Recent Literature on Medication Errors and Adverse Drug Events in Older Adults

Abstract

Medication errors and adverse drug events are common in older adults, but locating literature addressing these issues is often challenging. The objective of this article was to summarize recent studies addressing medication errors and adverse drug events in a single location to improve accessibility for individuals working with older adults. The authors conducted a comprehensive literature search for studies published in 2014 and identified 51 potential articles. After critical review, 17 studies were selected for inclusion based on innovation, rigorous observational or experimental study designs, and use of reliable, valid measures. Four articles characterizing potentially inappropriate prescribing and interventions to optimize medication regimens were annotated and critiqued in detail. We hope that health policy makers and clinicians find this information helpful in improving the quality of care for older adults.

INTRODUCTION

Many older adults have multiple medical conditions commonly treated with multiple medications.¹ Indeed, this past year we learned that over 50% of older adults in the United States take four or more medications, which is higher than rates reported in Canadian or the United Kingdom.² Unfortunately, the use of multiple medications can lead to medication errors and adverse drug events.¹ To address these unintended negative consequences, it is important to answer certain clinical questions: How is drug therapy being used in older adults? If potential medication errors are detected, what are the outcomes of drug therapy? And finally, are there interventions that are effective in modifying the use and outcomes of drug therapy?³

Unfortunately, literature addressing these questions is challenging to locate due dispersed publication worldwide and a lack of specific, standardized search terms with which to query journals. Therefore, the editors of the Journal of the American Geriatrics Society have graciously permitted us to publish a compilation of the latest literature regarding medication errors and adverse drug events in older adults for the third consecutive year. The objective of this article is to summarize recent studies in a single location to improve accessibility to individuals working with older adults. We hope you find the following articles informative and pertinent to your practice.

METHODS

A search of Medline^® and Google Scholar™ restricted to 2014 was conducted using a combination of the terms: adverse drug events, adverse drug reactions, adverse drug withdrawal events, therapeutic failure, aged, elderly, drug-related problems, medication-related problems, medication errors, polypharmacy, unnecessary drug use, suboptimal or inappropriate prescribing, under prescribing/utilization, medication non-adherence, medication administration or dispensing errors, medication monitoring omissions, and medication misadventures A manual search for relevant articles from the New England Journal of Medicine; Age and Aging; Annals of Internal Medicine; Annals of Pharmacotherapy; Clinical Pharmacology and Therapeutics; Journal of the American Medical Association; Journal of the American Medical Association Internal Medicine; Journal of Gerontology: Medical Science; Lancet; and Pharmacotherapy: The Journal of Human Pharmacology and Drug Therapy was also conducted. Additional articles identified by the authors were also considered.

The abstracts of identified articles were independently reviewed by two authors (JGN, JTH) for relevance. A preliminary list of potential articles for inclusion was then reviewed by all authors and consensus was reached regarding which to highlight. Preference was given to studies that were innovative and used rigorous observational or experimental study designs and reliable, valid measures. Review articles without an accompanying meta-analysis or articles focused exclusively on drug efficacy/effectiveness were excluded. Also excluded were articles published in the Journal of the American Geriatrics Society, Consultant Pharmacist, and Drugs and Aging as these journals are obvious places for readers to easily locate relevant material.

RESULTS

A preliminary list of 51 articles was developed. After each author reviewed the abstracts of these articles and consensus was reached regarding which articles to highlight, a total of 17 remained. Below we annotate and critique four studies.^4–7 The remaining article abstracts can be found in the online Appendix by categories, alphabetized by the first author’s last name.

Article 1: Improving Inappropriate Prescribing

The Eliminating Medications Through Patient Ownership of End Results (EMPOWER) trial was a pragmatic cluster randomized clinical trial in Montreal, Canada, designed to evaluate the effectiveness of personalized, direct patient education on benzodiazepine (BZD) discontinuation or dose reduction (defined as ≥ 25% decrease in dose sustained for ≥ 3 consecutive months) in adults older than 65. Patients taking at least 5 medications, including one chronic benzodiazepine, were recruited. The intervention group (n=148) received information about potential harms of BZD as well as a visual 21-week step-wise tapering protocol, while the control group (n=155) received usual care. At 6 months, 27% of the intervention group had achieved complete BZD cessation and an additional 11% had achieved dose reduction, compared with 5% and 6%, respectively, of the control group. Overall, nearly 4 patients would need to receive the intervention to achieve BZD discontinuation or dose reduction over 6 months. The authors concluded “direct-to-consumer education effectively elicits shared decision making around the overuse of medications that increase the risk of harm in older adults.”

