MEDICATION SAFETY. December 2010 and the application has been requested to be given priority. [Australian Incident 117, December 2010]

Transcription

1 MEDICATION SAFETY This series brings you up-to-date information about medication safety issues and strategies to prevent medication errors. It draws on Australian incidents and US experience, including (with permission) material from ISMP Medication Safety Alert! a bulletin published by the US Institute for Safe Medication Practices < This series is coordinated via the Committee of Specialty Practice in Medication Safety (Chair, Rosemary Burke, Director of Pharmacy, Concord Hospital, NSW). Australian incidents are collated and editorial recommendations made by Penny Thornton (Principal Advisor, Medication Safety, NSW Health; <medsafety@shpa.org.au>). AUSTRALIAN INCIDENTS Little green and white bottles! Two patients were admitted to hospital suffering from over-anticoagulation. One patient was found to have had dispensed Coumadin (warfarin) 5 mg instead of Coversyl (perindopril arginine) 5 mg at a community pharmacy. On reconciling the second s patient s own medicines, a hospital pharmacist found Coversyl and Coumadin tablets mixed in the same bottle when reconciling the second patient s medications. The problem relates to similarities in packaging used for Coumadin and Coversyl both come in white bottles with green lids and with Figure 1. Packaging similarity of Coumadin and Coversyl. green, black and white colouration on the manufacturers labels (Figure 1). Metoclopramide (Pramin) and amiloride (Kaluril) are also packed in a similar manner. Instances of this mix-up have occurred in several community pharmacies. Community pharmacies tend to store medications by alphabetical brand names and Coumadin and Coversyl are likely to be stored adjacent to one another in the dispensary. The chance of mix-up within hospital pharmacies is reduced as these products are stored by alphabetical generic names. Warfarin is a highrisk medication and the seriousness of these types of mix-ups cannot be underestimated. What can be done? Consider storing these products in a manner that reduces the risk of wrong product selection. Review the perindopril brand held in your pharmacy s inventory. To proactively avert this type of error, some pharmacies stock perindopril erbumine (APO- Perindopril) 4 mg instead of the Coversyl brand. Consider using alert stickers to ensure correct replacement of stock removed from pharmacy or imprest stock shelves. This includes other products that are similarly packaged. Continue to be vigilant regarding look-alike (and sound-alike) medicines and the potential for error. Risks with dispensing Coumadin/Coversyl have been communicated to community pharmacists via the Pharmaceutical Defense Limited. This issue will also be covered in the Pharmacy Guild s communication to members and in other pharmacy publications. Community pharmacists are being urged to store Coumadin under W and to use their scanners. The Australian Commission on Safety and Quality in Health Care has advised that the TGA has received an application for packaging changes from Servier (manufacturer of Coversyl) in mid December 2010 and the application has been requested to be given priority. [Australian Incident 117, December 2010] Care with chloral hydrate (again!) An incident has been reported involving a child who drank a full bottle of chloral hydrate syrup. This was a fatal overdose. Chloral hydrate bottles do not have childproof caps and until this situation is rectified, consider recapping all bottles with a child-proof cap at the point of dispensing, if possible. Always alert parents/carers to the need for safe storage to avoid children accessing what might be considered a nice flavoured drink. [Australian Incident 118, December 2010] US SAFETY BRIEFS Update on metal in drug patches An updated list of transdermal drug delivery systems with a metallic component was recently published (Hong I, et al. Hosp Pharm 2010; 45: 771-8l.). Serious burns may occur in patients undergoing magnetic resonance imaging (MRI) who have transdermal patches with metallic content applied to their skin. The metal acts as a conductor of radiofrequency pulses, inducing electric current that causes intense heat and burns. The US Food and Drug Administration (FDA) and ISMP issued an alert about burns from medication patches known to contain a metallic component in 2004 < acutecare/articles/ asp>. In 2009, the US FDA issued a public health advisory regarding removal of drug patches prior to MRI <www. fda.gov/newsevents/ Newsroom/PressAnnouncements/ucm htm>. The latest review, by pharmacists from the University of Illinois informs on the metallic content of available transdermal patches and recommendations regarding reapplication of patches after a scan. [Medication Safety Alert! 18 November 2010] Australian comment: Although the patches available in Australia may differ from those in the US, consider a blanket rule: Remove all patches before MRI even though some may cause no problems. Concerns are that the patch may not be replaced after the MRI or may be reapplied but has lost its stickiness. Allow for an extra unit usage if the patient has had an MRI as a new patch may need to be applied. Some dressings with a metallic thread may also cause the same burning effect. Top drugs that cause violence The link between violence and prescription drugs is the subject of a new study (Moore T, et al. PloS ONE. December 2010>. The study identified 31 drugs that are disproportionately associated with increased rates of reported violence. Case reports were obtained from the ISMP QuarterWatch database and composed of computer 52

