medication use or patient harm while the medicine is controlled by a health professional, patient or consumer.

Plain takeaway: something in the medication process went wrong that could have been prevented, whether or not anyone was hurt.

The definition is broad on purpose. Errors can start in prescribing, communicating the order, labeling, packaging, product names, compounding, dispensing, distribution, administration, patient education, monitoring or use. A confusing package design can contribute to an error just as a rushed person can.

Near miss: the error happened, but harm did not

A near miss, also called a close call, is an error that did not harm the patient only because it was caught in time or because of luck.

Plain takeaway: the same mistake, with a lucky or well-caught ending.

The patient may never receive the wrong item, or they may receive it and still not be harmed. For example, the patient notices the pills look different and asks before taking any. The error was real. What stopped harm was detection or chance, not a safe process.

Adverse event: something bad happened to the patient

This term has two common meanings, and you need to know which one a source uses.

• FDA reporting meaning: any undesirable experience associated with the use of a medical product in a patient. "Associated" means it happened in connection with using the product. It does not mean someone proved the product caused it.
• Patient-safety research meaning: harm caused by medical care rather than by the patient's underlying illness. Reviewers make a judgment about cause before applying this label.

Plain takeaway: an adverse event is about what happened to the patient, not about whether anyone made a mistake.

An adverse event can happen with no error at all. A patient can take the right medicine at the right dose and still have a bad reaction. An adverse event that resulted from an error is called a preventable adverse event.

Worked example: one mistake, several endings

A pharmacy stores 10 mg and 20 mg tablets of the same medicine side by side in similar bottles. A prescription is written for 10 mg.

1. The technician fills it with 20 mg tablets. The pharmacist catches the mistake at the final check and refills it correctly. This is a medication error and a near miss. A preventable mistake occurred, and early detection stopped it from reaching the patient.
2. Change one fact: the pharmacist misses it. The patient takes the 20 mg tablets for two days, has no symptoms, and the error is found at the next refill. This is still a medication error and a near miss. The patient was exposed, but no harm resulted, so the outcome was luck.
3. Change one more fact: the patient becomes dizzy and falls while taking the wrong strength. This is a medication error with an adverse event. Whether the dose actually caused the fall is a clinical judgment. You can report the association without making that judgment.
4. A different case: the prescription is filled correctly with 10 mg tablets, and the patient develops a rash. This is an adverse event with no medication error. Nothing in the process went wrong, yet something undesirable happened.

The tempting mistake is to treat a near miss as "nothing happened." A near miss exposes the same weak spot that would have caused harm on another day. The only difference was the outcome. That is why near misses are worth recording.

Question 1

A pharmacy prints a label with the wrong patient's name. The packing technician notices the mismatch before the bag leaves the counter. Which label or labels fit: near miss, medication error, adverse event?

Question 2

A patient correctly takes a newly prescribed medicine exactly as directed and later reports an upset stomach. A coworker says, "That's an adverse event, so someone made an error." What is wrong with that reasoning?

Where reports go: MedWatch and VAERS

MedWatch: FDA's reporting program for medical products

MedWatch is the FDA's program for reporting serious problems with human medical products and for sharing safety information. Health professionals, patients and consumers can all report to it.

Plain takeaway: MedWatch is how FDA hears that something went wrong with a medicine or other medical product, so it can watch for patterns and warn people.

MedWatch covers prescription and over-the-counter medicines, biologics such as blood products, medical devices, combination products such as prefilled syringes, auto-injectors and metered-dose inhalers, and cosmetics, among others. FDA uses these reports for ongoing safety monitoring. A report can lead to changes in how a product is used or designed, and MedWatch publishes safety alerts.

VAERS: the national early-warning system for vaccines

The Vaccine Adverse Event Reporting System (VAERS) is a national early-warning system for detecting possible safety problems with vaccines used in the United States. The CDC and FDA run it together. It accepts reports of adverse events after a person receives a vaccine, and anyone can report.

Plain takeaway: VAERS collects reports of things that happen after vaccination so experts can spot unusual patterns early.

MedWatch does not take vaccine reports. FDA directs vaccine reports to VAERS, tobacco product problems to a separate safety reporting portal, and animal drug and device problems to their own pathway.

A report is not proof

VAERS accepts reports without judging whether the vaccine caused the event. A VAERS report only shows that the event happened sometime after vaccination. Some events after vaccination are coincidences, so a report is not documentation that the vaccine caused the event. VAERS also depends on people choosing to report. Many events are never reported, and reports vary in completeness and accuracy.

