IN THE APRIL AND MAY issues of Nursing2008, we asked readers to respond to a survey about making and reporting medication errors. This survey was originally published in 2002, and the results were published in the September issue of Nursing 2003. By replicating the survey in 2008, we hoped to learn if practices and attitudes about medication safety have changed over the past 6 years.

Based on 1,296 responses to the 2008 survey, we noted many positives. For example, most nurses realize the importance of medication safety and error reporting. However, some risky behaviors still need to be addressed based on specific work settings or clinical populations.

This statement is false; error reporting is a tool for evaluating weaknesses in the entire medication administration process, not for assessing a nurse's competence. Responses to this question are identical to those given in the 2002 survey, which tells us that more than half of nurses still misunderstand the true value of error reporting.

By analyzing error reports, healthcare organizations focus on determining how an error occurred, not who made it. This paves the way to a better understanding of what processes need to be changed to prevent errors.

Evaluation of a nurse's competence shouldn't be based on medication errors. Anyone can make an error, even the most competent, careful nurse. Yet approximately three-quarters of respondents working in home healthcare or long-term-care facilities identified error reporting as a tool for measuring competency. Nurses working in specialty areas such as the ED, oncology, orthopedics, and intensive/critical care were more likely to recognize that medication administration competency should be assessed using other tools.

This statement is false. The vast majority of errors are multifaceted and the result of system failures.

Nurses agree that following the five rights of medication administration helps prevent errors, even more strongly now than in 2002. However, the five rights are still heavily valued as a tool to prevent all errors. Errors can still occur even when nurses diligently follow the five rights. Sometimes other system failures "downstream"-such as during the ordering, transcribing, and dispensing process-precipitate errors.

With the greater use of various tools and technology, such as bar coding, preprinted medication order forms, and computerized prescriber order entry, many potential errors have been decreased or eliminated. Observing the five rights can't prevent every error. For example, an error could occur if the lab reports a critical blood glucose level for the wrong patient or if a prescriber orders a cardiac medication for the wrong patient. In each instance, you could follow the five rights to a T, yet still give the wrong medication.

About a third of the responses to this statement reflect the continued sense, real or perceived, that disciplinary action may result either from the employer or other organizations (such as the state board of nursing) if a nurse reports an error. Surprisingly, we found a 13% increase in nurses believing this is true among those with less than 1 year's experience, compared with responses from this group in 2002. This implies that either healthcare organizations don't promote a just culture or that they're not fully educating these new nurses about the value of error reporting during orientation. The good news is that nurses working in long-term care, outpatient services/clinic, and offices and those with advanced degrees see the importance of never failing to report errors.

Healthcare facilities need to do a better job of educating nurses and other members of the healthcare team about a just culture that will help address underlying causes of the error. By analyzing the medication administration process, healthcare facilities that embrace a just culture model shift the focus from human to system failures by addressing the processes within the system that are broken. Accurate reporting of errors, potential errors, and near misses helps healthcare organizations identify, plan, and implement strategies to prevent similar errors in the future.

Adverse event (incident) reports don't reflect a nurse's true competence and shouldn't become part of the personnel file. Circumstances such as extreme negligence, chemical impairment, or intent to harm should be documented in the personnel file with other tools. We're glad to see that the perception (or, hopefully, the reality) that these reports go into the nurse's personnel file has decreased since 2002.

Overall, about 63% of nurses with less than 1 year's experience in 2008 reported that event reports go into their personnel file, compared with 46% of those with more than 1 year's experience. This suggests that less-experienced nurses may not really know what happens to these reports. But a significantly higher percentage of those working in long-term-care areas (68%) and LPNs (80%), who are the primary caregivers in long-term-care facilities, also stated that event reports were placed in their personnel files. This could indicate that many nurses lack knowledge about what happens to these reports or that long-term-care facilities lag behind acute care facilities in this policy area.

In 2008, an overwhelming majority of nurses (95%) identified this as a true statement, indicating that they understand the importance of analyzing errors that have been made. This perception has increased over the past 6 years. However, based on responses to other statements within this survey, it's evident that nurses struggle with consistently implementing the practice of error reporting because of organizational barriers. We also need to emphasize the importance of reporting dangerous situations, potential errors, and near misses. Along with actual errors that reach the patient, nurses need to be more proactive in reporting all events that might have led to an error.