“Deprescribing” is an increasingly important concept that may enable providers to avoid clinical inertia leading to persistent use of potentially inappropriate medications (PIM).^8,9 Although two other studies addressing BZD discontinuation using different interventions have been published, neither yielded statistically significant results.^10,11 This study is important in that it is the first examining the role of the patient in deprescribing BZD, and demonstrates that shared decision making is an effective strategy for reducing PIM use. Given appropriate information, older adults can play an integral role in identifying situations of overtreatment and can initiate cessation of potentially harmful medications. Even if there was no ultimate attempt to taper the BZD, 62% of participants in this study communicated concerns with their healthcare provider.

As with any study, however, there are some potential concerns. It is important to note that among individuals in the intervention group, adverse drug withdrawal events (e.g., rebound insomnia and anxiety) were high at 42%. However, it is unclear whether a valid instrument was used to measure insomnia adverse drug withdrawal events, and the rate of withdrawal events in the control group was not provided. The high withdrawal rate may be secondary to the rapidity of the tapering protocol or lack of supervision by a healthcare provider (38% of individuals attempting a taper did not discuss their intention with a physician or pharmacist). Notably, 5% of those individuals who discontinued a BZD received a highly-anticholinergic PIM drug (i.e., paroxetine, amitriptyline) as a substitute. As a next step, it would have been helpful to determine if the reduction in BZD use associated with this intervention had a favorable impact on other clinical outcomes such as cognitive impairment and falls.

Article 2: Documenting Potentially Unnecessary Prescribing

Tjia et al. conducted a national cross-sectional study of data from 460 nursing homes to determine the rate of potentially unnecessary medications in residents with advanced dementia and to identify associated resident- and facility-level factors.⁵ Unnecessary medications were defined as those considered “never appropriate” for use in advanced dementia according to explicit criteria developed using a Delphi survey.¹² More than half (53.9%) of nursing home residents with advanced dementia received at least 1 potentially unnecessary medication. Acetylcholinesterase inhibitors, memantine, and lipid-lowering agents were the most common unnecessary medications. Interestingly, living in a facility where at least 10% of residents had feeding tubes was the only factor associated with an increased risk of unnecessary drug use. The authors concluded “most nursing home residents with advanced dementia receive medications with questionable benefit that incur substantial associated costs.”

Unnecessary drug use in those with limited life expectancy has drawn considerable attention over the past decade.^12–16 The current study is among the first to examine this issue in such a large representative sample (n=5,406). The only prior study (published by Colloca et al.) included 1,449 Italian nursing home residents with dementia.¹⁷ Interestingly, that study employed the same explicit criteria but found a lower rate of potentially unnecessary drugs (26.8%). The most common potentially unnecessary medications were the same as in the current study, with the exception of a slightly higher rate for antiplatelets (9.9% vs 7.2%, respectively). In both studies residents with diabetes, heart failure, stroke, or a recent hospitalization had a higher risk for unnecessary medications. In the study by Tjia, a physician visit within the past 14 days was also associated with increased risk of questionable medications, whereas Colloca found having a geriatrician on site was protective.