This is why providers are encouraged to report clinically important events even when they are unsure whether the vaccine was responsible. The system is built to collect possibilities. Deciding cause comes later, from looking at many reports and other evidence.

Worked example: choosing the pathway

• A patient's metered-dose inhaler does not spray properly straight out of a new box. The inhaler is a combination product, so this goes to MedWatch.
• A patient faints shortly after receiving a vaccine at the pharmacy. This goes to VAERS. The fainting happened after vaccination. The report does not say the vaccine caused it.

Changed case, from a product-specific instruction: VAERS gives specific routing for nirsevimab, which the source identifies by the brand name Beyfortus. Nirsevimab is an injected antibody given to young children to protect against RSV, and it is not a vaccine. If it was given at the same visit as one or more vaccines, VAERS says to report to VAERS. If it was given alone, report to MedWatch. The pathway depends on whether vaccines were given at the same visit, not just on the product name. This is a rule for one named product. Do not extend it to other products.

Question 3

A patient reports that a new bottle of over-the-counter tablets contains crumbled, discolored tablets. Which program fits, and why not the other one?

Question 4

A week after a vaccine, a patient is diagnosed with a condition. A family member says, "If we report this to VAERS, that proves the vaccine did it." How do you correct this?

Serious outcomes: sorting what happened, not deciding why

FDA asks for reports of serious adverse events. "Serious" here describes the outcome for the patient. The reporter needs a suspicion that the outcome is linked to the product, not proof.

The FDA serious-outcome categories are:

• Death: report if you suspect the death was an outcome of the adverse event, and include the date if known.
• Life-threatening: the patient was at substantial risk of dying at the time of the event, or continued use of the product might have caused death.
• Hospitalization, initial or prolonged: the event led to a hospital admission or made an existing stay longer.
• Disability or permanent damage: the event caused a significant, persistent or permanent disruption of the person's normal life functions, body function, physical activities or quality of life.
• Congenital anomaly or birth defect: exposure to the product before conception or during pregnancy may have led to a harmful outcome in the child.
• Required intervention to prevent permanent impairment or damage: listed by FDA for devices. Medical or surgical treatment was needed to prevent permanent harm suspected to come from the product.
• Other serious, or important medical events: the event doesn't fit the categories above but may endanger the patient and may need treatment to prevent one of those outcomes. FDA's examples include breathing trouble from an allergic reaction treated in an emergency room, serious blood disorders, and seizures that did not lead to hospitalization. Developing drug dependence or drug abuse is also an example.

Worked example: the emergency room visit

A patient goes to the emergency room with a suspected reaction, is treated and is sent home the same day.

It is tempting to call this "hospitalization." But FDA says an emergency room visit that does not lead to admission is not in the hospitalization category by itself. It should be evaluated for another category, such as life-threatening or an important medical event.

Change one fact: the patient is admitted overnight. Now the hospitalization category applies.

What these categories are not

These categories sort outcomes for a report. They are not a checklist for deciding whether a patient needs help right now. A patient describing symptoms at the counter needs prompt attention from the pharmacist or emergency help under your workplace procedure, whether or not their symptom appears in any example above. A short list of examples never proves something is not urgent. See Recognize when the pharmacist must intervene.

Also, choosing a category is not deciding cause. You can correctly describe an outcome as "hospitalization" while the question of whether the medicine caused it stays open.

Question 5

A patient already in the hospital for surgery has a suspected reaction to a medicine, and their discharge is delayed by four days. Which category fits?

Question 6

A patient had a seizure soon after starting a new medicine, was treated in the emergency room and went home. A coworker says, "No admission, so it isn't serious." Is that right?

Who reports, what kind of problem, and how to follow up

Who takes part

• Patients and consumers can report to MedWatch themselves. FDA encourages them to take the form to their health care provider, who can add clinical details from the medical record. If the provider won't fill it out, the patient can still report directly.
• Health professionals report to MedWatch voluntarily. FDA states that a health care provider is not required to report to FDA through this voluntary route. FDA relies on these voluntary reports for its safety monitoring.
• Industry reports are mandatory for certain reporters, such as manufacturers, distributors, importers and user facility personnel, in an electronic format FDA can process.
• For VAERS, anyone may report. Health care providers are required by law to report certain events: those listed in the VAERS Table of Reportable Events within the stated time period, and events the manufacturer lists as a reason not to give further doses. Vaccine manufacturers must report all adverse events that come to their attention.