Most nurses agree that the use of technology will decrease medication errors, which is true to a degree. But the perception that instituting technology can reduce all errors is highly inflated. A computer is just a tool, and the information it provides is only as good as the information entered into the system. Nurses still need to compare the information available in the computer database with the patient's condition and other information about his health status. Nurses will always be that final safety check before the patient takes the medication. Nurses must also avoid at-risk behaviors related to technology, such as overriding alerts and using work-arounds-for example, duplicating a patient's ID band and placing it on the clipboard to scan the bar code.

According to our results, nurses with less than 1 year's experience agreed with this statement 88% of the time; more-experienced nurses, especially those with more than 16 years of experience, agreed with this statement only 79% of the time. This may imply that newer nurses are more confident that technology can prevent errors. Nurses with more experience seem to be wary of technology's ability to decrease errors.

Although two-thirds of the survey respondents said they "never" bypass built-in safety features, one- third reported they "sometimes" do so. The clinical settings where nurses were most cautious were home healthcare, long-term care, outpatient services/clinic, and office settings; 83% to 92% of respondents from these settings said they "never" bypass safety features. The likelihood of taking shortcuts increased in the intensive care/critical care areas and in pediatrics, where 46% and 42% of respondents (respectively) reported "sometimes" bypassing safeguards. This could reflect the need to administer medications quickly in critical situations. Unfortunately, bypassing safeguards can have particularly devastating outcomes in critical situations. Nurses must be diligent in applying all safety measures and basic fundamental skills consistently when administering medications, regardless of the situation or setting. If nurses are routinely bypassing safeguards, the healthcare organization needs to look at the root causes and address those issues.

Here we see a slight shift in the willingness of nurses to report errors when other practitioners are involved. This implies that nurses are only slowly beginning to realize the value of reporting errors and are becoming more comfortable reporting errors they find.

In general, our findings show that nurses with less than 1 year's experience are the least likely to report errors. This is most likely due to inexperience, a lack of self- confidence, and fear of being intimidated and stigmatized by coworkers.

In the past 6 years, significant improvements in reporting errors made by all practitioners have occurred in several areas, including home healthcare, long-term care, and outpatient services/clinic areas. Nurses working in surgical services and postoperative areas also reported significant increases in their willingness to report errors made by pharmacists and physicians.

Although these responses are an encouraging sign that reporting errors is increasing, much room for improvement remains. Nurses need to become more comfortable in reporting any error they find, including their own. The stigma associated with making an error must be overcome and a just culture approach cultivated in all care settings. Nurses must remember that by reporting errors, they're addressing the error itself, not who made it.

Every facility should have a disclosure policy to outline who's responsible for informing the patient and family about an error, who else needs to be informed about the error, and how the error and disclosure should be documented. Almost 40% of respondents reported that their facility has such a policy, but nearly half said they don't know. Although having a policy in place is recommended by The Joint Commission (TJC) and other organizations, many staff members remain unfamiliar with their facility's policy or the disclosure process unless they've been involved in an error.

More than 53% of respondents working in outpatient services/clinic areas reported they don't know if such a policy was in place. However, almost 60% of those working with geriatric patients were aware of such a policy. This can be correlated not just to disclosure of medication errors in the geriatric setting but also to policies for disclosing patient falls, a common hazard in this age-group.

Despite progress in this area, more discussion and education needs to take place. Nurse leaders need to ensure that their facilities develop a policy and communicate the process to all members of the staff during orientation-before an error occurs, not after the fact. Further discussion at staff meetings and role playing can familiarize nurses with the policy and help them become comfortable with it.

Here we see a shift from the 2002 survey, in which 31% of respondents stated they "never" disclosed errors. In 2008, only 25% said they never do so, suggesting a growing willingness to own up to mistakes. Even so, only 27% of respondents in 2008 say they "always" disclose errors. We'd like this number to be higher, but it remains low for various reasons: the stigma associated with disclosing an error, the lack of a disclosure policy, or certain organizational constraints. For example, a policy that requires managers to disclose errors impedes self- disclosure. This is another area that needs improvement.

