Journal of Gerontological Nursing

Geropharmacology 

Clinical Alerts to Decrease High-Risk Medication Use in Older Adults

Wilhelmina Lord-Adem, PharmD; Nicole J. Brandt, PharmD, MBA, CGP, BCPP, FASCP

Abstract

High-risk medications (HRMs) account for 14.6% to 54.6% of all medications used in older adults, and have been linked to >50% of adverse drug events (ADEs). HRM-related ADEs lead to increased morbidity and mortality, increased hospital length of stay, and financial costs for patients and health care systems. It has been well documented that incorporating information technology in patient care in the form of clinical alert systems can effectively decrease HRM use and improve patient safety. The current article seeks to identify and discuss clinical alert systems focusing on HRMs, their impact on prescribing for older adults, and challenges to the implementation of electronic decision systems. [Journal of Gerontological Nursing, 43(7), 7–12.]

Abstract

High-risk medications (HRMs) account for 14.6% to 54.6% of all medications used in older adults, and have been linked to >50% of adverse drug events (ADEs). HRM-related ADEs lead to increased morbidity and mortality, increased hospital length of stay, and financial costs for patients and health care systems. It has been well documented that incorporating information technology in patient care in the form of clinical alert systems can effectively decrease HRM use and improve patient safety. The current article seeks to identify and discuss clinical alert systems focusing on HRMs, their impact on prescribing for older adults, and challenges to the implementation of electronic decision systems. [Journal of Gerontological Nursing, 43(7), 7–12.]

High-risk medications (HRMs) can cause significant patient harm (Institute for Safe Medication Practices [ISMP], 2017) through the occurrence of adverse drug events (ADEs) (Arvisais et al., 2015), which can lead to increased morbidity and mortality (Makris et al., 2015). HRMs account for 14.6% to 54.6% of all medications used in older adults (i.e., individuals age 65 and older), and their use has been linked to >50% of all ADEs among all patient populations (Arvisais et al., 2015; ISMP, 2017; Lee et al., 2014). Studies have shown that among hospitalized older adults, ADEs are the most common cause of injury, accounting for approximately 20% of all injuries (Bates et al., 1998). Other studies indicate that older adults are more likely to visit emergency departments (EDs) for ADEs, and seven times more likely than younger patients to be admitted during such visits (Makris et al., 2015).

Due to age-related pathophysiologic, pharmacokinetic, and pharmacodynamic changes, older adults are more likely to experience ADEs than younger patients, and the effects are typically worse in older adults than in younger patients (Arvisais et al., 2015; Lee et al., 2014). Other factors contributing to the increased susceptibility of older adults to HRM-related ADEs are the presence of multiple comorbidities, polypharmacy, and functional impairment (Makris et al., 2015). Furthermore, researchers have found that 43% to 54% of all ADEs are preventable (Gray, Sager, Lestico, & Jalaluddin, 1998; Klopotowska et al., 2013).

In addition to patient harm, HRM-related ADEs have financial implications for patients and health care institutions. The Harvard Medical Practice Study identified ADEs as the primary source of injury among hospitalized patients, causing approximately 20% of all injuries in this patient population and leading to annual post-hospitalization costs of $2 billion, excluding malpractice costs (Bates et al., 1998).

Given the high cost of ADEs for patients and hospitals, it is imperative that health care systems implement measures to improve safety and the quality of care for all patients, especially older adults. It has been well documented that incorporating information technology in patient care in the form of clinical alert systems is an effective tool for optimizing drug therapy and improving patient safety. Although the initial investment may be high, the long-term benefits of clinical decision tools lead to substantial savings for patients and health care systems (Bates et al., 1998).

Clinical Decision Tools: Methods Focusing on High-Risk Medications

Clinical decision tools include clinical alert systems (CAS), clinical decision support systems (CDSS), and computerized prescription order entry (CPOE) (Table 1). CAS are software programs that provide real-time alerts, typically in the form of pop-ups, to clinicians at the point of care (Kesselheim et al., 2011). Aimed at preventing ADEs and decreasing prescribing errors, CAS alert clinicians to potential drug–drug interactions and inappropriate dosages. Evidence suggests that exposure of clinicians to pop-up alerts can result in up to a 12-fold decrease in prescribing errors (Klein, 2013). Some CAS programs, aimed at improving therapeutic care for older adults, integrate geriatric-specific criteria such as the American Geriatrics Society (AGS) Beers Criteria and Screening Tool to Alert to Right Treatment/Screening Tool of Older People's Prescriptions (START-STOPP) criteria to help optimize clinical care for this patient population (Kesselheim et al., 2011).