Who actually submits a report from your pharmacy depends on your role, your employer's procedures and your state. This lesson does not settle that.

What kind of product concern

MedWatch accepts reports of four kinds of problem. Each answers a different question:

• Serious reaction: What happened to the patient?
• Product quality problem: Is something wrong with the product itself, such as damaged, contaminated or mislabeled units?
• Therapeutic inequivalence or failure: Did the product fail to work as expected, or fail to perform like a product it was expected to match?
• Product use error: Did a mistake happen while using the product, for example because packaging or labeling was confusing?

A single event can involve more than one. A confusing label (product use error) might also lead to a reaction (serious reaction). For handling products you suspect are defective, see Recalls, tracing and suspect products.

Following up a VAERS report

Each VAERS report is given a VAERS identification number. VAERS sends it to the reporter in a confirmation letter, electronically or by mail, and the number is used to add information to that report later. Without it, new information can't easily be connected to the original report.

Other than the confirmation, VAERS contacts the reporter only if required fields are missing, and those requests come electronically or by mail. For reports classified as serious, VAERS tries to get medical records directly from the institutions that hold them.

VAERS suggests having ready: patient age, date of birth and sex; vaccine brand name and dose; date, time and place of vaccination; when the event started; symptoms and outcome; relevant test results; and the physician's contact information if applicable.

Information that identifies the vaccinated person or the reporter is not made public. Knowingly filing a false VAERS report violates federal law.

Question 7

A patient switched to a different manufacturer's version of a medicine and says it "stopped working" even though they are taking it exactly as before. Which kind of MedWatch concern does this describe, and what does a technician do with it?

Question 8

A clinic submitted a VAERS report last month. Today they received the patient's hospital discharge summary, which adds important details. What do they need to connect this new information to the original report?

From blame to system: reconstruct contributing conditions

When an error happens, the easy conclusion is "someone wasn't careful." Patient-safety analysis looks further. Root cause analysis is a structured way to work out how and why an event happened. Some prefer the name "systems analysis," because there is rarely just one cause.

Two terms help:

• Active error: the mistake made at the moment a person interacts with the system. Plain takeaway: the visible slip, such as grabbing the wrong bottle.
• Latent condition: a hidden weakness built into the system that makes the slip more likely. Plain takeaway: the setup that was waiting for someone to slip, such as look-alike bottles stored side by side.

Several weaknesses usually have to line up before harm reaches a patient. Removing one of them can stop the next event, even if people keep making ordinary human slips.

The analysis follows a sequence: collect information from records and from the people involved, reconstruct the event step by step, and then ask how it happened (active errors) and why (latent conditions). Kinds of contributing factors include organizational and management issues, work environment, team communication, staffing, the task itself and patient characteristics.

Worked example: reconstructing the wrong-strength fill

Return to the 10 mg and 20 mg case. Interviews and records show:

• The two strengths sit next to each other, and the bottles look alike.
• One technician was covering two stations during a rush.
• The barcode scanner at that station had been failing for a week, so staff were skipping the scan to keep up.

The active error was picking the 20 mg bottle. The latent conditions were the storage layout, the look-alike packaging, staffing during the rush, and a broken safety check that had quietly become normal to skip.

Compare two proposed fixes:

• "Remind everyone to be more careful." This is weak. Reminders and re-enforcing existing policy are a common reason improvement efforts fail, because they leave every latent condition in place.
• "Move the 20 mg bottles to a separate shelf with a divider and strength label, and repair the scanner." This is stronger. It changes the conditions themselves, so a tired person under pressure is less likely to make the same slip.

A good proposal names the specific condition it targets. "Separate the strengths" answers "they were stored side by side." "Be careful" does not answer anything in particular.

For more on these error types, see Prevent the wrong-patient and wrong-dose error and High-alert medicines and similar names.

Question 9

After a wrong-patient pickup, the manager's only action is a staff memo saying "Always verify patients." Using this lesson, why is this likely to be weak, and what should the review ask first?

Test one change: Plan-Do-Study-Act

A good idea still needs testing. Plan-Do-Study-Act (PDSA) is a method for trying a change on a small scale, watching what happens, learning from the result and trying again.

Plain takeaway: try one small change, measure it, learn, then adjust or expand.

