Decision Support for Quality Improvement Unit 6.3: Alerts and Clinical Reminders Welcome back to Decision Support for Quality Improvement. This unit is designed to provide you with a deeper understanding of the role of alerts and clinical reminders, their benefits, and potential hazards. Component 12/ Unit6.3 Health IT Workforce Curriculum

Health IT Workforce Curriculum Objectives Analyze the benefits and shortfalls of alerts and clinical reminders After completing unit 6.3, the learner should be able to analyze the benefits and shortfalls of alerts and clinical reminders. Component 12/ Unit6.3 Health IT Workforce Curriculum

Health IT Workforce Curriculum Reminders and Alerts “…the burden of reminders and alerts must not be too high…or alert fatigue may cause clinicians to override both important and unimportant alerts, in a manner that compromises the desired safety effect of integrating decision support into CPOE.” While decision support in the form of alerts and reminders has the potential to improve patient safety, reality is that, according to the current literature, these alerts are overridden by clinicians 49 - 96% of the time. This is a real concern for HIT professionals. 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(2), 138-147. Component 12/ Unit6.3 Health IT Workforce Curriculum

Health IT Workforce Curriculum Alerts and Reminders Nuisance Alert Alert Fatigue “…provides little perceived benefit to the prescriber at the time of the alert” “…arise when clinicians, either consciously or unconsciously, begin to systematically bypass CDS alerts without regard to their importance, enabling the possibility that a clinically important alert is missed” One of the main reasons given for over-riding alerts is perceived lack of relevance. The so-called Nuisance alerts provide little perceived benefit to the prescriber at the time of the alert. This causes great frustration on the part of clinicians and leads to a phenomenon called alert fatigue. Alert fatigue can result in the clinician systematically bypassing the alert, either consciously or subconsciously. The potential for missing a clinically important alert is high. Chaffee, BW (2010). Future of clinical decision support in computerized prescriber order entry. American Journal of Health System Pharmacists, 67: 932-935. Component 12/ Unit6.3 Health IT Workforce Curriculum

Responses to Clinical Reminders Tendency to perform an action when a warning system instructs the user to do so Compliance Tendency to refrain from performing an action when the warning system does not indicate that it is necessary Reliance There have been 4 types of responses identified with respect to clinicians and clinical reminders. Compliance is the tendency to perform an action when a warning system instructs the user to perform a corrective or preventive action, such as checking a hospitalized patient’s status when a clinical monitor issues an alert in the EHR. Reliance, on the other hand, is the tendency to refrain from performing an action when the warning system does not indicate that it is necessary, e.g., assuming that the patient’s status is normal and thus not checking the patient, when the monitor does not generate an alert in the EHR. Vashitz G, Meyer J, Parmet Y, Peleq R, Goldfarb D, Porath A, Gilutz H. (2009). Defining and measuring physicians' responses to clinical reminders. Journal of Biomedical Informatics; 42(2):317-26. Epub 2008 Oct 26. Component 12/ Unit6.3 Health IT Workforce Curriculum

Responses to Clinical Reminders Clinician performs an action even when not prompted by the reminder system Spillover Clinician refrains from performing an action due to a perceived threat to professional autonomy Reactance Two additional types of responses include Spillover and Reactance. Spillover can occur when there is a spread of responses merely due to increased awareness of the need for an action, even when the clinician is not prompted by the reminder system. For example, Zanetti’s research team found that reminders for intra-operative drugs increased re–dosing even in a control group which did not receive these reminders and concluded that this spillover occurred because of increased awareness to the importance of re-dosing. Reactance is the tendency for clinicians to experience a threat to their autonomy or freedom of choice in the presence of these systems. The clinician may consciously or unconsciously react by either ignoring the reminders or choosing a different course of action, trying to regain their sense of freedom and autonomy. Vashitz G, Meyer J, Parmet Y, Peleq R, Goldfarb D, Porath A, Gilutz H. (2009). Defining and measuring physicians' responses to clinical reminders. Journal of Biomedical Informatics; 42(2):317-26. Epub 2008 Oct 26. G. Zanetti, H.L. Flanagan, L.H. Cohn, R. Giardina and R. Platt, Improvement of intraoperative antibiotic prophylaxis in prolonged cardiac surgery by automated alerts in the operating room, Infect Control Hosp Epidemiol 24 (1) (2003), pp. 13–16. Component 12/ Unit6.3 Health IT Workforce Curriculum