Examples of Clinical Decision Support Tools

Table 1:

Examples of Clinical Decision Support Tools

Unlike CAS, which provide general medication-related information to clinicians, CDSS analyze patients' electronic medical records to provide patient- and medication-specific recommendations to clinicians (Kesselheim et al., 2011). Recommendations provided by CDSS include, but are not limited to:

CDSS targeting HRM use in older adults incorporate globally accepted criteria for monitoring drug therapy for older adults, such as the AGS Beers Criteria, START-STOPP criteria, scales of anticholinergic drug exposure, and the Drug Burden Index (Arvisais et al., 2015). One example is the Geriatric Risk Assessment MedGuide™ (GRAM), a CDSS software that was developed by the American Society of Consultant Pharmacists (ASCP) Foundation in collaboration with Brown University's Center for Gerontology and Health Care Research. A novel CDSS, GRAM enabled pharmacists to establish links between nursing home residents' ADEs and their medications. During a randomized controlled trial (RCT), use of this software resulted in a 58% reduction in medication-induced delirium and a 12% decrease in hospitalization among newly admitted residents to a long-term care facility (ASCP Foundation, 2017). The study also found that 12% of participants were less likely to die than those in the control group.

Another CDSS that has been proven effective for improving patient safety is Information Systems-enabled Outreach for Preventing Adverse Drug Events (ISTOP-ADE). Aimed at preventing non-adherence and decreasing the incidence, duration, and severity of ADEs, ISTOP-ADE combines an interactive voice response system (IVRS) with pharmacist care. The IVRS component contacts patients at predetermined times after starting a new medication to identify ADEs and connect patients with pharmacists as needed. Patients can also reach pharmacists at any time through the IVRS with questions about their medications. During a pilot study with 628 patients, ISTOP-ADE effectively contacted 90% of participants and identified 45% of ADEs, leading to more efficient management of 42% of all ADEs (Forster et al., 2015). Overall, integration of ISTOP-ADE led to reductions in the incidence and severity of ADEs and the number of ADEs requiring ED visits or inpatient care (Forster et al., 2015).

In addition to real-time alerts generated by CAS and CDSS, studies have shown that incorporating clinical alerts into the prescribing process in the form of CPOE is an effective strategy for improving patient safety. CPOE allows all prescriptions to be ordered online, thereby eliminating the problem of illegible medication orders and reducing the time required for transcription (Bates et al., 1998). In addition, some CPOE programs incorporate menu options with appropriate dose ranges, routes of administration, and relevant laboratory tests and results, which further improve patient safety and prescribing practices. For instance, the impact of CPOE on medication errors led to a 23% decrease in dosing errors, 56% decrease in serious allergy errors, and a significant decrease in drug–drug interactions among study participants (Bates et al., 1998).

Although many clinical decision tools such as GRAM and ISTOP-ADE function effectively as stand-alone programs, evidence suggests that combining them with other clinical decision tools further enhances their effectiveness at optimizing therapeutic care and preventing medication errors. An example is HARMLESS—High Alert Medication Recognition and Management System to Lower Errors and Secure Safety. A novel CDSS-CAS, it focuses on HRMs and provides pop-up notifications and order recommendations to clinicians at the point of decision making. Within 6 months of integrating it with the CPOE at a large teaching hospital in Korea, HARMLESS led to the prevention of 4,137 medication errors (Lee et al., 2014).