2 extracts of all adverse drug event reports received by the US FDA. Inclusion criteria were reports of violence and aggressive behaviour from drugs that were submitted to the US FDA from 2004 through the third quarter of During the 5-year study period, 484 drugs accounted for serious adverse event reports; of which 1937 (0.25%) were cases of violence. Thirty-one drugs met the criteria for a disproportionate association with violence, i.e of the 1937 (79%) cases. Included were 387 reports of homicide, 404 reports of physical assaults, 223 cases of violence-related symptoms, and 896 reports of homicidal ideation. The drug that accounted for the most reports of violence was Champix (varenicline). Also associated with violence were two selective serotonin reuptake inhibitors used to treat depression, FLUoxetine (Prozac) and PARoxetine (Aropax, Extine, Paxtine, Paroxo), the attention deficit hyperactivity disorder drug atomoxetine (Strattera), amphetamines and the benzodiazepine triazolam. These data provide new evidence that violent acts are associated with a relatively small group of drugs, and that systematic studies are needed to establish the incidence, confirm differences among drugs, and identify additional common features. [Medication Safety Alert! 13 January 2011] documenting administration at the scheduled time, but actually giving the drug early or late; altering drug administration schedules to avoid documenting late administration; pre-pouring medications ahead of time to accelerate the drug administration process; skipping double-checks due to time constraints; failing to use barcode scanning to avoid documenting late administration; and leaving medications in the room for the patient to take at the right time. These and other at-risk behaviours, which are unintended consequences of attempts to comply with the CMS 30-minute rule have led to harmful errors. On the other hand, for some critical patient populations, certain first doses and time-critical scheduled medications (e.g. rapid-acting insulin), delays in administration can result in harm. Therein lies the dilemma faced by nurses and pharmacists responsible for dispensing and administering medications in a timely and safe manner. A one-size-fits-all, inflexible requirement to administer all scheduled medications within 30 minutes before or after their scheduled time rule is ill-suited for the majority of medications that are not time critical. A rigid, standard requirement also does not encourage or allow health professionals to use critical thinking to prioritise patient care or to manage the unexpected, which is commonplace in healthcare. In a setting of competing priorities, administration of non-time-critical medications may be lower on the list of priorities relative to other more crucial time-sensitive clinical tasks. Since publication of the survey results, ISMP has been working towards documenting what we believe would represent best practices associated with timely administration of scheduled medications. Scheduled medications include all maintenance doses administered according to a standard, repeated cycle of frequency (e.g. q4h, QID, TID, BID, daily, weekly, monthly, annually). For the purpose of ISMP s guidelines, scheduled medications do not include: stat and now dose; first doses and loading doses; one-time doses; specifically timed doses (e.g. antibiotic for surgical patient 10 minutes before incision); on-call doses (e.g. pre-procedure sedation); time-sequenced or concomitant medications (e.g. chemotherapy and rescue drugs, n-acetylcysteine and iodinated contrast media); drugs administered at specific times to ensure accurate peak/trough/serum drug levels; investigational drugs in clinical trials; and prn medications. The Guidelines for Timely Administration of Medications represent a synthesis of ISMP expertise, review of available literature on the risks associated with early and delayed administration of maintenance drug therapy, and advice from survey participants and an expert clinical advisory group of more than 20 clinicians many of whom are practising in healthcare settings. The guidelines may seem detailed but timely medication administration is a multifaceted issue that cannot be managed appropriately with a single standard. We also plan to interact with CMS to discuss possible alterations to their guidelines to best facilitate both timely and safe drug administration. The Guidelines for Timely Administration of Medications (summary in Table 1) recommend: Guidelines for timely medication administration In our 9 September 2010 newsletter < letters/acutecare/articles/ pdf> we discussed the results of our survey regarding the requirement in the Centers for Medicare & Medicaid Services (CMS) Conditions of Participation Interpretive Guidelines to administer medications within 30 minutes before or after the scheduled time < som107ap_a_hospitals.pdf>, which directs surveyors to determine whether drugs [are] administered within 30 minutes of the scheduled time for administration. The survey elicited responses from almost nurses. Respondents made it clear that changes to drug delivery methods and gradual increases in the complexity of care, number of prescribed medications per patient, and number of patients assigned to each nurse have made the long-standing CMS 30-minute rule troublesome. Many nurses felt the rule was unsafe, impossible to follow, and often unnecessary from a clinical perspective. Aware that nurses were struggling to comply with the rule and that surveyors were sometimes citing hospitals for noncompliance, we conducted the survey to prepare for interaction with CMS leaders to discuss our concerns regarding the rule s unintended effects on patient safety. Preliminary interaction has taken place, and CMS has indicated openness to suggestions for improvement. Respondent nurses reported feeling pressured to take these and other short cuts to comply with the CMS 30- minute rule, such as: preparing IV solutions on the unit, and bypassing pharmacy verification and specialised aseptic techniques; administering medications before conducting a physical assessment of the patient and/or checking vital signs, lab values, weight, and allergy status; removing medications from automated dispensing cabinets via override without a pharmacist s review; borrowing medications from one patient to administer to another patient; not thinking critically about drug administration while rushing through verification of new orders and actual drug administration; 53

