Drugs Aging (2015) 32:227–233 DOI 10.1007/s40266-015-0244-2
SHORT COMMUNICATION
Impact of Computerized Physician Order Entry Alerts on Prescribing in Older Patients Paula E. Lester • Liliana Rios-Rojas • Shahidul Islam • Melissa J. Fazzari • Irving H. Gomolin
Published online: 10 March 2015 Ó Springer International Publishing Switzerland 2015
Abstract Background A computerized physician order entry (CPOE) system provides opportunity for real-time alerts to prescribers. Winthrop University Hospital began using CPOE in 2009. Objective We sought to improve prescribing among older hospitalized patients by adding alerts to the CPOE system for potentially inappropriate medications. Methods In January 2011, informational alerts were integrated into the CPOE system for selected high-risk medications: diphenhydramine, metoclopramide, and all antipsychotics. We evaluated the effect of these alerts on prescribing frequency by comparing the number of
Electronic supplementary material The online version of this article (doi:10.1007/s40266-015-0244-2) contains supplementary material, which is available to authorized users. P. E. Lester (&) L. Rios-Rojas Division of Geriatric Medicine, Winthrop University Hospital, 222 Station Plaza North, Suite 518, Mineola, NY 11501, USA e-mail: [email protected] P. E. Lester I. H. Gomolin Clinical Medicine, Stony Brook University Medical Center, Stony Brook, NY, USA S. Islam M. J. Fazzari Department of Biostatistics, Winthrop University Hospital, Mineola, NY, USA I. H. Gomolin Divisions of Geriatric Medicine and Clinical Pharmacology, Winthrop University Hospital, Mineola, NY, USA
prescriptions during the second quarters of 2010 (‘‘prealert’’) with the second quarters of 2011 through 2013 (‘‘post-alert’’). Prescribing patterns were evaluated through a pharmacy database of medication orders. Frequency of prescribing was adjusted for total discharges. A comparison was made to ages 18–64 years, and comparing ‘‘as needed’’ vs standing orders. Results In the 65 years of age and older group, there were significant reductions in prescription rates pre-alert vs post-alert for diphenhydramine (p \ 0.001) and metoclopramide (p \ 0.001). There was no decrease in prescription rates for antipsychotics in older patients (p = 0.80). In the younger comparison group, no decreases in prescription rates for those drugs were observed. Our analysis is based on numbers of written prescriptions and not actual doses administered; therefore, no conclusions concerning the effect of these alerts on communication or documentation of risk/benefits of these medications can be ascertained. Conclusion The data suggest that prescribing rates for drugs with the least efficacy and potential for harm and with alternative agents (i.e., diphenhydramine and metoclopramide) can be modified by CPOE alerts for older patients.
Key Points Real-time alerts for high-risk drugs in older patients can impact prescribing patterns. Alerts for high-risk medications are effective in altering prescribing behaviors when alternative agents are available.
228
1 Background Previous studies have demonstrated the impact of computerized physician order entry (CPOE) alerts to improve medication safety [1, 2]. One review of 10 studies found that the biggest impact occurred when the alerts were targeted at a limited number of potentially inappropriate medications [2]. Real-time computer reminders have been shown to reduce the use of potentially inappropriate sedative hypnotics through an alert with suggested alternatives in the acute hospital setting [3]. The Beers criteria for potentially inappropriate medication use identifies medications whose risks may outweigh their potential benefits for people aged 65 years and older [4]. This information can help prevent harmful medication adverse effects especially when safer alternatives may be available [1]. In 2010, Mattison et al. developed alerts for 16 drugs based on the Beer’s criteria with a recommendation ‘‘not to order’’ these drugs and for two drugs (lorazepam and ferrous sulfate) with recommendations for a decreased dose. They found a highly significant and immediately observed drop in the rate of use of ‘‘not recommended medications’’ (p \ 0.001) [1]. We developed ‘‘geriatric alerts’’ for selected medications when ordered in our hospital CPOE system for patients aged 65 years or older. These alerts included links to appropriate reference literature. We selected the following: diphenhydramine, metoclopramide, and antipsychotics (see Electronic Supplementary Material Table S1 for a list of included medications and the alert text). When ordering one of these drugs for anyone aged 65 years or older, a pop-up alert appears containing drugor class-specific text. Prescribers were required to click the pop-up box to acknowledge the alert and to proceed to the order entry. CPOE did not prevent ordering of the medication nor did it require specific documentation. Unlike the Mattison et al. study, we selected medications based on the high-risk potential and used Beer’s criteria as supporting evidence when available. Mattison et al. included diphenhydramine, but did not include metoclopramide. The only antipsychotic they included was thioridazine, which is a typical (first-generation) antipsychotic, whereas we included typical and atypical antispychotics, which are more widely used than thioridazine [1]. Diphenhydramine is on the Beer’s list and carries the following warning: highly anticholinergic; clearance reduced with advanced age, and tolerance develops when used as hypnotic; increased risk of confusion, dry mouth, constipation, and other anticholinergic effects/toxicity (strength of recommendation: strong) [4–6]. Urinary retention is a potentially serious complication of diphenhydramine use as well [7]. Diphenhydramine is commonly prescribed in a hospital setting to decrease transfusion
