Review of the Trigger Literature: Adverse Events Targeted and Gaps in Detection
Triggers and Targeted Injury Detection Systems (TIDS)
An important development in adverse event (AE) detection is the use of triggers, algorithms that use patient data to look for patterns consistent with a possible AE.A1,A4 In a trigger system, when a trigger flags a record, there is a method to further determine whether an AE occurred. In the case of action-oriented trigger systems, triggers are designed to support clinical interventions that prevent or mitigate iatrogenic harm. Trigger systems have been used in inpatient settings for rate detection and to signal providers to investigate a possible AE in real time.A5,A7 Recently, trigger systems have been used to detect AEs that occur in particular settings, such as emergency departmentsA8 or neonatal intensive care units,A7 or among specific patient groups, such as pediatric populations.A7,A9,A10 This paper reviews the literature on triggers developed as part of an outpatient trigger development project funded by the Agency for Healthcare Research and Quality (AHRQ).A11
This review summarizes the trigger literature published prior to January 1, 2008. In addition to literature from the project team, we conducted searches of information databases using standardized keywords. Forty-five references contained information on triggers or trigger systems. We also reviewed articles for background information on the leading causes and types of AEs.
Summary of Literature on Accounting Trigger Systems
Some triggers are designed to be used together as a trigger system, typically for the purpose of AE rate estimation or accounting.A3,A6,A12,A17 Most accounting trigger systems were developed by the Institute for Healthcare Improvement (IHI) and include information on implementation as well as guidelines on classifying the harm and/or preventability of AEs detected.A16 The objective of accounting trigger systems is not to test and improve the positive predictive value (PPV) of any individual trigger, but to estimate rates of AEs within the system.
Summary of Triggers Linked to Specific AEs
For this paper, our primary focus is on triggers and trigger systems that were linked to specific AEs or specific AE causes. (Therefore, we do not include triggers that were part of accounting trigger systems in this section.) The majority of the triggers linked to AEs were drug related (n = 364). Figure 1 shows the most frequent adverse drug events (ADEs) targeted by triggers or trigger systems in the literature. (Only ADEs with ≥5 triggers are shown.) In addition to the 23 ADEs shown in Figure 1, there were 88 other ADEs targeted by specific triggers. Triggers varied in the amount of detail or in the type of data used to detect a specific AE. For example, one of the triggers that targeted bleeding was "Vitamin K given," while another trigger that also targeted bleeding included information on the type of bleeding by specifying "International Normalized Ratio (INR) elevated or increasing."A18
Figure 2 presents the frequency of triggers designed to detect specific AEs that occurred because of medical mismanagement and progression of underlying disease. (Only AEs with ≥2 triggers are shown.) In addition to the 18 " medical management failure" events shown, there were 34 other AEs targeted by one trigger. AEs classified as medical mismanagement tend to be rare but harmful, and trigger development in this area is focused primarily on expanding the number of AEs detected, rather than refining the detection process.
Figure 3 shows the distribution of surgical AEs targeted by triggers in the literature. AEs resulting from inpatient and same-day surgeries are prevalent and costly;A19,A21 however, we found only 31 surgical triggers. Several of these triggers were not part of trigger systems and therefore did not have any mechanism for confirming that an AE occurred.
We found 27 triggers that could not be easily categorized. These triggers concerned global AEs (e.g., a natural language processing discharge summary review that used trigger words like "error"A22); crimes (e.g., infant abductionA23); or death/serious injury with an unspecified cause.A23
Summary of Literature on Triggers Linked to AE Causes
We also reviewed triggers and trigger systems linked to the cause of an event. In some cases, particularly with respect to medical mismanagement and surgical triggers, the event was specified and is therefore included in the previous section. There were 314 drug-related triggers that specified the drug that caused the ADE; types of causal drugs are shown in Figure 4. (Only causal agents with ≥2 triggers are shown.) One hundred drug-related triggers specify the targeted ADE but do not include the drug or drug class that may have been the causal agent.
Gaps and Future Directions for Trigger Development
Our review of the literature found that the majority of triggers and trigger systems were drug related. Based on the ADE prevalence literature, the most frequent drug-related triggers detect the most common ADEs in the population.A24 However, several drugs that cause high rates of ADEs in the outpatient setting are not in the trigger literature: contraceptives, and drugs used for skin, eye, and dental problems.A24 Future drug-related trigger system development should consider ADE detection in ambulatory settings, including primary and specialty care.
We found a wide variety of triggers designed to detect specific medical mismanagement AEs. Most of these triggers were designed as accounting triggers; however, there is also an opportunity to use the trigger language to develop action-oriented trigger systems. Only two articles specified a cause of medical mismanagement AEs.A25,A26 Diagnostic errors and failure to follow up are common causes of AEs, and more work needs to be done in developing action-oriented trigger systems that detect these types of events.
With the exception of the IHI,A14 surgical trigger systems have not yet been developed. While we found two articles with triggers that could be used in an inpatient action-oriented trigger system,A22,A23 there were no surgical triggers designed for outpatient surgery. Given the severity and nature of surgical AEs, future research should target the development of action-oriented surgical trigger systems for inpatient and outpatient care.
A7. Sharek PJ, Horbar JD, Mason W, et al. Adverse events in the neonatal intensive care unit: development, testing, and findings of an NICU-focused trigger tool to identify harm in North American NICUs. Pediatrics 2006 Oct;118(4):1332-1340.
A9. Ferranti J, Horvath MM, Cozart H, et al. Reevaluating the safety profile of pediatrics: a comparison of computerized adverse drug event surveillance and voluntary reporting in the pediatric environment. Pediatrics 2008 May;121(5):e1201-1207.
A10. Takata GS, Mason W, Taketomo C, et al. Development, testing, and findings of a pediatric-focused trigger tool to identify medication-related harm in US children's hospitals. Pediatrics 2008 Apr;121(4):e927-935.
A11. Rosen AK, Nebeker JR, Shimada S, et al. Development and Use of Ambulatory Adverse Event Trigger Tools. Rockville (MD): Agency for Healthcare Research and Quality; 2007. AHRQ Contract No. HHSA2902006000012, TO #3.
A18. Handler SM, Altman RL, Perera S, et al. A systematic review of the performance characteristics of clinical event monitor signals used to detect adverse drug events in the hospital setting. J Am Med Inform Assoc 2007 Jul-Aug;14(4):451-458.
A20. Johnson RG, Arozullah AM, Neumayer L, et al. Multivariable predictors of postoperative respiratory failure after general and vascular surgery: results from the patient safety in surgery study. J Am Coll Surg 2007 Jun;204(6):1188-1198.
A21. Neumayer L, Hosokawa P, Itani K, et al. Multivariable predictors of postoperative surgical site infection after general and vascular surgery: results from the patient safety in surgery study. J Am Coll Surg 2007 Jun;204(6):1178-1187.
A24. Thomsen LA, Winterstein AG, Sondergaard B, et al. Systematic review of the incidence and characteristics of preventable adverse drug events in ambulatory care. Ann Pharmacother 2007 Sep;41(9):1411-1426.
a VA Center for Health Quality Outcomes and Economic Research (CHQOER) and Boston University School of Public Health, Health Policy and Management Department.
b VA Salt Lake City Geriatrics Research, Education, and Clinical Center (GRECC); Department of Medicine, University of Utah; and Intermountain Institute for Healthcare Delivery.
Note: The views in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs.