Prevention of adverse drug events in hospitals

INTRODUCTION — Adverse drug events (ADEs) comprise the largest single category of adverse events experienced by hospitalized patients, accounting for about 19 percent of all injuries [1]. The occurrence of ADEs is associated with increased morbidity and mortality [2,3], prolonged hospitalizations [4], and higher costs of care [2,5]. In the United States, due to the importance of ADEs, the Joint Commission has established national patient safety goals requiring each health care organization to implement comprehensive medication reconciliation at every transition point (eg, admission, transfer, discharge) along the continuum of care [6].

This topic will focus on interventions to prevent ADEs caused by medication errors in the hospital setting. Specific issues related to hospital discharge, drug prescribing in older adults, adverse drug reactions (ADRs), and deprescribing are discussed in detail elsewhere. (See "Hospital discharge and readmission" and "Drug prescribing for older adults" and "Drug hypersensitivity: Classification and clinical features", section on 'Categories of adverse drug reactions' and "Deprescribing".)

DEFINITIONS

●Adverse drug events (ADEs) are defined as any injuries resulting from medication use, including physical harm, mental harm, or loss of function [7]. ADEs, compared with medication errors, are a more direct measure of patient harm.

●Medication errors refer to any mistakes occurring in the medication use process, regardless of whether an injury occurred or whether the potential for injury was present [7,8]. Approximately 1 in 100 medication errors result in an ADE, while 7 in 100 have the potential to do so [9]. Although relatively few medication errors result in ADEs, they provide important information for identifying opportunities to improve patient care.

●Preventable ADEs are ADEs resulting from a medication error that can be avoided. At least a quarter of all medication-related injuries are preventable [10]. Preventable ADEs include errors made by the clinician and systematic errors.

●Adverse drug reactions (ADRs), or non-preventable ADEs, are ADEs that occur due to pharmacologic properties of the drug. (See "Drug hypersensitivity: Classification and clinical features", section on 'Categories of adverse drug reactions'.)

●Potential ADEs are medication errors that pose a significant risk but do not cause harm to a patient [7,11]. Potential ADEs are also called near-miss errors or close calls. Potential ADEs include errors that are detected and intercepted by a patient or clinical staff before the patient is affected.

The relationship between ADEs, medication errors, preventable ADEs, non-preventable ADEs (ADRs), and potential ADEs is depicted in a figure (figure 1).

INCIDENCE — Many studies have examined the incidence of adverse drug events (ADEs), but estimates vary widely depending on definitions of events, units of analysis, detection methods (eg, voluntary reporting, chart review, computerized surveillance), clinical settings, and subjects studied. Most early data examining ADEs came from observational studies performed in the United States and Canada, usually in urban tertiary care hospitals. Subsequent studies in international settings have largely corroborated these results [12].

ADEs can be broadly classified as occurring prior to hospitalization (in the outpatient setting or in the emergency department), during hospitalization, and immediately following discharge (to home or to another facility).

●Prior to hospitalization – ADEs in the outpatient setting are an important cause of emergency department visits and hospital admissions. One study in a Canadian hospital found that 2.4 percent of emergency department visits were associated with an ADE, of which 29 percent were preventable and 42 percent required hospitalization [13]. One study in a United States hospital found that ADEs were responsible for 1.4 percent of hospital admissions, of which 28 percent were preventable [14].

●During hospitalization – Several studies have documented the incidence of ADEs during hospitalization. The ADE Prevention Study was one of the first rigorous prospective cohort studies of ADEs in the hospital setting, with 6.5 ADEs per 100 adult admissions [7]. More than one-fourth of ADEs were preventable. Other large observational studies reported rates of ADEs from 2.4 to 52 ADEs per 100 hospital admissions [2,15-18]. Inpatient transitions are times of increased risk. Nearly half of patients transferred from an intensive care unit (ICU) to a non-ICU experienced a medication error in a study of 58 ICUs in the United States and Netherlands [19].

●Hospital discharge – Many ADEs occur following discharge from the hospital [20]. The incidence of ADEs for patients discharged home was 11 percent in a study of one United States academic hospital, of which 27 percent were preventable [21]. Patients were less likely to experience an ADE if they remembered someone explaining the side effects of prescribed medications. (See "Hospital discharge and readmission", section on 'Medication reconciliation'.)