3 1. Identify a hospital-specific list of time-critical scheduled medications (those where early or delayed administration of maintenance doses of greater than 30 minutes may cause harm or result in substantial suboptimal therapy or pharmacological effect). While the list of time-critical scheduled medications will include a limited number of drugs, a universal list is not desirable because hospitals that treat different patient populations (e.g. mental health, oncology, pediatrics, premature infants) may need to include different medications to address respective risks. Similarly, some hospitals that serve very diverse patient populations may decide to identify both hospital-wide and unit-specific time-critical medications. Examples of time-critical scheduled medications include: immunosuppressives used to prevention solidorgan transplant rejection or to treat myasthenia gravis; and drugs that require administration within a specified short period of time before, after, or with meals, e.g. rapid-, short-, or ultra-short-acting insulins, certain oral antidiabetics (e.g. acarbose, repaglinide, glimepiride), alendronate and pancrelipase. Additionally, because some scheduled medications can be time-critical for certain patients given their diagnoses (e.g. scheduled parenteral anti-infectives for worsening sepsis), the list may include drugs that are time-critical only when used for a specific diagnosis or indication. Further, policies should allow prescribers, pharmacists, or nurses to declare any medication to be time-critical (i.e. must be given within 30 minutes before or after the scheduled time) by including this designation with the medication order and/or medication administration record (MAR) entry. 2. Establish guidelines that facilitate pharmacy order review, dispensing, and nurse administration of the hospital-identified, time-critical scheduled medications within 30 minutes before or 30 minutes after the scheduled time (or more exact timing when indicated, as with rapid-, short-, and ultrashort-acting insulin). MAR entries for hospital-identified timecritical scheduled medications should be clearly designated to remind staff that these drugs require meticulous attention to timely administration (e.g. administered at the designated time, or plus or minus 30 minutes from the scheduled time, if appropriate). 3. Establish guidelines for timely drug administration of scheduled non-time critical medications as follows: administer daily, weekly, or monthly medications plus or minus 2 hours from the scheduled time. (It is generally safe to administer daily/weekly/monthly medications within a timeframe that exceeds 2 hours, ISMP recommends keeping the timeframe to 2 hours before or after the scheduled time to prevent accidental omission of doses that might be more easily forgotten if delayed more than 2 hours.) administer medications given more frequently than daily but not more frequently than every 4 hours (e.g. BID, TID, q4h, q6h) plus or minus 1 hour from the scheduled time. administer medications given more frequently than every 4 hours (e.g. q1h, q2h, q3h) within 25% of the dosing interval (e.g. ± 15 minutes for hourly doses, ± 30 minutes for every 2 hours dosing, ± 45 minutes for every 3 hours dosing). As an alternative to the guidelines in item 3 (above), hospitals with emars may choose to establish a standard allowable percent of difference between the scheduled and administration times not to exceed 25% for all non-time-critical medications if the emars can alert nurses in advance to an impending late dose and the allowable timeframe for administration. 4. Although not associated with the timing of scheduled medications, hospitals should also define targeted timeframes for administering first doses and loading doses of key medications, such as IV antiinfectives, IV anticoagulants, and IV antiepileptics where timeliness is critical (e.g. ED patient with suspected sepsis should not wait several hours for administration of a prescribed anti-infective). Where electronic prescribing systems are available, the prescriber can also be queried regarding when the first dose should be administered given the standard administration times. While timely administration of first or loading doses of these drugs may be critical, many are not necessarily time-critical when it comes to subsequent maintenance doses. The targeted timeframes for first or loading doses of medications should be accompanied by procedures that facilitate achievement of the administration time goals. [Medication Safety Alert! 13 January 2011] Adjust Pradaxa dose for renal impairment We hope health professionals are paying as much attention to safe prescribing of the oral direct thrombin inhibitor Pradaxa (dabigatran etexilate mesylate) as they would warfarin. In a recent report, Pradaxa was prescribed by a cardiologist for a patient with Stage IV chronic kidney disease. The dose was not adjusted for the patient s renal impairment. Instead of the recommended dose of 75 mg BID, the patient received 150 mg BID. On a follow-up visit with the cardiologist, the patient complained of weakness, black tarry stools, and was found to have a low haemoglobin and haematocrit. The cardiologist immediately sent the patient to a local ED. The patient was admitted to the intensive care unit with a massive gastrointestinal bleed and given multiple units of fresh frozen plasma and packed red blood cells. His haematocrit and haemoglobin returned to safe levels, allowing transfer to a progressive care unit. Among the precautions for Pradaxa are several drug interactions and a recommended dose of 75 mg BID for patients with a creatinine clearance of 15 to 30 ml/min. Dosing recommendations for patients with a creatinine clearance below 15 ml/min or on dialysis are not provided on the label. 54