Impact of High-Risk Medication Clinical Alerts on Adverse Drug Events

Numerous studies have demonstrated the impact of CAS, CDSS, and CPOE on patient care: fewer errors in prescribing, lower HRM use, improved patient safety, decreased health care costs, and prevention of ADE-related hospitalizations (ASCP Foundation, 2017; Bates et al., 1998; Forster et al., 2015; Lee et al., 2014). Researchers following hospitalized older adults in an acute care geriatric ward found that within 1 year of implementation, INTERCheck, a novel CDSS designed specifically to help improve prescribing for older adults, led to a significant decrease in the use of HRMs and incidence of ADEs (Ghibelli et al., 2013).

In addition to identifying clinically significant alerts, decreasing the use of HRMs in older adults, and preventing HRM-related ADEs, CAS/CDSS/CPOE have also been shown to result in more appropriate prescribing. In a 2001 pilot study of older adults who were prescribed HRMs during admission to a tertiary care hospital, implementation of a newly developed CDSS correctly identified and flagged the medication regimens for 22% of participants (Peterson et al., 2014). Pharmacists reviewed the flagged regimens and made recommendations to prescribing physicians. In 78% of cases, physicians implemented the pharmacists' recommendations (Peterson et al., 2014), leading to more appropriate prescribing. This conclusion, in addition to the findings of other studies, suggests that patients stand to benefit immensely from clinicians' use of clinical decision support tools.

Challenges to Implementing Clinical Alerts

Despite documented evidence of the effectiveness of clinical decision support tools in improving therapeutic care, not all health systems have adapted them. Numerous studies have identified success factors for the implementation of CDSS into patient care: presentation of the right information, in the right system, in the right place, and at the right time (Scheepers-Hoeks, Grouls, Neef, Ackerman, & Korsten, 2011) (Table 2). In addition, a 2005 study by Kawamoto and Garg found that clinical decision support tools are more likely to be successful if they provide automatic recommendations for action at the point of care and do not interfere with established workflow patterns at institutions where they are adapted (Klein, 2013). Health systems that fall short in any of these parameters are likely to fail in their implementation of clinical decision support programs. Other factors hindering widespread adaptation of clinical decision support tools include lack of interoperability between existing software and new programs and lack of time to train staff to use new health information programs (Scheepers-Hoeks et al., 2011). Alert fatigue, lack of capital, and lack of buy-in from clinicians have also been identified as chief challenges to implementation of clinical decision programs.

Success Factors for Implementation of Clinical Decision Support Systems

Table 2:

Success Factors for Implementation of Clinical Decision Support Systems

Alert fatigue, a situation where CDSS generate such a large number of clinically insignificant alerts that users disregard a high proportion of them, has been identified as the biggest challenge to the widespread adaptation of CAS and CDSS (Kesselheim et al., 2011). In institutions that implement them, alert fatigue is the primary reason for the high proportion of alert overrides by clinicians (Van der Sijs, Aarts, Vulto, & Berg, 2006). In one study, it was noted that a CDSS generated alerts for topical medications with limited systemic absorption, which led clinicians to override >50% of all alerts due to the lack of clinical significance (Kuperman et al., 2007). Another study found an even greater rate of alert overrides due to the presentation of clinically insignificant alerts: 49% to 96% of all alerts (Van der Sijs et al., 2006).

At the University of California San Francisco (UCSF) Medical Center, a clinical decision support tool was noted to generate pop-up alerts for pharmacists for approximately 50% of all medication orders (Wachter, 2017). A 2011 study at the same institution found that approximately 2,507,822 monthly intensive care unit alerts were generated at UCSF Medical Center, a majority of which were false. As a result, clinicians routinely disregarded or overrode alerts, leading to patient deaths in some cases. The threat to patient safety posed by habitual override or disregard of clinical alerts is not unique to UCSF Medical Center. Between 2005 and 2010, alert fatigue led to approximately 216 deaths in the United States (Wachter, 2017).

Strategies for eliminating alert fatigue include enhancing the specificity of alerts by designing clinical decision tools that generate fewer alerts and tailoring them to specific practice settings and user groups (Kesselheim et al., 2011). This enhancement would significantly decrease the number of insignificant alerts and result in fewer overrides, which could potentially eliminate inadvertent overrides of clinically significant alerts.