Nurses over age 21 who've been in nursing for more than 1 year are becoming more likely to report their errors or mistakes directly to the patient or family. But a closer analysis of the data reveals that practice setting also plays an important role in determining whether a nurse is likely to disclose errors. For those who work in a hospital, for example, only 23% reported that they "always" disclose the mistake. On the other hand, in home healthcare and long-term-care facilities nurses responded they "always" disclose errors 54% and 43% of the time, respectively. This may be due to the closer relationships nurses establish with patients and family when they work with them daily in a long-term-care facility or regularly in the home healthcare setting, as well as policies and procedures established by their organizations.

If a nurse must take an oral or telephone order, she must follow these steps to ensure accuracy:

* Obtain the patient's medical record and write down the order while still on the phone with the prescriber.

* Confirm the patient's name.

* Read back the name of the drug; if unfamiliar with the drug name, ask the prescriber to spell it.

* Confirm the dose by stating each number individually (for example, "two five," not "twenty-five").

* Ask for the drug indication or communicate an understanding of the intended purpose and make sure that the order makes sense based on the patient's condition. For example, hydrALAZINE, which is used to treat hypertension, can easily be confused with hydrOXYzine, which is used to treat anxiety. When taking an oral order in the presence of the prescriber, hand the medical record to the prescriber to write down the order except in special circumstances (such as when sterile conditions must be maintained). Never take oral or telephone orders for chemotherapy.

In 2002, only 44% of respondents said they never use the abbreviation U for units. In 2008, this percentage swelled to 84%.

In looking at practice settings and clinical areas, we learned that hospitals in particular have done an outstanding job in eliminating this abbreviation. Almost 88% of nurses working in hospitals responded that they "never" use it. However, those in the long-term-care setting and geriatric areas responded that 33% and 27% of the time, respectively, they still abbreviate units "sometimes" or "always"!! Older adults and patients in long-term-care facilities are frequently prescribed medications that are measured in units, such as insulin and heparin. Nurses who practice in these areas need to eliminate the use of U as an abbreviation for units because the potential for error is extremely high.

When administering "high-alert" drugs, I have another practitioner independently double-check my work.

In 2002, this question was asked generally about opiates, concentrated electrolytes, anticoagulants, and insulin; in 2008, responses were specific to three high-alert drugs: opioids, heparin, and insulin. Here's how the results compare.

High-alert drugs are defined as drugs that, when misused, can cause serious injury or death. Because the consequences of misuse can be devastating, practitioners should routinely conduct independent double checks before administering these drugs.

In 2002, 58% of respondents said they "always" have another practitioner double-check their work. When this study was replicated in 2008, we altered this question slightly to ask specifically about opioids, heparin, and insulin. Surprisingly, when we averaged the responses to these three categories, we found similar responses to the first choice: 58% stated they "always" have another practitioner check their work. But we were disturbed by the rising percentage of 2008 respondents who say they "never" employ an independent double check for high-alert drugs: 16% in 2008, compared with only 5% in 2002.

This 11% increase in nurses who never use double checks for these drugs surprised us. With numerous recommendations about the importance of double-checking high-alert drugs, we expected this practice to have changed dramatically in the other direction. Also, two-thirds of the respondents stated that they "always" have another practitioner check their work for heparin and insulin, but only 42% use this method when administering opioids.

Because we approached this issue differently in 2008 than in 2002, the correct answer in 2008 is "always" and in 2002, "never."

Oral liquid medications that must be drawn up in a syringe should always be administered with a syringe specifically designed for oral medications, not with a parenteral medication syringe, spoon, or other device. Using a parenteral medication syringe has resulted in oral medications being given I.V.

Because 51% of 2008 respondents stated that they use oral syringes only "sometimes," we wonder if other factors may contribute to the persistence of this risky practice. For example, oral syringes may not be available in certain practice settings.

We did find that 57% of the nurses working in pediatrics and 52% of nurses working in home healthcare "always" utilize oral syringes to administer oral medicine. Failing to use oral syringes every time is a significant problem, especially in pediatric settings where accurate dosing is critical. However, further improvements are necessary in all clinical and practice settings. Easy, low-cost practice changes that would significantly improve patient safety include having pharmacy dispense oral medications in syringes and stock oral syringes in all patient-care units.