P. E. Lester et al.
reactions, but studies of its use for the prevention of allergic and febrile non-hemolytic transfusion reactions have not shown effectiveness [4, 8, 9]. Diphenhydramine is also used in the hospital setting as a sedative or hypnotic; however, there are safer alternatives available as a sleep aid [10, 11]. Research in younger healthy men shows that antihistamines with high blood–brain barrier-crossing efficiency (such as diphenhydramine) were significantly associated with sleepiness and psychomotor performance decline the next day [12]. It is likely that such effects would be even greater in the older patients. Metoclopramide is an anti-emetic and gastroprokinetic agent [13]. It can also be used shortterm to treat heartburn caused by gastroesophageal reflux in people who have used other medications without relief of symptoms. Metoclopramide can cause extrapyramidal effects including tardive dyskinesia and the risk may be further increased in frail older adults (strength of recommendation: strong) [4, 13]. The risk of extrapyramidal effects increases with duration of use and is greatest in women, diabetic patients, and older adults. In fact, a US Food and Drug Administration (FDA) black-box warning specifies that ‘‘the elderly, especially elderly women, are most likely to develop this [TD] condition’’ [13]. There are alternative anti-emetic agents and anti-reflux agents available with safer side-effect profiles [14–16]. While there are no alternative pro-kinetic agents for gastroparesis, there are non-pharmacologic treatments (e.g., smaller meals) and safer anti-emetics for the relief of nausea [16]. Antipsychotics are widely used in geriatric psychiatric disorders, especially for delirium-related agitation in the hospital setting [17]. Although there is no psychotropic medication that is approved by the FDA for the treatment of psychosis in dementia, a number of consensus statements have recommended the use of atypical antipsychotic agents as a first-line pharmacologic approach to treatment [17]. However, there is evidence of an increased risk of stroke and death in patients treated with either typical or atypical antipsychotics and the FDA has mandated a blackbox warning that ‘‘elderly patients with dementia-related psychosis treated with antipsychotic drugs are at an increased risk of death’’ [18, 19]. Antipsychotics were added to the Beer’s criteria as a medication to avoid in dementia patients (strength of recommendation: strong) [4]. Other adverse outcomes are attributed to antipsychotics as well. A recent editorial reveals that observational studies conducted to date demonstrate a two- to threefold increased risk of pneumonia in a dose-dependent fashion with conventional or atypical antipsychotics compared with nonuse, with a possibly higher risk attributable to atypical antipsychotics [20]. Additionally, the risk of venous thromboembolism is increased with typical and atypical antipsychotics, though the actual rate varies with the type
CPOE Alerts and High-Risk Medications in Older Patients
of drug and is highest just after starting the antipsychotic [21]. Oral atypical antipsychotic medication use has been shown to increase the risk for acute kidney injury requiring hospitalization [22].
2 Objectives Our study aimed to evaluate the impact of these ‘‘geriatric alerts’’ in the CPOE on ordering patterns of diphenhydramine, metoclopramide, and antipsychotics in the hospital assessing prescription rates as measured by the number of prescriptions divided by the number of hospital discharges in the second quarter of a given calendar year over a 3-year period. We compared rates of ordering before the alert (2010) and after the alerts were implemented in January of 2011 (2011–2013). Our goal was to reduce the use of these medications in the older patients through increased awareness of their potential adverse outcomes supported by published evidence. We intentionally did not suggest alternative pharmacologic agents or non-pharmacologic interventions as each patient has specific needs and a medical history and requires treatment catered to their individual health status. Furthermore, because these highrisk drugs are inherently dangerous to use, substitution with another agent by definition, would most likely be less dangerous.
3 Methods Winthrop Hospital is a community teaching institution located in Long Island, NY, USA with residencies, fellowships, and a physician’s assistant/nurse practitioner service. All clinicians, including attending physicians are able to enter orders through CPOE. CPOE became operational by 2009. In January 2011, a series of alerts were built into CPOE for selected medications. These pop-up alerts contained links to articles relevant to the alert [4, 9, 23] (see Electronic Supplementary Material Table S1). The alerts require the user to click ‘‘noted’’ and then the user either cancels their order or continues with prescribing the medication. The alert does not require explanation for the use of the medication. See Electronic Supplementary Material Table S2 for a snapshot of a diphenhydramine alert. We evaluated the effect of these alerts on prescribing frequency for diphenhydramine, metoclopramide, and all antipsychotics by comparing the number of patients for whom these medications were prescribed during the second quarters of 2010 through 2013. Prescribing patterns were evaluated by ascertaining the pharmacy database, which contained all medication orders since the introduction of CPOE. If patients were prescribed medication of the same
229
type multiple times during their admission, only the first occurrence was counted in the total number prescribed. This was done to avoid counting medications multiple times for the same patient. Because the alerts went into effect in January of 2011, we compared 2010 data as the baseline to use in years 2011 through 2013. Prescription rates for each year were computed as the number of prescriptions adjusted by the total number of discharges during each second quarter of interest. We also compared prescribing for the same time period for patients aged 18–64 years to serve as a comparison group. In secondary analyses, older patients were further stratified by age (65–74 years, 75–84 years, 85 years and over). We hypothesized that use in the younger comparison group (who did not receive the alerts when ordering medications) would not result in a reduction in prescribed numbers. Additional analysis was performed to assess changes in prescribing patterns over time such as transition to more PRN (as needed) orders instead of standing orders. The observed prescription rates for diphenhydramine, metoclopramide, and all antipsychotics were analyzed via Poisson regression models, using the natural log of the number of discharges as the offset, and including time and age as model covariates. Specific trends in prescription rates over time were examined: whether there was a reduction in the average 3-year (post-alert) rate from 2011 to 2013 compared with pre-alert (2010) levels by age group, a reduction from 2010 when comparing only 2013 rates, and a linear reduction in rates over time (2010–2013) by age group. One-sided p values corresponding to t tests for each effect of interest were presented and values B0.025 were considered statistically significant. One-sided analysis was used as the study was designed to assess for a decrease in prescription rates. Data were analyzed using SAS 9.2 (SAS Institute, Cary, NC, USA). Institutional review board approval was not obtained for this study as it was initially designed as a quality improvement project.
4 Results Our results are notable because the alerts only reduced prescription rates for medications for which there were safer alternatives (metoclopramide and diphenhydramine for sleep) or the medication was not needed (diphenhydramine as a pre-medication for blood transfusions). Antipsychotic orders were not reduced, probably because providers lack alternative treatments for psychosis and delirium. We also found no carry-over effect to the younger population, which suggests the alert needs to be specific to patient characteristics. Among patients aged older than 65 years, there were statistically significant reductions in prescription rates following the introduction of
230
P. E. Lester et al.
Fig. 1 Prescription rates of the three drug classes in 65 and older cohort
Fig. 2 Prescription rate for diphenhydramine stratified by age
these alerts for diphenhydramine (26.9 % in 2010 vs 20 % in 2013; p \ 0.001) and metoclopramide (16.7 % in 2010 vs 12.5 % in 2013; p \ 0.001). There was no decrease observed in the rate of prescriptions for antipsychotics in the older patients (8.8 % in 2010 vs 9.2 % in 2013; p = 0.80) (Fig. 1). In the younger group (B65 years old), no decrease in prescription rates for metoclopramide, diphenhydramine, or antipsychotics were observed. See Table 1 for hospital discharges by age group and year. Analysis of the older cohort with additional age stratification was performed with groups defined as ages
65–74 years, 75–84 years, and 85 years and over. This analysis revealed a greater impact of the alert on the reduction of rate of prescriptions for diphenhydramine (Fig. 2) and metoclopramide (Fig. 3) for those with advancing age. Those aged 85 years and over had a larger absolute reduction in prescription rates for diphenhydramine (22.5 % in 2010 vs 12.5 % in 2013; p \ 0.001), compared with those aged 75–84 years (24.2 vs 18.4; p \ 0.001) and aged 65–74 years (30.5 vs 27.8; p = 0.12). Similarly, for metoclopramide, those aged 85 years and over had the largest rate reduction (11.8 % in 2010 vs 5.6 % in 2013; p \ 0.001), compared with those aged 65–74 where rates remained stable (20.6 vs 21.4; p = 0.68). Analysis of antipsychotics in 2010 (pre-alert) showed the highest prescription rate in those aged 85 years and over (14.4 %), with lower rates seen in each successively younger age group (p \ 0.001; prescription rates are: 7.4 % in 75- to 84-year-old individuals, and 4.1 % in 65- to 74-year-old individuals, and 2.6 % in \65-year-old individuals) (Fig. 4). Secondary analysis revealed no significant change in types of orders from 2010 to 2013, i.e., there was no significant decrease in standing orders or increase in PRN order rates (data not shown).
Table 1 Hospital discharges by age group and year (second quarter) Year
Age group (years)
Hospital discharges (n)
2010
\65
4243
65–74
1036
75–84
1325
2011
2012
2013
85?
898
\65
4412
65–74
1107
75–84 85?
1304 1000
\65
3909
65–74
1065
75–84
1113
85?
873
\65
4051
65–74
1089
75–84
1127
85?
913
5 Discussion We found a statistically significant decrease in the number of orders for patients aged 65 years and over for diphenhydramine and metoclopramide after implementation of CPOE ‘‘geriatric alerts’’. There was no reduction in the rate of prescription for these drugs in our comparison group of 18- to 64-year-old individuals. This demonstrates that alerts of this nature can have a meaningful impact on
CPOE Alerts and High-Risk Medications in Older Patients
Fig. 3 Prescription rate for metoclopramide stratified by age
Fig. 4 Prescription rate for antipsychotics stratified by age
prescriber patterns. Perhaps the availability of alternative agents helped reduce use. For example, there are safer options than diphenhydramine to treat insomnia and safer options then metoclopramide for nausea/vomiting in the older population. Additionally, the alert provided specific reference to lack of efficacy for the use of diphenhydramine as prophylaxis prior to blood transfusions. Our study is unique because it was not exclusively based on Beer’s criteria (as in Mattison et al. [1)] and it did not force users to obtain approval to use the medications or suggest alternative agents (as in Agostini et al. [3]). There was no reduction in antipsychotic order prescription rate for older patients as a result of the ‘‘geriatric alerts’’. For some patients, the antipsychotic may have been continued from outpatient regimens, and might have been
231
inappropriate to discontinue during an acute hospital stay. In such cases, the onus should be on the original prescribing provider to address potential adverse effects including the FDA warning with caregivers or patients. Moreover, antipsychotics are often ordered in the hospital because of acute aggressive, psychotic, or delirious symptoms. As such, prescribing patterns were likely not to be reduced as the drugs are used for an acute symptom without safer and established alternatives to use. This is consistent with nursing home data that showed no reduction in the use of antipsychotics since the FDA black-box warning [24]. Our study was not designed to evaluate whether the alert increased the discussion or documentation of the FDA black-box warning with caregivers or if additional non-pharmacologic interventions were attempted. Subgroup analysis showed a lower use for diphenhydramine and metoclopramide in the oldest group of patients aged 85 years and older, with incrementally higher rates of use in younger age groups. This suggests that the older patients are more frail and at greater peril for adverse reactions from these potentially high-risk medications. In comparing antipsychotic use in those patients aged 85 years and older from 2011 to 2013, the linear trend downward is statistically significant and not seen in the younger age groups. The higher prescription rate of antipsychotics in the 85 years of age and older group is not surprising because the highest rates of dementia and delirium are in the older population and thus they are more likely to have behavioral symptoms for which these drugs are prescribed. In August 2012, consistent with newly revised Beer’s criteria [4], we revised the antipsychotic alert (see Electronic Supplementary Material Table S1). In 2013, there was no significant drop in antipsychotic use among patients aged older than 65 years even after this alert was implemented. Our findings suggest that despite awareness of the risks of antipsychotics, these drugs are used in hospital because of perceived drug efficacy and lack of satisfactory alternatives [25]. There was no carry-over effect of reducing prescription rates for orders of diphenhydramine, metoclopramide, or the antipsychotics in the comparison group of individuals aged 18–64 years. This suggests that such alerts are seen as targeted specifically to the older patient population aged over 65 years. Of note, the alerts we used for metoclopramide and the antipsychotics advise caution about long-term use. Because our study did not evaluate the number of days these medications were prescribed for, or their use following hospitalization, we cannot assess whether long-term use of metoclopramide or antipsychotics were influenced by the alert. Our study was not designed to assess whether there
232
was better documentation about informing patients of the FDA warning on metoclopramide or antipsychotics. Better documentation and more discussion with patients about the risks associated with these drugs would be another benefit of the alerts, even if there were no reduction in hospital prescribing. There was no shift in prescribing patterns from standing to PRN as a result of the alerts. One might have expected a shift to PRN as a way for providers to try to limit use of high-risk drugs, but still have them available for patients. The lack of shift suggests that providers recognize the risk of the drugs (whether administered standing or PRN) and choose alternatives instead, when possible. Our study has limitations. The analysis is based on number of written prescriptions and not actual doses administered. Additionally, the study was not designed to include a chart review so no conclusions concerning the effect of these alerts on communication or documentation of risk/benefits of these medications can be ascertained. We did not assess clinical outcomes as to effectiveness of alternative hypnotics or anti-emetics. However, the reduction of diphenhydramine and metoclopramide orders implies that patients were managed without them. Of note, there were no specific educational programs targeting any of these medications in the older patients, which might have impacted prescribing patterns. Like similar studies, we also lack the ability to determine if adverse drug events were reduced by use of this warning system and whether the medications that were ordered were in fact, clinically required.
6 Conclusion This study demonstrates that CPOE alerts for older hospitalized patients can be used to increase awareness about the efficacy and risk of certain medications and can have a significant impact on drug prescribing rates for some medications, notably those with alternative available agents. Such alerts can be targeted to high-risk drugs in vulnerable populations and can include supportive references. It is also important to keep the resource links updated. It is imperative to carefully select the medications to receive such alerts to prevent ‘‘alert fatigue,’’ in which the users actually cease reading the alerts because they are overwhelmed by a multitude of them. Future research may show the impact on clinical outcomes and should include assessment of the impact of the alerts on the documentation of discussions with patients and caregivers about the risks and benefits of these medications. Acknowledgments The authors gratefully acknowledge the contributions of Steve Gerson RPh, MPA of Pharmacy, and Denise Strand
P. E. Lester et al. BSN RN of Clinical Decision Support for their contributions to this project. There was no funding for this project. Contributions: All authors provided substantial contributions to the conception and design of the work, and the acquisition, analysis, and interpretation of data. Dr. Lester drafted the manuscript with critical revisions provided for important intellectual content by all other authors. All authors approve the manuscript version submitted. The authors have no potential conflicts of interest that are directly relevant to the content of this study.
References 1. Mattison MLP, Afonso KA, Ngo LH, Mukamal KJ. Preventing potentially inappropriate medication use in hospitalized older patients with a computerized provider order entry warning system. Arch Intern Med. 2010;170:1331–6. 2. Lainer M, Mann E, So¨nnichsen A. Information technology interventions to improve medication safety in primary care: a systematic review. Int J Qual Health Care. 2013;25(5):590–8. 3. Agostini JV, Zhang Y, Inouye SK. Use of a computer-based reminder to improve sedative-hypnotic prescribing in older hospitalized patients. J Am Geriatr Soc. 2007;55(1):43–8. 4. American Geriatrics Society 2012 Beers Criteria Update Expert Panel. American Geriatrics Society updated Beers criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2012;60:1–15. 5. Ochs KL, Zell-Kanter M, Mycyk MB. Hot, blind, and mad: avoidable geriatric anticholinergic delirium. Am J Emerg Med. 2012;30(3):514. 6. Fortin MP, Rouch I, Dauphinot V, et al. Effects of anticholinergic drugs on verbal episodic memory function in the elderly: a retrospective, cross-sectional study. Drugs Aging. 2011;28:195–204. 7. Shimizu T, Tokuda Y. Acute urinary retention. Intern Med. 2011;19(1):2255. 8. Kennedy LD, Case LD, Hurd DD, et al. A prospective, randomized, double-blind controlled trial of acetaminophen and diphenhydramine pretransfusion medication versus placebo for the prevention of transfusion reactions. Transfusion. 2008;48:2285–91. 9. Wang SE, Lara PN Jr, Lee-Ow A, et al. Acetaminophen and diphenhydramine as premedication for platelet transfusions: a prospective randomized double-blind placebo-controlled trial. Am J Hematol. 2002;70:191–4. 10. Montgomery P, Lilly J. Insomnia in the elderly. Clin Evid. 2007;2007:2302. 11. Tu RH, Grewall P, Leung JW, et al. Diphenhydramine as an adjunct to sedation for colonoscopy: a double-blind randomized, placebo-controlled study. Gastrointest Endosc. 2006;63:87–94. 12. Katayose Y, Aritake S, Kitamura S, et al. Carryover effect on next-day sleepiness and psychomotor performance of nighttime administered antihistaminic drugs: a randomized controlled trial. Hum Psychopharmacol. 2012;27(4):428–36. 13. REGLAN Injection (metoclopramide injection, USP) Rx only FDA. [Online]. 2013. http://www.accessdata.fda.gov/drugsatfda_ docs/label/2010/017862s063lbl.pdf. Accessed 10 Dec 2013. 14. Jordan K, Gralla R, Jahn F, Molassiotis A. International antiemetic guidelines on chemotherapy induced nausea and vomiting (CINV): content and implementation in daily routine practice. Eur J Pharmacol. 2013;722:197–202. 15. Kovac AL. Update on the management of postoperative nausea and vomiting. Drugs. 2013;73:1525–47. 16. Scorza K, Williams A, Phillips JD, Shaw J. Evaluation of nausea and vomiting. Am Fam Physician. 2007;76:76–84.
CPOE Alerts and High-Risk Medications in Older Patients 17. Alexopoulos GS, Streim J, Carpenter D, Docherty JP. Using antipsychotic drugs in older patients. J Clin Psychiatry. 2004;65 (Suppl 2):5–99. 18. US FDA: Postmarket drug safety information for patients and providers–public health advisory: deaths with antipsychotics in elderly patients with behavioral disturbances. [Online]. 2010. http:// www.fda.gov/drugs/drugsafety/postmarketdrugsafetyinformation forpatientsandproviders/ucm053171. Accessed 17 Feb 2015. 19. Schneider LS, Dagerman KS, Insel P. Risk of death with atypical antipsychotic drug treatment for dementia: meta-analysis of randomized placebo-controlled trials. JAMA. 2005;294:1934–43. 20. Gambassi G, Sultana J, Trifiro` G. Antipsychotic use in elderly patients and the risk of pneumonia. Expert Opin Drug Saf. 2015;14(1):1–6. 21. Ishiguro C, Wang X, Li L, Jick S. Antipsychotic drugs and risk of idiopathic venous thromboembolism: a nested case-control study
233
22.
23.
24.
25.
using the CPRD. Pharmacoepidemiol Drug Saf. 2014;23(11): 1168–75. Hwang YJ, Dixon SN, Reiss JP, et al. Atypical antipsychotic drugs and the risk for acute kidney injury and other adverse outcomes in older adults: a population-based cohort study. Ann Intern Med. 2014;161(4):242–8. Fick DM, Cooper JW, Wade WE, et al. Updating the Beers criteria for potentially inappropriate medication use in older adults. Arch Intern Med. 2003;163:2716–24. Lester P, Kohen I, Stefanacci RG, Feuerman M. Antipsychotic drug use since the FDA black box warning: survey of nursing home policies. JAMA. 2011;18:573–7. Zirker W, Dorokhine I, Knapp CM, et al. Haloperidol overdosing in the treatment of agitated hospitalized older people with delirium: a retrospective chart review from a community teaching hospital. Drugs Aging. 2013;3:639–44.
SHORT COMMUNICATION
Impact of Computerized Physician Order Entry Alerts on Prescribing in Older Patients Paula E. Lester • Liliana Rios-Rojas • Shahidul Islam • Melissa J. Fazzari • Irving H. Gomolin
Published online: 10 March 2015 Ó Springer International Publishing Switzerland 2015
Abstract Background A computerized physician order entry (CPOE) system provides opportunity for real-time alerts to prescribers. Winthrop University Hospital began using CPOE in 2009. Objective We sought to improve prescribing among older hospitalized patients by adding alerts to the CPOE system for potentially inappropriate medications. Methods In January 2011, informational alerts were integrated into the CPOE system for selected high-risk medications: diphenhydramine, metoclopramide, and all antipsychotics. We evaluated the effect of these alerts on prescribing frequency by comparing the number of
Electronic supplementary material The online version of this article (doi:10.1007/s40266-015-0244-2) contains supplementary material, which is available to authorized users. P. E. Lester (&) L. Rios-Rojas Division of Geriatric Medicine, Winthrop University Hospital, 222 Station Plaza North, Suite 518, Mineola, NY 11501, USA e-mail: [email protected] P. E. Lester I. H. Gomolin Clinical Medicine, Stony Brook University Medical Center, Stony Brook, NY, USA S. Islam M. J. Fazzari Department of Biostatistics, Winthrop University Hospital, Mineola, NY, USA I. H. Gomolin Divisions of Geriatric Medicine and Clinical Pharmacology, Winthrop University Hospital, Mineola, NY, USA
prescriptions during the second quarters of 2010 (‘‘prealert’’) with the second quarters of 2011 through 2013 (‘‘post-alert’’). Prescribing patterns were evaluated through a pharmacy database of medication orders. Frequency of prescribing was adjusted for total discharges. A comparison was made to ages 18–64 years, and comparing ‘‘as needed’’ vs standing orders. Results In the 65 years of age and older group, there were significant reductions in prescription rates pre-alert vs post-alert for diphenhydramine (p \ 0.001) and metoclopramide (p \ 0.001). There was no decrease in prescription rates for antipsychotics in older patients (p = 0.80). In the younger comparison group, no decreases in prescription rates for those drugs were observed. Our analysis is based on numbers of written prescriptions and not actual doses administered; therefore, no conclusions concerning the effect of these alerts on communication or documentation of risk/benefits of these medications can be ascertained. Conclusion The data suggest that prescribing rates for drugs with the least efficacy and potential for harm and with alternative agents (i.e., diphenhydramine and metoclopramide) can be modified by CPOE alerts for older patients.
Key Points Real-time alerts for high-risk drugs in older patients can impact prescribing patterns. Alerts for high-risk medications are effective in altering prescribing behaviors when alternative agents are available.
228
1 Background Previous studies have demonstrated the impact of computerized physician order entry (CPOE) alerts to improve medication safety [1, 2]. One review of 10 studies found that the biggest impact occurred when the alerts were targeted at a limited number of potentially inappropriate medications [2]. Real-time computer reminders have been shown to reduce the use of potentially inappropriate sedative hypnotics through an alert with suggested alternatives in the acute hospital setting [3]. The Beers criteria for potentially inappropriate medication use identifies medications whose risks may outweigh their potential benefits for people aged 65 years and older [4]. This information can help prevent harmful medication adverse effects especially when safer alternatives may be available [1]. In 2010, Mattison et al. developed alerts for 16 drugs based on the Beer’s criteria with a recommendation ‘‘not to order’’ these drugs and for two drugs (lorazepam and ferrous sulfate) with recommendations for a decreased dose. They found a highly significant and immediately observed drop in the rate of use of ‘‘not recommended medications’’ (p \ 0.001) [1]. We developed ‘‘geriatric alerts’’ for selected medications when ordered in our hospital CPOE system for patients aged 65 years or older. These alerts included links to appropriate reference literature. We selected the following: diphenhydramine, metoclopramide, and antipsychotics (see Electronic Supplementary Material Table S1 for a list of included medications and the alert text). When ordering one of these drugs for anyone aged 65 years or older, a pop-up alert appears containing drugor class-specific text. Prescribers were required to click the pop-up box to acknowledge the alert and to proceed to the order entry. CPOE did not prevent ordering of the medication nor did it require specific documentation. Unlike the Mattison et al. study, we selected medications based on the high-risk potential and used Beer’s criteria as supporting evidence when available. Mattison et al. included diphenhydramine, but did not include metoclopramide. The only antipsychotic they included was thioridazine, which is a typical (first-generation) antipsychotic, whereas we included typical and atypical antispychotics, which are more widely used than thioridazine [1]. Diphenhydramine is on the Beer’s list and carries the following warning: highly anticholinergic; clearance reduced with advanced age, and tolerance develops when used as hypnotic; increased risk of confusion, dry mouth, constipation, and other anticholinergic effects/toxicity (strength of recommendation: strong) [4–6]. Urinary retention is a potentially serious complication of diphenhydramine use as well [7]. Diphenhydramine is commonly prescribed in a hospital setting to decrease transfusion
P. E. Lester et al.
reactions, but studies of its use for the prevention of allergic and febrile non-hemolytic transfusion reactions have not shown effectiveness [4, 8, 9]. Diphenhydramine is also used in the hospital setting as a sedative or hypnotic; however, there are safer alternatives available as a sleep aid [10, 11]. Research in younger healthy men shows that antihistamines with high blood–brain barrier-crossing efficiency (such as diphenhydramine) were significantly associated with sleepiness and psychomotor performance decline the next day [12]. It is likely that such effects would be even greater in the older patients. Metoclopramide is an anti-emetic and gastroprokinetic agent [13]. It can also be used shortterm to treat heartburn caused by gastroesophageal reflux in people who have used other medications without relief of symptoms. Metoclopramide can cause extrapyramidal effects including tardive dyskinesia and the risk may be further increased in frail older adults (strength of recommendation: strong) [4, 13]. The risk of extrapyramidal effects increases with duration of use and is greatest in women, diabetic patients, and older adults. In fact, a US Food and Drug Administration (FDA) black-box warning specifies that ‘‘the elderly, especially elderly women, are most likely to develop this [TD] condition’’ [13]. There are alternative anti-emetic agents and anti-reflux agents available with safer side-effect profiles [14–16]. While there are no alternative pro-kinetic agents for gastroparesis, there are non-pharmacologic treatments (e.g., smaller meals) and safer anti-emetics for the relief of nausea [16]. Antipsychotics are widely used in geriatric psychiatric disorders, especially for delirium-related agitation in the hospital setting [17]. Although there is no psychotropic medication that is approved by the FDA for the treatment of psychosis in dementia, a number of consensus statements have recommended the use of atypical antipsychotic agents as a first-line pharmacologic approach to treatment [17]. However, there is evidence of an increased risk of stroke and death in patients treated with either typical or atypical antipsychotics and the FDA has mandated a blackbox warning that ‘‘elderly patients with dementia-related psychosis treated with antipsychotic drugs are at an increased risk of death’’ [18, 19]. Antipsychotics were added to the Beer’s criteria as a medication to avoid in dementia patients (strength of recommendation: strong) [4]. Other adverse outcomes are attributed to antipsychotics as well. A recent editorial reveals that observational studies conducted to date demonstrate a two- to threefold increased risk of pneumonia in a dose-dependent fashion with conventional or atypical antipsychotics compared with nonuse, with a possibly higher risk attributable to atypical antipsychotics [20]. Additionally, the risk of venous thromboembolism is increased with typical and atypical antipsychotics, though the actual rate varies with the type
CPOE Alerts and High-Risk Medications in Older Patients
of drug and is highest just after starting the antipsychotic [21]. Oral atypical antipsychotic medication use has been shown to increase the risk for acute kidney injury requiring hospitalization [22].
2 Objectives Our study aimed to evaluate the impact of these ‘‘geriatric alerts’’ in the CPOE on ordering patterns of diphenhydramine, metoclopramide, and antipsychotics in the hospital assessing prescription rates as measured by the number of prescriptions divided by the number of hospital discharges in the second quarter of a given calendar year over a 3-year period. We compared rates of ordering before the alert (2010) and after the alerts were implemented in January of 2011 (2011–2013). Our goal was to reduce the use of these medications in the older patients through increased awareness of their potential adverse outcomes supported by published evidence. We intentionally did not suggest alternative pharmacologic agents or non-pharmacologic interventions as each patient has specific needs and a medical history and requires treatment catered to their individual health status. Furthermore, because these highrisk drugs are inherently dangerous to use, substitution with another agent by definition, would most likely be less dangerous.
3 Methods Winthrop Hospital is a community teaching institution located in Long Island, NY, USA with residencies, fellowships, and a physician’s assistant/nurse practitioner service. All clinicians, including attending physicians are able to enter orders through CPOE. CPOE became operational by 2009. In January 2011, a series of alerts were built into CPOE for selected medications. These pop-up alerts contained links to articles relevant to the alert [4, 9, 23] (see Electronic Supplementary Material Table S1). The alerts require the user to click ‘‘noted’’ and then the user either cancels their order or continues with prescribing the medication. The alert does not require explanation for the use of the medication. See Electronic Supplementary Material Table S2 for a snapshot of a diphenhydramine alert. We evaluated the effect of these alerts on prescribing frequency for diphenhydramine, metoclopramide, and all antipsychotics by comparing the number of patients for whom these medications were prescribed during the second quarters of 2010 through 2013. Prescribing patterns were evaluated by ascertaining the pharmacy database, which contained all medication orders since the introduction of CPOE. If patients were prescribed medication of the same
229
type multiple times during their admission, only the first occurrence was counted in the total number prescribed. This was done to avoid counting medications multiple times for the same patient. Because the alerts went into effect in January of 2011, we compared 2010 data as the baseline to use in years 2011 through 2013. Prescription rates for each year were computed as the number of prescriptions adjusted by the total number of discharges during each second quarter of interest. We also compared prescribing for the same time period for patients aged 18–64 years to serve as a comparison group. In secondary analyses, older patients were further stratified by age (65–74 years, 75–84 years, 85 years and over). We hypothesized that use in the younger comparison group (who did not receive the alerts when ordering medications) would not result in a reduction in prescribed numbers. Additional analysis was performed to assess changes in prescribing patterns over time such as transition to more PRN (as needed) orders instead of standing orders. The observed prescription rates for diphenhydramine, metoclopramide, and all antipsychotics were analyzed via Poisson regression models, using the natural log of the number of discharges as the offset, and including time and age as model covariates. Specific trends in prescription rates over time were examined: whether there was a reduction in the average 3-year (post-alert) rate from 2011 to 2013 compared with pre-alert (2010) levels by age group, a reduction from 2010 when comparing only 2013 rates, and a linear reduction in rates over time (2010–2013) by age group. One-sided p values corresponding to t tests for each effect of interest were presented and values B0.025 were considered statistically significant. One-sided analysis was used as the study was designed to assess for a decrease in prescription rates. Data were analyzed using SAS 9.2 (SAS Institute, Cary, NC, USA). Institutional review board approval was not obtained for this study as it was initially designed as a quality improvement project.
4 Results Our results are notable because the alerts only reduced prescription rates for medications for which there were safer alternatives (metoclopramide and diphenhydramine for sleep) or the medication was not needed (diphenhydramine as a pre-medication for blood transfusions). Antipsychotic orders were not reduced, probably because providers lack alternative treatments for psychosis and delirium. We also found no carry-over effect to the younger population, which suggests the alert needs to be specific to patient characteristics. Among patients aged older than 65 years, there were statistically significant reductions in prescription rates following the introduction of
230
P. E. Lester et al.
Fig. 1 Prescription rates of the three drug classes in 65 and older cohort
Fig. 2 Prescription rate for diphenhydramine stratified by age
these alerts for diphenhydramine (26.9 % in 2010 vs 20 % in 2013; p \ 0.001) and metoclopramide (16.7 % in 2010 vs 12.5 % in 2013; p \ 0.001). There was no decrease observed in the rate of prescriptions for antipsychotics in the older patients (8.8 % in 2010 vs 9.2 % in 2013; p = 0.80) (Fig. 1). In the younger group (B65 years old), no decrease in prescription rates for metoclopramide, diphenhydramine, or antipsychotics were observed. See Table 1 for hospital discharges by age group and year. Analysis of the older cohort with additional age stratification was performed with groups defined as ages
65–74 years, 75–84 years, and 85 years and over. This analysis revealed a greater impact of the alert on the reduction of rate of prescriptions for diphenhydramine (Fig. 2) and metoclopramide (Fig. 3) for those with advancing age. Those aged 85 years and over had a larger absolute reduction in prescription rates for diphenhydramine (22.5 % in 2010 vs 12.5 % in 2013; p \ 0.001), compared with those aged 75–84 years (24.2 vs 18.4; p \ 0.001) and aged 65–74 years (30.5 vs 27.8; p = 0.12). Similarly, for metoclopramide, those aged 85 years and over had the largest rate reduction (11.8 % in 2010 vs 5.6 % in 2013; p \ 0.001), compared with those aged 65–74 where rates remained stable (20.6 vs 21.4; p = 0.68). Analysis of antipsychotics in 2010 (pre-alert) showed the highest prescription rate in those aged 85 years and over (14.4 %), with lower rates seen in each successively younger age group (p \ 0.001; prescription rates are: 7.4 % in 75- to 84-year-old individuals, and 4.1 % in 65- to 74-year-old individuals, and 2.6 % in \65-year-old individuals) (Fig. 4). Secondary analysis revealed no significant change in types of orders from 2010 to 2013, i.e., there was no significant decrease in standing orders or increase in PRN order rates (data not shown).
Table 1 Hospital discharges by age group and year (second quarter) Year
Age group (years)
Hospital discharges (n)
2010
\65
4243
65–74
1036
75–84
1325
2011
2012
2013
85?
898
\65
4412
65–74
1107
75–84 85?
1304 1000
\65
3909
65–74
1065
75–84
1113
85?
873
\65
4051
65–74
1089
75–84
1127
85?
913
5 Discussion We found a statistically significant decrease in the number of orders for patients aged 65 years and over for diphenhydramine and metoclopramide after implementation of CPOE ‘‘geriatric alerts’’. There was no reduction in the rate of prescription for these drugs in our comparison group of 18- to 64-year-old individuals. This demonstrates that alerts of this nature can have a meaningful impact on
CPOE Alerts and High-Risk Medications in Older Patients
Fig. 3 Prescription rate for metoclopramide stratified by age
Fig. 4 Prescription rate for antipsychotics stratified by age
prescriber patterns. Perhaps the availability of alternative agents helped reduce use. For example, there are safer options than diphenhydramine to treat insomnia and safer options then metoclopramide for nausea/vomiting in the older population. Additionally, the alert provided specific reference to lack of efficacy for the use of diphenhydramine as prophylaxis prior to blood transfusions. Our study is unique because it was not exclusively based on Beer’s criteria (as in Mattison et al. [1)] and it did not force users to obtain approval to use the medications or suggest alternative agents (as in Agostini et al. [3]). There was no reduction in antipsychotic order prescription rate for older patients as a result of the ‘‘geriatric alerts’’. For some patients, the antipsychotic may have been continued from outpatient regimens, and might have been
231
inappropriate to discontinue during an acute hospital stay. In such cases, the onus should be on the original prescribing provider to address potential adverse effects including the FDA warning with caregivers or patients. Moreover, antipsychotics are often ordered in the hospital because of acute aggressive, psychotic, or delirious symptoms. As such, prescribing patterns were likely not to be reduced as the drugs are used for an acute symptom without safer and established alternatives to use. This is consistent with nursing home data that showed no reduction in the use of antipsychotics since the FDA black-box warning [24]. Our study was not designed to evaluate whether the alert increased the discussion or documentation of the FDA black-box warning with caregivers or if additional non-pharmacologic interventions were attempted. Subgroup analysis showed a lower use for diphenhydramine and metoclopramide in the oldest group of patients aged 85 years and older, with incrementally higher rates of use in younger age groups. This suggests that the older patients are more frail and at greater peril for adverse reactions from these potentially high-risk medications. In comparing antipsychotic use in those patients aged 85 years and older from 2011 to 2013, the linear trend downward is statistically significant and not seen in the younger age groups. The higher prescription rate of antipsychotics in the 85 years of age and older group is not surprising because the highest rates of dementia and delirium are in the older population and thus they are more likely to have behavioral symptoms for which these drugs are prescribed. In August 2012, consistent with newly revised Beer’s criteria [4], we revised the antipsychotic alert (see Electronic Supplementary Material Table S1). In 2013, there was no significant drop in antipsychotic use among patients aged older than 65 years even after this alert was implemented. Our findings suggest that despite awareness of the risks of antipsychotics, these drugs are used in hospital because of perceived drug efficacy and lack of satisfactory alternatives [25]. There was no carry-over effect of reducing prescription rates for orders of diphenhydramine, metoclopramide, or the antipsychotics in the comparison group of individuals aged 18–64 years. This suggests that such alerts are seen as targeted specifically to the older patient population aged over 65 years. Of note, the alerts we used for metoclopramide and the antipsychotics advise caution about long-term use. Because our study did not evaluate the number of days these medications were prescribed for, or their use following hospitalization, we cannot assess whether long-term use of metoclopramide or antipsychotics were influenced by the alert. Our study was not designed to assess whether there
232
was better documentation about informing patients of the FDA warning on metoclopramide or antipsychotics. Better documentation and more discussion with patients about the risks associated with these drugs would be another benefit of the alerts, even if there were no reduction in hospital prescribing. There was no shift in prescribing patterns from standing to PRN as a result of the alerts. One might have expected a shift to PRN as a way for providers to try to limit use of high-risk drugs, but still have them available for patients. The lack of shift suggests that providers recognize the risk of the drugs (whether administered standing or PRN) and choose alternatives instead, when possible. Our study has limitations. The analysis is based on number of written prescriptions and not actual doses administered. Additionally, the study was not designed to include a chart review so no conclusions concerning the effect of these alerts on communication or documentation of risk/benefits of these medications can be ascertained. We did not assess clinical outcomes as to effectiveness of alternative hypnotics or anti-emetics. However, the reduction of diphenhydramine and metoclopramide orders implies that patients were managed without them. Of note, there were no specific educational programs targeting any of these medications in the older patients, which might have impacted prescribing patterns. Like similar studies, we also lack the ability to determine if adverse drug events were reduced by use of this warning system and whether the medications that were ordered were in fact, clinically required.