Lack of capital is another important barrier to the implementation of clinical decision support tools. It is estimated that the cost of implementing a CPOE is $2.8 million, with annual maintenance costs of approximately $500,000 (Bates et al., 1998). Many institutions, especially smaller facilities, find this cost prohibitive. In addition to the financial costs, health care institutions incur other costs in the form of lost productivity and man-hours required for training. However, studies have shown that the long-term benefits of clinical decision support programs offset the initial costs within a few years. At Brigham and Women's Hospital, where preventable ADEs cost the hospital approximately $3 million per year, implementation of CPOE resulted in a 19% decrease in preventable ADEs, leading to annual savings of $500,000 (Bates et al., 1998). The hospital also saved on malpractice costs and patient injuries, leading to improved quality of care and patient safety and additional financial savings. In fact, by some estimates, CPOE-related prevention of ADEs could lead to annual savings of $5 to $10 million for hospitals (Bates et al., 1998). A crucial step toward obtaining capital to fund adoption of clinical decision support tools is to obtain the support of key stakeholders, such as hospital and institution administrators (Scheepers-Hoeks et al., 2011; Trivedi et al., 2009). The support of these personnel ensures that the input of end-users will be taken into consideration in designing clinical decision tools and that clinicians have advocates at upper levels of management promoting the implementation of software that will facilitate patient care.

In many institutions, clinicians may be reluctant to incorporate clinical decision support tools into their practice due to: lack of confidence in computerized decision support systems, reluctance to learn a new operating system, and perceived loss of autonomy (Trivedi et al., 2009). In 2011, Steve Balt, a physician who once lauded the benefits of clinical decision programs, stated:

While I appreciate the ability to enter patient data (and my notes) into a system that is instantly accessible by any provider in my office at any time, and write prescriptions with a few clicks of my mouse, I've begun to resent the ways in which EHRs tell me how to practice, particularly when (a) they give recommendations that I would employ anyway (thereby wasting my time), or (b) they give recommendations that deviate from what I believe is right for the patient.

Many institutions attempt to address the problem of alert fatigue by requiring pharmacists to review all alerts and make recommendations to physicians. In one RCT at an academic ED in Indiana, CPOE was used for all prescriptions. During the study, pharmacists reviewed patients' medication profiles and provided 114 recommendations to physicians, of which only 43% were accepted and implemented (Terrell et al., 2009). When physicians implemented pharmacists' recommendations, the prescription of HRMs decreased from 5.4% to 3.4% (Terrell et al., 2009). To help increase clinician buy-in, administrators must include end-users in the planning and development of CDSS. Not only would this help identify alerts that are relevant to various teams within institutions, but it would also help determine where a support tool will fit into the work-flow (Klein, 2013). Obtaining the input of end-users throughout the planning and design states would help ensure the development of tailored alerts, which would in turn decrease the number of clinically insignificant alerts, thereby helping decrease alert fatigue, an important obstacle to successful implementation of clinical support tools.

Implications for Practice and Future Directions

The benefits of clinical decision support tools have been clearly demonstrated by numerous studies: their use has led to decreased dosing and prescribing errors, decreased HRM use, fewer side effects, improved patient care, and financial savings for patients and health care systems. However, many obstacles—including lack of finances, alert fatigue, and lack of buy-in from clinicians—have hindered their widespread adaptation. With the use of electronic health systems becoming the norm, it is imperative that hospitals and health care systems integrate clinical decision support into their practices to optimize patient care. It is equally important to provide adequate training to physicians, nurses, pharmacists, and other professionals involved in the care of older adults on the risks associated with HRM use. More often than not, these professionals are the end-users of CDSS, and their input must be obtained and incorporated into the design of CDSS as a means to optimize their use in patient care.

Currently, there are few studies outlining effective ways for hospitals and health care systems to overcome these challenges to benefit from the integration of information technology into patient care. Furthermore, studies on the impact on nurses' use of CDSS have shown mixed results (Randell, Mitchell, Dowding, Cullum, & Thompson, 2007). Additional studies on this area are required and will assist health care facilities in designing and implementing effective clinical decision tools to improve patient safety. Further studies are also needed to determine ways to increase physician acceptance of pharmacists' recommendations for changes to patients' medication regimens.