• Plan: State the one change you will test, the result you hope for and how you will measure it, who is involved, and the time limit.
• Do: Carry out the change and write down what you observe, including surprises and whether everything went as planned.
• Study: Compare the results with what you hoped for.
• Act: Use what you learned to improve the change and test again, or to expand it once it works.

Keep each cycle small: one step, as short as practical, involving only part of the workplace. Each adjustment and re-test is a new cycle. Expand to the whole workplace only after the change works at small scale.

Worked example: testing the shelf change

These numbers are invented for teaching.

Plan: At fill station 1 only, move the 20 mg bottles to a separate labeled shelf for one week. Hoped-for result: fewer wrong-strength picks of this medicine caught at final check. In the four weeks before, final check caught 2 per week on average.

Do: The shelf move happens. Staff mention that reaching the new shelf is slightly slower. On day 3, a delivery is restocked into the old spot next to the 10 mg bottles.

Study: One wrong-strength pick was caught that week, and it came from a bottle restocked in the old spot.

The quick conclusion would be "the shelf change failed." That misreads the test. The change was not fully carried out, because restocking undid it for part of the week. Before judging an idea, check whether it actually happened as planned.

Act: Add the new location to the restocking instructions and run Cycle 2 at the same station.

Changed case: Suppose Cycle 2 runs as planned and shows zero catches for one week. That is encouraging, but one week at one station is a small sample. A good next step is another cycle or a longer one before expanding to every station. Keep watching the speed concern too, because a change that slows work a lot may tempt people into new workarounds.

Question 10

A pharmacy's improvement plan reads: "Starting Monday, all 12 technicians at all three locations will use a new double-check process forever." Using PDSA, what is wrong with this plan?

Practice on your own

Question 11

A label is printed with "take twice daily" when the prescription says once daily. The patient reads the label at home, thinks it seems wrong, calls the pharmacy before taking any doses, and the label is corrected. Classify the event and explain each label you apply or reject.

Question 12

For each case, name the reporting pathway: (a) a prefilled syringe's needle shield cannot be removed; (b) a patient develops a high fever the day after a vaccine; (c) a medicated shampoo causes a severe scalp reaction.

Question 13

A coworker says: "MedWatch reporting is voluntary for health professionals under federal rules, so our state board and our employer can't require us to report anything." What is wrong with this reasoning?

Question 14

A pharmacy has twice dispensed an oral liquid with the wrong measuring device. Review shows both liquids arrive in identical boxes, the devices are kept in one unlabeled bin, and the errors happened during the evening shift, when one technician fills and bags alone. (a) Name one active error and two latent conditions. (b) Propose one focused improvement and design a first PDSA cycle for it.

What to remember

• A medication error is a preventable mistake in the medication process, whether or not it harms anyone. A near miss is an error that caused no harm because of detection or luck. An adverse event is about the patient's undesirable experience, and it can happen with or without an error.
• In FDA reporting, "adverse event" means an experience associated with a product, not one proven to be caused by it. Reports describe suspicions. They are not findings of cause.
• MedWatch is FDA's program for problems with medicines, biologics, devices, combination products and cosmetics: serious reactions, product quality problems, therapeutic inequivalence or failure, and product use errors. VAERS is the CDC and FDA early-warning system for events after vaccination. A VAERS report is not proof the vaccine caused the event.
• FDA's serious-outcome categories describe outcomes for reporting. An ER visit without admission is checked against the other categories. The categories are not triage rules: a patient in distress needs prompt help regardless.
• Anyone can report to MedWatch and VAERS. Some reporters have mandatory duties. Federal voluntary routes do not cancel state or workplace requirements. The VAERS ID links follow-up information to a report.
• Look past the active error to the latent conditions, propose a fix that targets a specific condition, and test it with a small Plan-Do-Study-Act cycle. Before judging the result, check whether the change was actually carried out.

Review: Define the measure before claiming improvement

• Test one change: Plan-Do-Study-Act

Review: Capture persistent impairment or a suspected birth defect

• Serious outcomes: sorting what happened, not deciding why

Review: An emergency visit is not identical to admission

• Serious outcomes: sorting what happened, not deciding why

Review: Voluntary reporting is not limited to a prescriber

• Who reports, what kind of problem, and how to follow up

Review: Seriousness is an outcome classification

• Serious outcomes: sorting what happened, not deciding why

Review: Keep the report identifier for additional information

• Who reports, what kind of problem, and how to follow up

Review: Improve the step, not just the instruction to concentrate

• From blame to system: reconstruct contributing conditions