6 Conclusion This study demonstrates that CPOE alerts for older hospitalized patients can be used to increase awareness about the efficacy and risk of certain medications and can have a significant impact on drug prescribing rates for some medications, notably those with alternative available agents. Such alerts can be targeted to high-risk drugs in vulnerable populations and can include supportive references. It is also important to keep the resource links updated. It is imperative to carefully select the medications to receive such alerts to prevent ‘‘alert fatigue,’’ in which the users actually cease reading the alerts because they are overwhelmed by a multitude of them. Future research may show the impact on clinical outcomes and should include assessment of the impact of the alerts on the documentation of discussions with patients and caregivers about the risks and benefits of these medications. Acknowledgments The authors gratefully acknowledge the contributions of Steve Gerson RPh, MPA of Pharmacy, and Denise Strand
P. E. Lester et al. BSN RN of Clinical Decision Support for their contributions to this project. There was no funding for this project. Contributions: All authors provided substantial contributions to the conception and design of the work, and the acquisition, analysis, and interpretation of data. Dr. Lester drafted the manuscript with critical revisions provided for important intellectual content by all other authors. All authors approve the manuscript version submitted. The authors have no potential conflicts of interest that are directly relevant to the content of this study.
References 1. Mattison MLP, Afonso KA, Ngo LH, Mukamal KJ. Preventing potentially inappropriate medication use in hospitalized older patients with a computerized provider order entry warning system. Arch Intern Med. 2010;170:1331–6. 2. Lainer M, Mann E, So¨nnichsen A. Information technology interventions to improve medication safety in primary care: a systematic review. Int J Qual Health Care. 2013;25(5):590–8. 3. Agostini JV, Zhang Y, Inouye SK. Use of a computer-based reminder to improve sedative-hypnotic prescribing in older hospitalized patients. J Am Geriatr Soc. 2007;55(1):43–8. 4. American Geriatrics Society 2012 Beers Criteria Update Expert Panel. American Geriatrics Society updated Beers criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2012;60:1–15. 5. Ochs KL, Zell-Kanter M, Mycyk MB. Hot, blind, and mad: avoidable geriatric anticholinergic delirium. Am J Emerg Med. 2012;30(3):514. 6. Fortin MP, Rouch I, Dauphinot V, et al. Effects of anticholinergic drugs on verbal episodic memory function in the elderly: a retrospective, cross-sectional study. Drugs Aging. 2011;28:195–204. 7. Shimizu T, Tokuda Y. Acute urinary retention. Intern Med. 2011;19(1):2255. 8. Kennedy LD, Case LD, Hurd DD, et al. A prospective, randomized, double-blind controlled trial of acetaminophen and diphenhydramine pretransfusion medication versus placebo for the prevention of transfusion reactions. Transfusion. 2008;48:2285–91. 9. Wang SE, Lara PN Jr, Lee-Ow A, et al. Acetaminophen and diphenhydramine as premedication for platelet transfusions: a prospective randomized double-blind placebo-controlled trial. Am J Hematol. 2002;70:191–4. 10. Montgomery P, Lilly J. Insomnia in the elderly. Clin Evid. 2007;2007:2302. 11. Tu RH, Grewall P, Leung JW, et al. Diphenhydramine as an adjunct to sedation for colonoscopy: a double-blind randomized, placebo-controlled study. Gastrointest Endosc. 2006;63:87–94. 12. Katayose Y, Aritake S, Kitamura S, et al. Carryover effect on next-day sleepiness and psychomotor performance of nighttime administered antihistaminic drugs: a randomized controlled trial. Hum Psychopharmacol. 2012;27(4):428–36. 13. REGLAN Injection (metoclopramide injection, USP) Rx only FDA. [Online]. 2013. http://www.accessdata.fda.gov/drugsatfda_ docs/label/2010/017862s063lbl.pdf. Accessed 10 Dec 2013. 14. Jordan K, Gralla R, Jahn F, Molassiotis A. International antiemetic guidelines on chemotherapy induced nausea and vomiting (CINV): content and implementation in daily routine practice. Eur J Pharmacol. 2013;722:197–202. 15. Kovac AL. Update on the management of postoperative nausea and vomiting. Drugs. 2013;73:1525–47. 16. Scorza K, Williams A, Phillips JD, Shaw J. Evaluation of nausea and vomiting. Am Fam Physician. 2007;76:76–84.
CPOE Alerts and High-Risk Medications in Older Patients 17. Alexopoulos GS, Streim J, Carpenter D, Docherty JP. Using antipsychotic drugs in older patients. J Clin Psychiatry. 2004;65 (Suppl 2):5–99. 18. US FDA: Postmarket drug safety information for patients and providers–public health advisory: deaths with antipsychotics in elderly patients with behavioral disturbances. [Online]. 2010. http:// www.fda.gov/drugs/drugsafety/postmarketdrugsafetyinformation forpatientsandproviders/ucm053171. Accessed 17 Feb 2015. 19. Schneider LS, Dagerman KS, Insel P. Risk of death with atypical antipsychotic drug treatment for dementia: meta-analysis of randomized placebo-controlled trials. JAMA. 2005;294:1934–43. 20. Gambassi G, Sultana J, Trifiro` G. Antipsychotic use in elderly patients and the risk of pneumonia. Expert Opin Drug Saf. 2015;14(1):1–6. 21. Ishiguro C, Wang X, Li L, Jick S. Antipsychotic drugs and risk of idiopathic venous thromboembolism: a nested case-control study
233
22.
23.
24.
25.
using the CPRD. Pharmacoepidemiol Drug Saf. 2014;23(11): 1168–75. Hwang YJ, Dixon SN, Reiss JP, et al. Atypical antipsychotic drugs and the risk for acute kidney injury and other adverse outcomes in older adults: a population-based cohort study. Ann Intern Med. 2014;161(4):242–8. Fick DM, Cooper JW, Wade WE, et al. Updating the Beers criteria for potentially inappropriate medication use in older adults. Arch Intern Med. 2003;163:2716–24. Lester P, Kohen I, Stefanacci RG, Feuerman M. Antipsychotic drug use since the FDA black box warning: survey of nursing home policies. JAMA. 2011;18:573–7. Zirker W, Dorokhine I, Knapp CM, et al. Haloperidol overdosing in the treatment of agitated hospitalized older people with delirium: a retrospective chart review from a community teaching hospital. Drugs Aging. 2013;3:639–44.