References

  • American Society of Consultant Pharmacists Foundation. (2017). GRAM Research Project. Retrieved from http://www.ascpfoundation.org/programs/gram_research.cfm
  • Arvisais, K., Bergeron-Wolff, S., Bouffard, C., Michaud, A., Bergeron, J., Mallet, L. & Cossette, B. (2015). A pharmacist–physician intervention model using a computerized alert system to reduce high-risk medication use in elderly inpatients. Drugs & Aging, 32, 663–670. doi:10.1007/s40266-015-0286-5 [CrossRef]
  • Bates, D.W., Leape, L.L., Cullen, D.J., Laird, N., Petersen, L.A., Teich, J.M. & Seger, D.L. (1998). Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA, 280, 1311–1316. doi:10.1001/jama.280.15.1311 [CrossRef]
  • Forster, A.J., Erlanger, T.T., Jennings, A., Auger, C., Buckeridge, D., van Walraven, C. & Tamblyn, R. (2015). Effectiveness of a computerized drug-monitoring program to detect and prevent adverse drug events and medication non-adherence in outpatient ambulatory care: Study protocol of a randomized controlled trial. Trials, 16, 2. doi:10.1186/1745-6215-16-2 [CrossRef]
  • Ghibelli, S., Marengoni, A., Djade, C., Nobili, A., Tettamanti, M., Franchi, C. & Pasina, L. (2013). Prevention of inappropriate prescribing in hospitalized older patients using a computerized prescription support system (INTERcheck®). Drugs & Aging, 30, 821–828. doi:10.1007/s40266-013-0109-5 [CrossRef]
  • Gray, S.L., Sager, M., Lestico, M.R. & Jalaluddin, M. (1998). Adverse drug events in hospitalized elderly. Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 53, M59–M63. doi:10.1093/gerona/53A.1.M59 [CrossRef]
  • Institute for Safe Medication Practices. (2017). ISMP high-alert medications. Retrieved from http://www.ismp.org/Tools/highAlertMedicationLists.asp
  • Kesselheim, A.S., Cresswell, K., Phansalkar, S., Bates, D.W. & Sheikh, A. (2011). Clinical decision support systems could be modified to reduce “alert fatigue” while still minimizing the risk of litigation. Health Affairs, 30, 2310–2317. doi:10.1377/hlthaff.2010.1111 [CrossRef]
  • Klein, C. (2013). Changing before our eyes: An evolution of clinical decision support in healthcare. Retrieved from https://pdfs.semanticscholar.org/f07e/f218efca75f5deecefea054e2c98226030c5.pdf
  • Klopotowska, J.E., Wierenga, P.C., Stuijt, C.C., Arisz, L., Dijkgraaf, M., Kuks, P. & Smorenburg, S.M. (2013). Adverse drug events in older hospitalized patients: Results and reliability of a comprehensive and structured identification strategy. PLoS One, 8, e71045. doi:10.1371/journal.pone.0071045 [CrossRef]
  • Kuperman, G.J., Bobb, A., Payne, T.H., Avery, A.J., Gandhi, T.K., Burns, G. & Bates, D.W. (2007). Medication-related clinical decision support in computerized provider order entry systems: A review. Journal of the American Medical Informatics Association, 14, 29–40. doi:10.1197/jamia.M2170 [CrossRef]
  • Lee, J., Han, H., Ock, M., Lee, S.I., Lee, S. & Jo, M.W. (2014). Impact of a clinical decision support system for high-alert medications on the prevention of prescription errors. International Journal of Medical Informatics, 83, 929–940. doi:10.1016/j.ijmedinf.2014.08.006 [CrossRef]
  • Makris, U., Mortensen, E., Pugh, M., Alvarez, C., Turner, B., Aung, K. & Berlowitz, D. (2015). Exposure to high-risk medications is associated with worse outcomes in older veterans with chronic pain. American Journal of the Medical Sciences, 350, 279–285. doi:10.1097/MAJ.0000000000000552 [CrossRef]
  • Peterson, J.F., Kripalani, S., Danciu, I., Harrell, D., Marvanova, M., Mixon, A. & Powers, J.S. (2014). Electronic surveillance and pharmacist intervention for vulnerable older inpatients on high-risk medication regimens. Journal of the American Geriatrics Society, 62, 2148–2152. doi:10.1111/jgs.13057 [CrossRef]
  • Randell, R., Mitchell, N., Dowding, D., Cullum, N. & Thompson, C. (2007). Effects of computerized decision support systems on nursing performance and patient outcomes: A systematic review. Journal of Health Services Research & Policy, 12, 242–249. doi:10.1258/135581907782101543 [CrossRef]
  • Scheepers-Hoeks, A.M., Grouls, R.J., Neef, C., Ackerman, E.W. & Korsten, E. (2011). Success factors and barriers for implementation of advanced clinical decision support systems. In Jao, C.S. (Ed.), Efficient decision support systems: Practice and challenges in biomedical related domain. Rijeka, Croatia: InTech.
  • Terrell, K.M., Perkins, A.J., Dexter, P.R., Hui, S.L., Callahan, C.M. & Miller, D.K. (2009). Computerized decision support to reduce potentially inappropriate prescribing to older emergency department patients: A randomized, controlled trial. Journal of the American Geriatrics Society, 57, 1388–1394. doi:10.1111/j.1532-5415.2009.02352.x [CrossRef]
  • Trivedi, M.H., Daly, E.J., Kern, J.K., Grannemann, B.D., Sunderajan, P. & Claassen, C.A. (2009). Barriers to implementation of a computerized decision support system for depression: An observational report on lessons learned in “real world” clinical settings. BMC Medical Informatics and Decision Making, 9, 6. doi:10.1186/1472-6947-9-6 [CrossRef]
  • Van der Sijs, H., Aarts, J., Vulto, A. & Berg, M. (2006). Overriding of drug safety alerts in computerized physician order entry. Journal of the American Medical Informatics Association, 13, 138–147. doi:10.1197/jamia.M1809 [CrossRef]
  • Wachter, B. (2017). Beware of the robot pharmacist. Retrieved from https://backchannel.com/beware-of-the-robot-pharmacist-4015ebf13f6f

