Cochrane in CORR®: Reducing Medication Errors for Adults in Hospital Settings

Importance of the Topic

Preventable medication errors may be caused by routine failure (ignorance or inadvertence), insufficient practical skills, or shortcomings in complex systems and medication-delivery procedures [7, 16]. The fallout from these errors may vary: Some errors may lead to no harm, others to temporary or permanent harm, and some are fatal [1, 11]. In addition, medication errors contribute to increased length of hospital stay, treatment burden, and costs [3, 14]. In the United States, medication errors are the third most common cause of death [8, 12], and in Europe, in-hospital medication errors occur in 0.3% to 9.1% of prescriptions [6].

Medication errors are more likely to happen in surgical departments than in medical departments [2, 4], with the omission of preadmission medication being the most common error in scheduled surgeries [8, 13]. Many orthopaedic patients are older and present for care taking a large number prescription medications; polypharmacy in older patients is known to be an especially severe problem, accounting for a large number of serious complications [13, 15]. This highlights the importance of preventing medication errors during the perioperative period [8, 13, 15].

Various interventions have been investigated to reduce or prevent medication errors in hospital settings [10]. These include interventions that target specific sources of error, such as information technology–based interventions (like electronic barcoding and medication reconciliation) and those that seek to reduce the effect of distractions or interruptions on prescribers. Educational interventions and those involving prescriber-pharmacist partnerships have also been tried.

Given the large number and variety of such interventions, a synthesis like the recent Cochrane review by Ciapponi et al. [5] is especially welcome. In it, the authors investigated the effects of interventions in reducing medication errors in hospitalized adults compared with usual care or alternative interventions.

Upon Closer Inspection

The authors included 51 randomized trials and 14 time-series studies consisting of 110,875 adult patients in hospital settings, mostly from high-income countries. They found moderate-quality evidence that improved computerized physician order entry (CPOE)/clinical decision support systems (CDSS) were associated with a lower likelihood of medication errors and adverse drug events. Medication reconciliation (MR) was associated with lower likelihood of adverse drug events (moderate quality). The benefits in the reduction of medication errors and adverse drug events from the remaining interventions (barcoding, feedback on prescribing errors, and using dispensing system) were based on low-quality evidence. However, none of these interventions reduced the length of hospital stay, in-hospital mortality, or improved patients’ quality of life (low-to-moderate quality of evidence).

Despite the comprehensive review, several limitations on its analysis deserve mention. Almost all comparisons are informed by a limited number of studies, and thus the review authors decided to combine data from different hospital settings, studies with different durations of follow-up, and time-series analyses with randomized trials. There was also important variability in the methods of each intervention and the unit of analysis (patient and prescriptions). These have led to high statistical heterogeneity and imprecise effect estimates in most comparisons and may have resulted in the overestimation of benefits from the interventions. In addition, given the limited available evidence, it is hard to argue whether the lack of statistical significance for some outcomes (such as length of stay, quality of life, or mortality) was due to the inefficacy of interventions or the low statistical power. Even with the shortcomings mentioned, the current review is the best available evidence, and surgeons and healthcare providers should consider the highest quality evidence available when making clinical decisions.

Take-home Messages

We believe that the most robust and important findings of this Cochrane review were: (1) that MR performed by a pharmacist (the process of comparing medication orders to the medications taken by the patient) and (2) improved CPOE and CDSS may result in a reduction of medication errors and adverse drug events. Although the Cochrane review found that in aggregate these may result in little to no difference in patients’ length of stay, quality of life, or all-cause mortality, this is not necessarily surprising. Death and severe complications are rare endpoints; finding statistical differences in them would be surprising in the context of trials of new interventions, and not all important medication errors would be expected to prolong length of stay. It seems safe to expect that even modest reductions in the frequency of serious medication errors would decrease harm, and we believe this is a sufficient goal for interventions of these kinds.

Given these findings, orthopaedic surgeons and healthcare decision makers should consider using pharmacist-led medication reconciliation (in which pharmacists work together with surgeons, patients, families, and care providers to consistently communicate accurate and comprehensive medication information across transitions of care), improved CPOE or alternative and more comprehensive CDSS, or a combination of these methods in operating rooms and surgical wards to improve short-term outcomes such as medication errors and medication-related adverse events.