There are potential limitations to both studies. The cross-sectional design does not allow one to know when unnecessary drugs were prescribed (e.g., prior to or during the nursing home stay). Moreover, the criteria used to measure unnecessary drug use did not include certain medications classes other experts consider relevant, including cardiovascular drugs (i.e., angiotensin converting enzyme inhibitors, anticoagulants, angiotensin receptor blockers, antihypertensives), anticholinergics (e.g., antivertigo, tricyclic antidepressants), anti-infectives, bisphosphonates, central nervous system stimulants, diabetes drugs, expectorants, genitourinary antispasmodics, minerals, proton pump inhibitors, and vitamins.^14–16 Using only explicit criteria as opposed to valid and reliable implicit measures such as the Medication Appropriateness Index does not take patient specific factors or patient/caregiver preferences into account. Finally, neither study examined the consequences of discontinuing these questionable medications. It is encouraging, however, that one recent reverse randomized controlled trial discontinuing lipid-lower agents in individuals with end stage cancer did not result in an increase in adverse cardio/cerebrovascular events.¹⁸

Article 3: Determining the Epidemiology of Medication Non-Adherence

Tamblyn and colleagues examined the occurrence of primary nonadherence (defined as having a prescription written but not filled at the pharmacy within 9 months) in 15,961 older Canadians receiving care from 131 primary care physicians. Almost a third (31.3%) of new prescriptions were not filled, with highest nonadherence for medication classes used to treat headache, ischemic heart disease, and depression. Primary non-adherence was less likely among individuals who were older in age, received prescriptions without copayments, and had more visits with the prescribing physician. The authors concluded “primary nonadherence is common and may be reduced by lower drug costs and payments, as well as increased follow-up care with prescribing physicians for patients with chronic conditions.”

This is one of the first studies examining primary nonadherence in elderly patients across medical conditions and medication classes. Primary medication nonadherence is concerning in that it may lead to greater health care utilization, morbidity, and mortality. This study did not measure whether primary nonadherence resulted in these adverse health outcomes. It is important to keep in mind, however, that some nonadherence may have been “intelligent” if the prescription was for a PIM where risk associated with its use may outweigh the benefit; in this situation, fewer adverse health outcomes may be expected. Higher rates of nonadherence with more expensive medications is consistent with another recent study highlighted in the Appendix, and may be due to the fact that physicians often do not know the costs of medications.^19,20

Article 4: Reducing a Specific Type of ADR - Drug Induced Falls

A randomized controlled pilot study by Michalek et al. examined the impact a drug regimen review and corresponding feedback to the prescribing physician had on suboptimal prescribing and falls. Of the 114 older hospitalized patients from Germany enrolled in this study, 56 patients were randomized to have their medication profiles reviewed and medications adjusted based on the Fit fOR The Aged (FORTA) criteria. Compared to the control group, the intervention group had significantly lower rates of over-prescribing (i.e., unnecessary medications) and under-prescribing (i.e., omissions of necessary drugs). Additionally, fewer patients in the intervention group versus control group experienced falls (3.4 % vs 21.4 %, respectively). The authors concluded that although medication quality improved and fall rate decreased in the intervention group, “[t]hese encouraging results must be interpreted carefully.”

This study joins 13 other randomized controlled trials designed to reduce potentially inappropriate prescribing and subsequent falls in older adults.^21–32 Including the current study, only 7 have achieved these objectives.^{23,24,26,29,30,32} Moreover, the current trial is the only successful single intervention study conducted in a hospital, where falls are common and can lead to considerable morbidity and additional health care costs.³³

However, like all studies this one has some limitations. Although not possible in a pilot study in one hospital, future multicenter studies should randomize physicians and use them as the unit of analyses to determine the impact of the intervention. Secondly, the FORTA criteria may be obsolete with the recent publication of two new European measures for inappropriate prescribing.^34,35 Finally, as this study was conducted in one hospital in Germany, generalizability to other health care settings or other health care systems is unknown.

CONCLUSIONS

Although there were a number of articles published this year that addressed medication errors and adverse drug events in older adults, the selected articles describe particularly interesting advancements. One study highlights a novel intervention that recruits patient participation to withdraw potentially inappropriate benzodiazepines. Another demonstrates that medications unlikely to improve quality of life are commonly prescribed to older adults with limited life expectancy. The third study quantifies primary medication nonadherence in older adults and examines factors associated with this particular drug-related problem. A final intervention study in hospitalized elders showed that reviewing drug regimens and providing feedback to physicians not only improved suboptimal prescribing, but also reduced the rate of falls. We hope that health policy makers and clinicians find this information helpful in improving the quality of care for older adults.