Examples of Clinical Decision Support Tools

TypeExampleResource
Clinical alert system (CAS)aAmerican Geriatrics Society Beers Criteriahttps://geriatricscareonline.org/ProductAbstract/american-geriatrics-society-updated-beers-criteria-for-potentially-inappropriate-medication-use-in-older-adults/CL001
Screening Tool to Alert to Right Treatment/Screening Tool of Older People's Prescriptions (START-STOPP) criteriahttps://www.ncbi.nlm.nih.gov/pubmed/25324330
Clinical decision support system (CDSS)Geriatric Risk Assessment MedGuide (GRAM)https://www.ahrq.gov/downloads/pub/advances/vol4/Lapane.pdf
Information Systems-enabled Outreach for Preventing Adverse Drug Events (ISTOP-ADE)https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4326368
CDSS-CASHigh Alert Medication Recognition and Management System to Lower Errors and Secure Safety (HARMLESS)https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5391194

Success Factors for Implementation of Clinical Decision Support Systems

FactorComponent
Right messageAccurate content
Reliable messages
Easy and actionable messages
Inclusion of references in the message
Right timeSave time
Integration into workflow
High systems speed
Right placeDeliver message at the point of care
Active alerting mechanism
Right systemElectronic availability of data in electronic medical record
Integration with other systems
Maintenance of the system and content
Authors

Dr. Lord-Adem is Post-Doctoral Fellow, and Dr. Brandt is Professor, University of Maryland School of Pharmacy; Dr. Brandt is also Executive Director, Peter Lamy Center on Drug Therapy and Aging, Baltimore, Maryland.

The authors have disclosed no potential conflicts of interest, financial or otherwise.

The authors thank Drs. Monica Tong and Stephanie Ozaleas for their contributions to the intellectual content of the manuscript and their dedication to furthering medication use safety for older adults.

Address correspondence to Wilhelmina Lord-Adem, PharmD, Post-Doctoral Fellow, University of Maryland School of Pharmacy, 220 Arch Street SGO-125, Baltimore, MD 21201; e-mail: wilhelmina.adem@umaryland.edu.

10.3928/00989134-20170614-04

Sign up to receive

Journal E-contents