High-risk settings — ADEs can occur in any area of the hospital, including the emergency department [22-25], hospital wards, and the operating room [26-31]. Two of the most common settings where ADEs occur are in the ICU [7,32-34] and during off hours (nights and weekends).

ADEs are more common in ICUs than in general medical and surgical units for several reasons [35,36]. First, patients in ICUs receive more medications than those in general medical and surgical units. Second, most medications in the ICU are given intravenously, leading to a higher risk of errors due to miscalculation of doses and infusion rates. Third, most patients in the ICU have multiple severe comorbidities and often are sedated, making them unable to check their own drugs or report on medication-related symptoms. Lastly, patients in the ICU are at increased risk of unintentional discontinuation of medications for chronic diseases (eg, antiplatelets, anticoagulants, statins), leading to ADEs over the long term [37].

ADEs may be more likely to occur overnight and during weekends, partly due to less staffing and supervision. In one pediatric hospital, medication error rates were higher during evening and night shifts compared with day shifts, and higher during weekends compared with weekdays [38].

High-risk populations — Younger children, older adults, and individuals with multiple comorbidities are at particular risk for ADEs.

Children are at high risk for medication errors and ADEs mainly because of the need to tailor doses to age, weight, or body mass index [39]. Younger children are also not usually able to identify potential errors themselves compared with older patients. In one study, 2.3 ADEs, 10 potential ADEs, and 55 medication errors occurred per 100 admissions at two urban teaching hospitals in the United States [40]. Another study reported a higher ADE rate (11.1 ADEs per 100 patients) in 12 United States hospitals [41].

Older adult patients are vulnerable to medication errors and ADEs due to their multiple comorbidities, diminished physiologic reserve, and more frequent use of multiple drugs. A meta-analysis of observational studies found that older adult patients were four times more likely to be hospitalized for an adverse drug reaction (ADR) than younger adults [42]. Another study found that 15 percent of hospitalized patients 70 years of age and older experienced at least one ADE, of which more than half were judged to be preventable [43]. The prevention of ADEs in the older adult population is discussed in detail elsewhere. (See "Drug prescribing for older adults", section on 'Preventing adverse drug events'.)

High-risk drugs — Several drug classes have been consistently associated with ADEs in studies of hospitalized patients (table 1). Anticoagulants, anti-hyperglycemic agents, sedatives, narcotics, antibiotics, antipsychotics, and chemotherapeutic agents are among the leading drug classes associated with ADEs in adults [7,44-46]. These drugs have several common features including narrow therapeutic windows, the potential for idiosyncratic physiologic responses, and variable dosing regimens. Several of the medications require close monitoring and dose adjustment, with the potential for calculation and timing errors. The most common medication classes causing ADEs in children include narcotics, antibiotics, and electrolytes/fluids [40,41,47,48].

DETECTION METHODS — The commonly used methods for detecting adverse drug events (ADEs) include voluntary reporting, chart review, computerized surveillance, and direct observation [49]. Each method detects different types of ADEs, and no single detection method is considered a gold standard for tracking ADEs. Multiple methods may be required to approximate the true ADE rate.

●Voluntary reporting by clinicians detects a small fraction of ADEs. Whether a given event is reported depends on individual attitudes, safety culture of the organization or hospital unit, ease of completing the report, fear of punishment, and concern about liability exposure [50]. Voluntary reporting identifies the lowest number of errors compared with chart review and automated surveillance [51,52].

●Medical record or chart review is a more systematic method for identifying ADEs, detecting many more ADEs compared with voluntary reporting (65 versus 4 percent) and computerized surveillance (65 versus 45 percent) [51]. However, chart review is costly and time-consuming, and certain ADEs are not reliably recorded in the medical record due to variable standards for documentation, liability concerns, lack of clinician awareness of the ADE, and incomplete record retrieval [51,53].

The use of an ADE trigger tool such as the Trigger Tool for Measuring ADEs developed by the Institute for Healthcare Improvement can facilitate manual and automated chart reviews and increase detection of ADEs [54-57]. Trigger tools include a list of clinical “clues” that an ADE may have occurred. Examples include the use of antidote medications such as naloxone for opioid-related ADEs or abnormal laboratory tests (ie, renal function or transaminase elevation) that may indicate medication-related toxicity. Trigger tools provide instructions for collecting the data to measure the number of ADEs per 1000 doses and the percentage of admissions with an ADE [58].