The practice of "borrowing" a drug from another patient's bin or the automated dispensing machine increases the risk of a medication error. Sometimes the patient's medication is missing for a reason. For example, the pharmacist wants to clarify the order with the prescriber before dispensing, an automatic stop order for the drug was entered, or another nurse already gave the medication to the patient and didn't document it. If a drug is missing, always investigate why before administering the medication. Rule of thumb: No borrowing, no stashing.

Comparing responses in 2002 and 2008, we're glad to see some improvement in this practice. Even so, 43% of respondents stated that they "sometimes" borrow a missing medication from another patient's bin or the automated dispensing machine. Clearly many nurses still need more education about why this is an unsafe practice.

Nurses must carefully check for drug allergies before administering any new medication to a patient. Almost three-quarters of 2008 respondents say they "always" follow this critical safety process, indicating a slight improvement in practice since our original survey.

Many facilities have implemented computerized prescriber order entry and pharmacy computer systems that incorporate reviewing and verifying allergies in the initial entry screens before medications can be cleared for retrieval from the automated dispensing machine or dispensed. But nurses must also again review and verify drug allergies in case the step was missed. When administering a new medication, ask the patient again if she has any allergies.

Nurses must compare the patient's MAR against two patient identifiers before giving medication.

When identifying a patient by name, don't ask a yes-or-no question such as, "Are you John Smith?" Instead, ask him to state his name.

We're glad to see that over three-quarters of respondents (77%) stated that they "always" utilize two identifiers, which is a 20% increase since our 2002 survey.

Over the course of a shift or after working with the same patient a few days in a row, a nurse may become familiar with the patient and stop following this practice. But sticking to it every time is essential because distractions and mix-ups can easily occur.

One area where nurses' comfort level gets in the way of patient safety is in long-term-care settings. Only 70% of respondents working in long-term care said they "always" check two patient identifiers against the patient's MAR-the lowest percentage among all the work settings surveyed. Most likely, this is because patients may spend months to years in the facility and nurses regard them as residents, not patients.

We were also disturbed by responses from respondents working in pediatrics. About 33% stated that they "sometimes" or "never" use two identifiers to check their patients' identity before medication administration. With infants and children, a high-risk population, we need to be more, not less, vigilant.

We're very pleased to see the 30% increase in responses from nurses stating that they "always" take the MAR into the patient's room when they administer medication. This very important safety measure gives nurses another opportunity to double-check that they're giving the correct medication to the right patient and also serves as reference at the point of care. Most likely, this change in practice is related to advances in technology and the wider use of handheld scanners and bar coding.

However, improvement is needed in several areas. Among nurses working in long-term-care settings, 56% said they don't take the MAR to the bedside. And very surprisingly, only 60% of nurses who work with pediatric patients responded that they bring the MAR to the bedside only "sometimes" or "never." Again, we can't overemphasize the importance of accuracy and safety in such high-risk populations as pediatrics.

By opening a medication package only when you arrive at the bedside, not before, you can eliminate the potential for numerous errors. We're pleased to see the dramatic shift in practice over the past 6 years, with 24% more nurses now reporting that they "never" remove medication from the unit-dose package before entering the patient's room.

Long-term-care facilities may represent an exception to this trend: More than 67% of those working in this practice setting reported that they "always" open the unit-dose package before entering the patient room. Responses from this practice setting were about the same in 2002; we're unsure why. We know that some long-term-care facilities use "bingo cards" (multidose packs of medication tablets). Nurses working in settings without unit-dose packaging can't follow best practices for medication administration. In the time a nurse opens a medication package, puts it in a cup, and delivers it to the patient, he could become distracted and give it to the wrong patient.

But the most likely reasons this practice setting lags behind are probably the lack of technology implementations in long-term-care facilities or basic work-flow issues. We recommend that these organizations look into their medication administration practices and identify ways they can improve the process.

Best practice is to take the MAR into the patient's room and document medication administration immediately after administering it. This removes the risk of erroneously documenting something you didn't do (for example, because you got called away or the patient refused the medication) or failing to document medication administration after the fact.

In general, the respondents in 2008 answered in a similar fashion to those responding in 2002. We recognize limitations to the organizational work flow in some facilities contribute to documentation practices remaining the same, but with such technology advances as electronic MARs and bar coding, the practice of documenting medication administration during the process will be easier to achieve. We encourage healthcare administrators to push their organizations to implement these technology systems.

Figure. No caption available.

Pushing for safer practices

Results from this survey show movement toward safer medication practices. Contributing greatly to this encouraging trend is a greater push for safer medication practices by regulatory agencies such as the FDA, standards organizations such as TJC, nonprofit medication safety groups such as the ISMP, the Institute for Healthcare Improvement, the National Quality Forum, and others.

In 2003, the implementation of TJC's first set of National Patient Safety Goals emphasized many processes associated with safer medication practices. To meet these goals, many facilities have begun to implement significant practice changes in the medication use process.

Progress has also been made in implementing special precautions for high-alert drugs, such as differentiation (tall-man letters, overwraps, purchasing from different manufacturers), package reminders (for example, highly concentrated or for oral use only), and system redundancies such as independent double checks.

But much remains to be done, as hundreds of nurses reminded us in written comments submitted with their responses (see Nurses speak out). As nurses, we can never stop striving to make changes that help keep our patients safe. We hope you'll use information from this survey to address the risks at your facility and in your own practice so we can continue to reduce medication errors.

Respondent snapshot

The mean age of survey respondents increased by approximately 3 years between the 2002 and 2008 surveys, to age 45. Even more interesting, 40% of respondents in 2008 were over age 50, compared with 25% of respondents 6 years ago. In addition, only 10% of respondents fell into the 21-to-30 age-group, compared with 17% 6 years ago. These findings correlate with current trends in nursing employment: More older nurses are remaining in the workforce well into their 50s and fewer young people are entering the profession.

Here's a sketch of a typical respondent to the 2008 survey:

* clinical setting: hospital (75%)

* clinical practice area: medical-surgical (33%) or intensive care/critical care (16%).

* years of experience: 16 or more (55%)

* education: BSN/BS degree (36%) or AD (23%).

About this survey

A limitation of this study is that respondents were self-selected. Because we didn't survey a predetermined sample, we can't say how well participants represent nurses in general. Participants responded based on their personal experiences, on policies and procedures within their organization, and on resources available to them. In some cases, nurses working in management, education, or advanced practice roles may have based their responses on how they believe nurses should practice rather than on what actually occurs. Controlling for these variables was beyond the scope of this study.

Nurses speak out

Along with survey responses, we received hundreds of detailed and passionate written comments from nurses worried about medication errors. As one nurse succinctly noted, "Making a medication error is one of my biggest fears." A recurring theme was how stress, fatigue, and understaffing raise the risk of errors in their work environment every day. Many noted that the stigma associated with making an error still strongly influences whether they report an error or near miss. Here's a sampling of comments:

* "Despite computerized prescriber order entry systems, MDs are still allowed to handwrite their orders while nurses are left to interpret and enter the orders."

* "Better (and mandatory!!) nurse-to-patient staffing ratios would help ensure that a nurse has time to do all five rights of medication administration. Sometimes all he or she has time for is two or three rights."

* "I have six to eight pediatric patients with I.V. meds. Last week I inadvertently missed one patient's I.V. antibiotic dose."

* "I work with a computerized system. All it does is change the type of errors. It may decrease some errors, but it takes three to five times longer to administer the medications."

* "I work in a hospital that has instituted scanner bars and scanners to dispense medications. This has dramatically reduced our medication errors. I really like this system."

* "Our unit director discourages us from writing event reports on our fellow employees. She states, 'We should work as a team.'"

* "People let machines do the thinking for them and sometimes ignore built-in checkpoints. Technology doesn't replace clinical judgment and safety protocols."

* "I've had to wait 2 to 3 hours for a missing med. It takes so long to fill out the necessary paperwork, then wait for it to be delivered."

* "Many errors by physicians aren't considered errors because they're caught by pharmacists and nurses. When a nurse makes a mistake, she's all alone."

* "In my facility, it sometimes seems as though nursing and pharmacy are on opposing teams. A true QI [quality improvement] process rather than an 'I gotcha' attitude would motivate nurses to report errors more frequently."

References

Cohen H, Robinson ES, Mandrack M. Getting to the root of medication errors: Survey results. Nursing. 2003;33(9):36-45.

Cohen M. Medication Errors. 2nd ed. Washington, DC: American Pharmacists Association; 2007.