Four Types of Alerts/Reminders Drug Alerts Lab Test Alerts Practice Reminders Administrative Reminders For the purpose of this discussion, we will look at experience with 4 types of alerts or reminders: drug alerts, lab test alerts, practice reminders, and administrative reminders. Component 12/ Unit6.3 Health IT Workforce Curriculum

Health IT Workforce Curriculum Basic Drug Alerts Drug allergy warnings Drug-drug interactions Duplicate medication or therapeutic duplication alert Basic medication order guidance Kuperman and colleagues identify 4 types of basic drug alerts. Drug allergy warnings are generated on ordering a drug to which the patient has a documented drug allergy. Drug-drug interactions are generated when the mode of action of one drug is known to be affected by simultaneously prescribing a second drug. Duplicate medication or therapeutic duplication alerts are generated when the patient is already receiving the medication just ordered or a different drug in the same therapeutic category. Basic medication order guidance is an alert that provides dosing information with default dosing being the most appropriate initial dosing. Component 12/ Unit6.3 Health IT Workforce Curriculum

Health IT Workforce Curriculum Advanced Drug Alerts Drug-Lab alerts Drug-Condition interactions Drug-Disease Contraindication alerts Drug-condition alerts aimed at appropriate prescribing There are 8 identified types of advanced drug alerts. Drug-lab alerts are generated when drug administration requires close monitoring of laboratory results before and/or after administration. Drug-condition interactions are generated to raise awareness of specific prescribing for particular clinical conditions. Drug-disease contraindication alerts are generated to warn against prescribing a certain drug in a specific disease or condition. Drug-condition alerts aimed at appropriate prescribing are generated to encourage prescribing a certain drug in a specific clinical condition. Component 12/ Unit6.3 Health IT Workforce Curriculum

Health IT Workforce Curriculum Advanced Drug Alerts Drug-age alerts Drug-formulary alerts Dosing guidelines Complex prescribing alerts Drug-age alerts are generated to discourage prescribing of a certain drug in the elderly or in children. Drug-formulary alerts are generated to notify the prescriber that a particular brand or drug is neither included nor recommended in the formulary at the prescribing location. Dosing guidelines are alerts that take into account complex patient characteristics such as age, renal/liver function, pregnancy or female of childbearing potential, pediatric weight-based dosing, drug utilization restrictions, or clinical indication. Finally, complex prescribing alerts are generated with combined features of basic and advanced alerts. Component 12/ Unit6.3 Health IT Workforce Curriculum

Evidence to Support Drug Alerts Systematic review examined 20 studies that evaluated the impact of efficacy of computerized drug alerts and prompts 23 of 27 alert types identified demonstrated benefit Improving prescribing behavior Reducing error rates Greatest potential for affecting prescribing Drug-drug interaction alerts Drug-disease contraindication alerts Dosing guidelines based on age Angela Schledlbauer and her colleagues performed a systematic review of the evidence that supports use of computerized alerts and prompts to improve drug prescribing behaviors. Most of the studies reviewed showed positive benefits. 23 of 27 drug alert types that were identified demonstrated benefits in terms of improving prescriber behavior and reducing error rates. They found that the greatest potential for benefit was with drug-drug interaction and drug disease contraindication alerts, and age-related dosing guidelines. Schedlbauer, A, Prasad V, Mulvaney C, Phansalkar S, Stanton W, Bates DW, Avery AJ (2009). What evidence supports the use of computerized alerts and prompts to improve clinicians' prescribing behavior? Journal of the American Medical Informatics Association;16(4):531-8. Epub 2009 Apr 23. Component 12/ Unit6.3 Health IT Workforce Curriculum

Improving Adoption of Drug Alerts Shah and colleagues studied way to improve clinician acceptance of drug alerts in ambulatory care settings Designed a selective set of drug alerts for the ambulatory care setting using a criticality leveling system Minimized workflow disruptions by designating only critical to high-severity alerts to be interruptive to clinician workflow Alert levels: 1: clinician could not proceed with the prescription without eliminating the contraindication; 2: clinicians could proceed if provided an over-ride reason 3: alert displayed at top of screen in red; did not hinder workflow A group of northeastern researchers wanted to find a way to improve clinician acceptance of drug alerts in ambulatory care settings. They designed a selective set of drug alerts using a criticality leveling systems. Level 1 alerts were the most critical and a clinician could not proceed with the prescription without eliminating the contraindication. Level 2 alerts were very important, but the clinician could proceed if he or she provided a reason for overriding the alert. Level 3 alerts were designed to not disrupt workflow, but to provide important information for the clinician to take into consideration. Through this research, Shah’s team was able to demonstrate that by designating only critical to high-severity alerts to be interruptive of clinician workflow they were able to minimize workflow disruptions and improve clinician acceptance of drug alerts. Shah NR, Seger AC, Seger DL, Fiskio JM, Kuperman, GJ, Blumenfeld B, Recklet EG, Bates DW, & Gandhi TK (2006). Improving Acceptance of Computerized Prescribing Alerts in Ambulatory Care. Journal of the American Medical Informatics Association; 13(1): 5–11. Component 12/ Unit6.3 Health IT Workforce Curriculum

Basic Laboratory Alerts Drug-laboratory alerts Duplicate laboratory testing alert Basic laboratory test order guidance Public health situational awareness Basic laboratory alerts include: drug-lab alerts, which as stated previously, are generated when drug administration requires close monitoring of laboratory results before and/or after administration. Duplicate laboratory testing alerts are generated when the patient has already had the lab test ordered. Basic laboratory test order guidance is an alert that provides ordering information with respect to the particular lab test being ordered. Finally, alerts can be generated to notify clinicians of public health alerts so that they can initiate appropriate laboratory testing during local disease outbreaks. Component 12/ Unit6.3 Health IT Workforce Curriculum

Evidence to Support Lab Alerts Research examined the impact of a clinical decision support system that generated reminders of previous lab test results Found that the proportion of unnecessarily repeated tests dropped significantly Features of the Alert Alert was automatically prompted and was part of the clinician workflow User could not deactivate the alert output Most recent laboratory result for viral serology test an its date was automatically retrieved from the patient’s EHR Alert was displayed at the time and location of decision making (before the user ordered an unnecessarily repeated test French researchers recently examined the impact of a serology-clinical decision support system providing point of care reminders of previous existing blood test results. These reminders were embedded in a Computerized Physician Order Entry at a university teaching hospital. The clinical decision support system was implemented in the Cardiovascular Surgery department of the hospital in order to decrease inappropriate repetitions of viral serology tests (specifically, Hepatitis B Virus). The researchers found that the proportion of unnecessarily repeated tests immediately dropped after implementation of the alert and remained stable. Features of the alert that enhanced its success were: the alert was automatically prompted and was part of clinician workflow, the user could not deactivate the alert output, the most recent laboratory result for viral serology tests and its date was automatically retrieved from the patient's EHR, and the alert was displayed at the time and location of decision making (ie, before the user ordered an unnecessarily repeated test). BMC Health Services Research, 9;10:70. Nies J, Colombet I, Zapletal E, Gillaizeau F, Chevalier P, Durieux P. (2010) Effects of automated alerts on unnecessarily repeated serology tests in a cardiovascular surgery department: a time series analysis. Component 12/ Unit6.3 Health IT Workforce Curriculum

Health IT Workforce Curriculum Practice Reminders Provides recommended treatment Guiding Checks prescriptions against clinical practice guidelines Critiquing Helps provider follow the patient Monitoring There are three ways in which practice reminders can be used as clinical decision support. They can guide the clinician to provide the recommended treatment. They can critique the clinicians course of action by comparing it against guideline recommendations, and they can help with patient follow-up. Lami JB, Ebrahiminia V, Riou C, Seroussi B, Bouaud J, Simon C, Dubois S, Butti A, Simon G, Favre M, Falcoff H, Venot A. et al. (2010). How to translate therapeutic recommendations in clinical practice guidelines into rules for critiquing physician prescriptions? Methods and application to five guidelines. BMC Medical Informatics and Decision Making. 2010 May 28;10:31. Component 12/ Unit6.3 Health IT Workforce Curriculum

Practice Reminders Challenges Incorrect Guidelines Too generic guideline Patient data inconsistency Inappropriate action Potential Risk There are a number of known challenges to incorporating clinical practice guidelines into clinical decision support systems. These were identified over 10 years ago, and are still applicable today. First, reminders can be false alarms, given as a result of an incorrectly implemented guideline. Next, reminders can be false due to the fact that the guideline is not specific enough to address the particular patient’s condition. For example, a guideline may not specify exceptions and a patient who falls within the exception may have the guideline inappropriately applied. Third, patient data in the EHR can be incomplete or inconsistent, resulting in the reminder either firing inappropriately or not firing at all. Fourth, the actions that the reminder generates may not be clinically appropriate for the particular patient, and finally, the reminder itself may generate risk to the patient. De Clercq PA, Blom JA, Hasman A, Korsten HHM. A strategy for developing practice guidelines for the ICU using automated knowl- edge acquisition techniques. Journal of Clinical Monitoring; 15: 109-117 Component 12/ Unit6.3 Health IT Workforce Curriculum

Administrative Reminders Guides prescribers to document to support appropriate medical coding Medical Coding Guides the collection of QI indicator data Quality Improvement Decision support can also be used to provide administrative reminders such as use of suggested coded problems for particular patient populations to support billing and specific guidance on documentation to support collection of QI indicator data. Component 12/ Unit6.3 Health IT Workforce Curriculum

Success Factors: Alerts Specificity Alert clinically important for the patient Sensitivity Alert generated in all dangerous cases There are several success factors that have been identified to aid in gaining clinician acceptance of alerts and reminders. Alerts should be specific. That is they should be clinically important for the patient. Action should follow the alert and the alert should be accurate for the right patient (taking patient characteristics such as age, clinical condition, and gender into account). Alerts should also be sensitive. They should fire in all dangerous cases. 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(2), 138-147. Component 12/ Unit6.3 Health IT Workforce Curriculum

Success Factors: Alerts Information Content Clear, concise, unambiguous Justification noted Further information accessible Alternative actions presented Information contained in the alert should be clear and concise. Justification should be noted and additional information should be readily accessible. Alternative actions should be presented 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(2), 138-147. Component 12/ Unit6.3 Health IT Workforce Curriculum

Success Factors: Alerts Workflow Directed to right person at right time Specialty-specific; Knowledge-specific Avoid repetition Safe, efficient handling High threshold Reasons for non-compliance Promotes action Speed; Screen design; minimize work Alerts should be directed to the right person at the right point in their workflow. Specialists should not receive alerts that tell them what they already know. Physicians in training should receive more alerts than specialists. Annoying repetition should be prevented. In addition, overriding fatal alerts should never be easy. Reasons for over-riding the alert should be requested. Alerts should promote action rather than stopping intended actions. Alerts should not slow the system down and the user should not be required to scroll endlessly or perform multiple keystrokes to accomplish a task. Screen design should be logical. 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(2), 138-147. Component 12/ Unit6.3 Health IT Workforce Curriculum

Health IT Workforce Curriculum Summary Alerts and reminders have the potential to improve patient safety Types of alerts and reminders include: drug and lab test alerts, practice reminders, and administrative reminders Nuisance alerts provide little perceived benefit to the prescriber at the time of the alert, causing clinician frustration and alert fatigue Successful alerts are specific, sensitive, clear and concise. Successful alerts support clinical workflow and allow for safe, efficient responses. In summary, alerts and reminders have the potential to improve patient safety. Types include drug and lab test alerts, practice reminders, and administrative reminders. Alerts and reminders also have the potential to compromise patient safety. Nuisance alerts provide little perceived benefit to the prescriber at the time of the alert and can cause frustration ad alert fatigue, which can , in turn, result in medical error. Successful alerts are specific, sensitive, clear and concise, support clinical workflow and efficiency. Component 12/ Unit6.3 Health IT Workforce Curriculum