●Computerized surveillance detects many events not captured by voluntary reporting [51,52]. Automated surveillance can be used to monitor a large patient population continuously, with fewer labor hours than chart review. However, few hospitals have access to these types of automated surveillance systems.

●Direct observation by trained staff is regarded as the most effective method to detect medication administration errors [36]. However, it is also among the most expensive approaches. A study comparing voluntary reporting, chart review, and direct observation found that direct observation identified the greatest number of errors [15].

●Reports by patients and family members show promise as a detection method that complements other approaches. This method is well demonstrated in research studies, but its performance in operational settings requires further study [59-61].

INTERVENTIONS — Several interventions have been used to prevent adverse drug events (ADEs) and can generally be categorized as provider- or system-based interventions. Provider- and system-based interventions should be used together to optimally prevent ADEs.

A series of steps is required to manage medications: drug prescribing, transcribing, dispensing, administering, and monitoring. Although medication errors can occur at any step, most interventions are directed at the steps where ADEs tend to occur, which include drug prescribing, administration, and monitoring [40,41,47].

Provider-based approach — The individual clinician should ideally review the medication list at each patient encounter. Issues to consider (eg, dose, directions, drug interactions, and side effects) when reviewing medications are summarized in a table (table 2) and an algorithm (algorithm 1). Provider-based interventions to prevent ADEs are similar for the inpatient and outpatient settings. Some have noted value in including the medication’s indication (eg, hypertension, hyperlipidemia, pain) on the prescription or medication order [62].

Avoid and be vigilant of high-risk drugs — Several drug classes have been consistently associated with ADEs in studies of hospitalized patients (table 1). Clinicians prescribing high-risk drugs should be vigilant for adverse effects and should consider discontinuing these drugs or replacing with drugs that are less likely to cause adverse effects (eg, changing narcotics to acetaminophen or nonsteroidal antiinflammatory drugs [NSAIDs] in children, older adults, and patients with a history of substance abuse who may be more susceptible to adverse outcomes) (see 'High-risk drugs' above). Organizations such as the Centers for Disease Control and Prevention (CDC) offer safe prescribing checklists for opioids for chronic pain [63]. Intravenous drugs are particularly susceptible to “wrong rate” and “wrong time” errors [64].

Discontinue unnecessary drugs — The risk of ADEs increases with number of medications taken [32,65,66]. In addition to potentially harmful drugs, patients are often taking drugs that were previously given for a prior indication and are no longer needed (eg, proton pump inhibitors for gastrointestinal stress prophylaxis). It is likely that reducing the number of medications a patient is taking will lead to fewer ADEs [67]. One approach to assessing whether a drug is truly necessary for a given patient is presented in an algorithm (algorithm 1). Deprescribing is discussed in detail elsewhere. (See "Deprescribing".)

Consider drugs as a cause of any new symptom — Many preventable ADEs are simply due to clinicians failing to discontinue or adjust medications appropriately in the presence of significant side effects. In one prospective cohort of 661 patients, there were 51 cases of medication-related symptoms, in which 32 (63 percent) were attributed to clinicians failing to respond to patient symptoms while 19 (37 percent) were attributed to patients failing to report significant symptoms to the clinician [53]. Drug-induced side effects should be considered when a patient presents with a new complaint. Clinicians should discontinue or reduce the dose of medications thought to cause significant adverse effects. (See "An approach to the patient with drug allergy", section on 'Identification of the suspect drug'.)

Avoid treating side effects with another drug — Clinicians are often tempted to start medications to treat new symptoms (eg, hydrocortisone or hydroxyzine for opioid-induced pruritus), even when symptoms are clearly related to the initiation of new medication. The following table lists common medications used to treat adverse effects from other medications (table 3). Clinicians should consider discontinuing the inciting drug and replacing with alternative therapy (eg, physical therapy rather than pain medications to treat musculoskeletal pain, lifestyle modification rather than high doses of lipid-lowering drugs for hypercholesterolemia). (See "An approach to the patient with drug allergy", section on 'Options for future treatment'.)