4 Nicotine nitroglycerin patch mix-up A pharmacist told us about two incidents in which nicotine transdermal patches were dispensed to patients instead of nitroglycerin patches. The patches were stored alphabetically by generic name in the pharmacy where technicians manually pulled the medications; they were not dispensed by an automated robot. The nicotine and nitroglycerin patches wound up side-by-side and were in similar-looking packages. Since the incident, the pharmacy has changed and re-labelled the shelving locations of these products and made staff aware of the newly located stock. Nicotine patches kept in some clinical areas have also been removed since providing nicotine in place of nitroglycerin, or nitroglycerin in place of nicotine, is likely to cause serious adverse effects. the expected dose in 16% of syringes, suggesting that unrecognised medication errors may be a major source of morbidity and mortality in resuscitated patients, including children. The types of errors that occur during codes are varied. However, numerous studies have suggested that the most common types of errors that occur during codes are dosing errors, drug selection errors, drug preparation errors, administration technique errors and omissions. Around a quarter of these errors originate during dispensing/preparation of the drug; about half originate during administration; and about 10% occur during prescribing. In addition, a study of mock pediatric resuscitations showed that approximately 1 in 5 verbal orders did not specify an exact dose and around half did not include the route of administration. The wrong dose prescribing errors that occurred during this study included 10-fold overdoses, total daily doses Preventing medication errors during codes Problem: Cardiac and/or respiratory resuscitation is an extremely stressful situation during which health professionals have little time for discussion and verification of the patient s treatment plan, including medications. During these medical emergencies, often referred to as a code blue every second counts and every ordered as a single dose, and ordering the wrong concentration of dextrose for an infant. Furthermore, around 12% of the ordered drugs in this study were never administered. Examples of actual errors that have been reported to the ISMP Medication Errors Reporting Program or errors published in the referenced journal articles are provided in Table 1. Most often, these errors errant action or inaction can result in patient harm or involve reliance on human calculations, death. While studies on medication errors during codes have documented variable incidence rates from less than 1% when studying reported events to 15% from observations during simulations two facts are clear: patient harm from medication errors during code conditions is high. Serious errors are often missed, even if the patient died or is harmed, because there was no suspicion that the drugs caused or contributed to the morbidity or mortality. One study documented that medication errors during codes are 39 times more likely to result in harm and 51 times more likely to result in death than non-code related medication errors. This finding is not surprising because many of the drugs administered during medical emergencies are high-alert and patients are at their most vulnerable. Another study analysed the actual drug content in syringes used during code simulations and found substantial deviations from miscommunication, protocol deviation, knowledge deficit, or a dispensing device problem. Commonly reported contributing factors include: look-alike packaging or drug names; disorganised and non-standard code carts; excessive stock in code carts; distractions caused by an hectic environment; poorly communicated verbal orders; inexperienced staff; alternative drugs in code carts during a drug shortage; confusing or missing information about drugs; and multiple concentrations of a drug in code cart drawers. A 2008 study also highlighted that patients involved in codes are not the only ones who may be affected by medication errors. Other patients assigned to staff who attend the code may also be victims of errors, i.e. collateral damage. These medication errors, which were rarely harmful, were mostly missed respiratory treatments that were supposed to be administered by respiratory therapists who were busy attending a code. Safe Practice Recommendations: Ideally, reducing the incidence of emergency resuscitations would simultaneously reduce the incidence of errors during codes. Interventions such as ensuring adequate patientto-staff ratios for appropriate monitoring and the ability to activate rapid response teams to address ongoing concerns with the patient s condition before it deteriorates into a code situation can help. However, there will always be unstable patients who are vulnerable to codes, and patients who arrive at the hospital in full cardiac and/or respiratory arrest. Thus, organisations need to address medication error risks that may be present during these unique and often chaotic situations. One study that examined medication errors associated with codes showed that action to reduce the risk of similar errors was documented for only 29% of the errors. When action was taken, the most common strategies were to inform staff who made the error and provide additional education both low leverage strategies. Less than 5% of the errors resulted in system- or policy-level action. Table 2 provides a list of strategies to decrease the risk of medication errors during codes. The list can be used to assess current practices and establish an action plan with high-leverage strategies that are more likely to reduce the risk of errors during codes. 55

5 56