Avoid drug-drug interactions — The risk of an adverse event due to drug-drug interactions is substantially higher when more medications are being prescribed [68,69]. Most drug-drug interactions involve commonly used medications. As an example, the risk of bleeding with warfarin therapy is increased with coadministration of NSAIDs, selective serotonin reuptake inhibitors (SSRIs), omeprazole, lipid-lowering agents, and/or amiodarone [68]. In one case-control study of older patients, hospitalization for hypoglycemia was six times more likely with concomitant administration of trimethoprim-sulfamethoxazole, digoxin toxicity was 12 times more likely with clarithromycin, and hyperkalemia was 20 times more likely with a potassium-sparing diuretic [69].

Specific interactions may be determined using the drug interactions program included in UpToDate. This tool can be accessed from the UpToDate online search page or through the individual drug information topics in the section on drug interactions.

Adjust dosing based on age and creatinine clearance — A common cause of ADEs is failure to properly adjust doses for age and renal insufficiency. Adjusting appropriately for age is particularly important in pediatric and geriatric populations. Renal impairment can occur at any age but becomes more common with advancing age. For patients with stable renal function, creatinine clearance can be estimated according to published formulas which factor age into the calculation (calculator 1). Due to decreased muscle mass in older adults, however, serum creatinine levels may not adequately reflect renal function. As a general rule, the initial dose for starting medications in older adults should be significantly reduced, and titrated up as tolerated by monitoring side effects or drug levels.

The drug database (Lexicomp) available through UpToDate includes appropriate dose adjustments for age and renal function. Dose adjustments can be accessed by searching on any individual drug.

Address non-adherence — As patients may have concerns about taking certain medications or may not be taking medications as directed, the review of a patient’s medication list can allay patient fears and improve patient adherence to medications. Patients should be educated on the indications for each medication, possible side effects, and alternative options that may decrease the risk of ADEs. Factors affecting medication adherence (eg, communication barriers, inadequate follow-up) and specific methods to improve adherence (eg, motivational interviewing) are discussed in detail elsewhere. (See "Adherence to lipid-altering medications and recommended lifestyle changes", section on 'Strategies for increasing adherence'.)

System-based approach — Most medication errors and ADEs are caused by poorly designed systems that either induce errors or make them difficult to detect. [70]. While human error is usually the proximal event leading to an ADE, another fundamental fault is in the system that permits the error to reach the patient [71]. As an example, when a clinician prescribes tobramycin for a patient with renal impairment but fails to adjust the dose for estimated creatinine clearance, the clinician’s error could result in an ADE. However, a properly designed system could prevent this event by providing the clinician with an automated alert.

Computerized physician order entry — Computerized physician order entry (CPOE) refers to a variety of computer-based systems that facilitate the medication ordering process [72]. Some CPOE systems provide more than prescription-writing capabilities with fixed fields and drop-down menus for drug, dose, route, and frequency. These systems are further enhanced by the integration of clinical decision support systems (CDSS) of varying sophistication. CDSS includes interactive software designed to assist health professionals in making clinical choices. CPOE with CDSS targets the prescribing stage of the medication use process and improves medication safety in the following ways [73,74]:

●Providing a means for standardization of practice

●Improving the completeness and legibility of orders

●Alerting clinicians to drug allergies, drug-drug interactions, and cumulative dose-limits

●Updating clinicians with the most current medication information

●Providing dosage adjustment calculations based on patient characteristics

●Timely communication of critical changes in a patient’s condition, in turn facilitating appropriate adjustments

Several studies have found that CPOE with CDSS significantly prevents medication errors and ADEs [75-80]. In one study, the implementation of CPOE with CDSS reduced the serious medication error rate by 55 percent and potential ADE rate by 84 percent [78]. Another study reported that CPOE with CDSS improved the rate of appropriate prescribing of nephrotoxic drugs compared with CPOE only (51 versus 30 percent) [79].

Some studies have reported unintended consequences related to the introduction of CPOE [81-85], especially among CPOE systems with limited decision support. Examples include inflexible ordering formats generating incorrect orders [81] and juxtaposition errors in which clinicians click on the adjacent patient name or medication from a list and inadvertently enter the wrong order [83,86]. These studies highlight the importance of proper design, incorporating decision support, staff training, and the need to monitor new technologies carefully to identify unanticipated problems.

Electronic medication administration record — Electronic innovations have also sought to address medication errors related to transcribing. Medication orders that are manually transcribed to a paper medication administration record and used by nurses to administer medications are prone to causing errors and ADEs. When CPOE has an electronic interface with an electronic medication administration record (eMAR), the need for medication transcription is eliminated, removing errors caused by interpretation and translation [73,87,88]. Moreover, eMAR helps organize medication administration schedules and prompts nurses to ensure timely medication administration [73].

Few studies have evaluated the effects of eMAR on drug transcription and administration errors. One study found that CPOE combined with eMAR eliminated all transcribing errors, compared with an error rate of 11 percent in CPOE with manual medication documentation [89]. Another study similarly found that transcription errors were completely eliminated on hospital units that used eMAR, compared with a rate of 6 percent on units that did not use the eMAR [90].

Bar coding — Bar codes can be affixed to medications and patient wristbands in order to ensure matching between patients and their medications at the time of drug administration. Additional interfaces with CPOE and eMAR allow for a closed-loop system that confirms a match between medication orders, medication preparation (including dispensing), and patients receiving medications. Most importantly, bar coding provides the final opportunity to intercept medication errors before drug administration. Bar coding can reduce administration errors and potential ADEs [90-92].

Smart pumps — Smart pumps are used to reduce errors associated with intravenous medication administration through their built-in safety features, such as safety alerts, clinical calculators, dose limits, and drug libraries. However, smart pumps have not been consistently found to prevent ADEs. While one study in a pediatric hospital found that the combination of smart pumps, standard drug concentrations, and improved labeling led to a 73 percent reduction in reported medication-infusion errors [93], most other studies reported no significant impact of smart pumps on serious medication errors and ADEs [94-96].

Several barriers undermine the effective implementation of smart pumps, such as inconsistencies in the smart pump drug libraries and bypassing of safety alerts during administration [94,97-100]. A study of 100 hospitals using smart pumps from the same manufacturer suggested substantial variability in drug names, dosing units, dose limits, and concentrations within the same library, which raise the risk of errors and ADEs [97]. Another study found high medication discrepancy rates for infusion pumps (24 percent for medications and 42 percent for fluids) [99]. In addition, one randomized time-series trial found that nurses bypassed drug libraries and alerts 25 percent of the time and gave medications without having a documented order in 8 percent of cases [94]. Improving decision support software, implementing use standards, and monitoring compliance are essential to overcome these workarounds.

Pharmacist interventions — Pharmacists can play an important role in preventing ADEs through various interventions, such as patient counseling at admission and discharge, medication reconciliation, daily medication review with face-to-face clinician contact, pharmacist participation on rounds, and drug class-specific pharmacist services (eg, anticoagulation services) [101]. Pharmacist review of medications and timely communication with clinicians decreases in-hospital medication error and ADE rates [102,103]. Pharmacist interventions at the end of hospitalization lead to fewer medication-related emergency department visits and hospital readmissions [104,105] and lower preventable ADE rates [105,106]. In addition, many studies indicate that the inclusion of pharmacists on rounds as a member of the care team is associated with reduced medication error and ADE rates in intensive care units (ICUs) [107,108] and in general medical or surgical units [16,109].

Medication reconciliation — Medication reconciliation is a process that identifies medication discrepancies, informs prescribing decisions, and prevents medication errors that could harm patients [17]. Unintentional medication discrepancies at the time of transitions occur commonly at the time of admission (54 percent) [18], during intra-hospital transfer (62 percent) [110], and at discharge (71 percent) [111]. About one-third of these discrepancies have the potential to cause patient discomfort or clinical deterioration [18,110,111]. A systematic review of 26 controlled studies of variable quality, including 10 randomized trials, found that medication reconciliation consistently reduced discrepancies, with a decrease in actual and potential ADEs [112]. Interventions were most effective that involved pharmacy staff intensively and that targeted high-risk patients.

The process of medication reconciliation has three steps [6,113]:

●Verification – Reviewing the patient’s medication use history and developing an accurate list of medications.

●Clarification – Ensuring that the medications and doses are appropriate and using the current list when writing medication orders.

●Reconciliation – Identifying any discrepancies between medication ordered for patients and those on the list, making appropriate changes to the orders, documenting any changes, and communicating the updated list to the next provider within or outside the hospital. Patients should also be provided written information on the medications when discharged from the hospital.

Of note, however, in 2014 a major provider of medication history services to hospitals determined that one data source was missing special characters such as a decimal point, forward slash, or percentage in some records (“.”, “/”, “%”); this would cause a drug previously dispensed as “ramipril 2.5 mg capsules” to be reported as “ramipril 25 mg capsules” [114]. Thus, when systems based solutions are implemented, organizations and providers must remain vigilant about the quality and integrity of their data sources. Medication reconciliation at time of hospital discharge is discussed in detail elsewhere. (See "Hospital discharge and readmission", section on 'Medication reconciliation'.)

Other — Educational programs may modestly decrease rates of medication errors and ADEs [13,115,116]. There was a 14 percent reduction in the reported rate of ADEs in hospitalized older adult patients after the implementation of an education intervention in which care teams were provided specific recommendations on medication use in older adults and how to identify and prevent ADEs [116].

Another study found that web-based education, point-of-care standardized dosing references, a zero-tolerance policy for incomplete or incorrect drug orders, and prescriber performance feedback reduced prescribing errors by 49 percent for pediatric patients [115].

Developing an accurate allergy list and storing allergy information in a single repository are essential to prevent medication errors and ADEs related to drug allergies [117,118]. In one study, the number of allergy errors decreased more than 80 percent following the implementation of CPOE with allergy-related decision support [119].

Developing and implementing medication safety best practices is crucial to preventing similar errors from happening again. In 2014, the Institute for Safe Medication Practices (ISMP) released its first consensus-based medication safety best practices for hospitals and has made biennial updates since. The report focuses on specific medication errors that continue to cause fatal and life-threatening events in patients despite repeated warnings from ISMP publications and encourages hospitals to implement best practices such as safeguard against errors with oxytocin use, maximizing the use of barcode verification prior to medication, and vaccine administration [120].

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Patient safety in the operating room".)

SUMMARY AND RECOMMENDATIONS

●Adverse drug events (ADEs) are defined as any injuries resulting from medication use, including physical harm, mental harm, or loss of function. ADEs, compared with medication errors, are a more direct measure of patient harm. Medication errors refer to any mistakes occurring in the medication use process, regardless of whether an injury occurred or whether the potential for injury was present. (See 'Definitions' above.)

●ADEs can occur at any time and place in the hospital: prior to being admitted (as an outpatient or in the emergency department), during the hospital stay (in medical or surgical units, intensive care units [ICUs], operating room), and immediately following discharge (to home or to another facility). (See 'Incidence' above.)

●Children are at high risk for medication errors and ADEs, mainly because of the need to tailor doses to their age, weight, or body mass index. Younger children are not usually able to identify potential errors themselves. Older adult patients are also vulnerable to medication errors and ADEs due to multiple comorbidities, diminished physiologic reserve, and use of multiple drugs. (See 'High-risk populations' above and 'High-risk drugs' above and "Drug prescribing for older adults".)

●A multifaceted approach should be used to prevent medication errors and ADEs in hospitals. This includes both provider- and system-based interventions. (See 'Interventions' above.)

●The individual clinician should review the medication list at each patient encounter and adjust medications accordingly to decrease risk of ADEs. This approach includes avoiding high-risk drugs, unnecessary drugs, drugs causing side effects, and drug-drug interactions (algorithm 1). The clinician should also address non-adherence and appropriately dose drugs based on age and creatinine clearance. (See 'Provider-based approach' above.)

●Most medication errors and ADEs are caused by poorly designed systems that either induce errors or make them difficult to detect. System-based interventions to prevent medication errors include computerized physician order entry (CPOE) systems with clinical decision support, electronic medication administration record (eMAR), bar coding, smart pumps, pharmacist counseling with patients and clinicians, medication reconciliation, and educational initiatives. (See 'System-based approach' above and "Hospital discharge and readmission".)