Background and definition
Health care organizations use a variety of strategies to detect safety hazards in order to prevent harm. These methods are often referred to by various names such as Targeted Injury Detection Systems and most commonly as triggers. Triggers have become a widely used way to retrospectively analyze medical records in order to identify errors and adverse events, measure the frequency with which such events occur, and track the progress of safety initiatives over time. Triggers alert patient safety personnel to possible adverse events so they can review the medical record to determine if an actual or potential adverse event has occurred. The main value of triggers is efficiency, since a complete review of every medical record to find adverse events is laborious and expensive, even in the era of electronic medical records. Triggers provide a way of screening medical records for possible harm and identifying cases that merit a more detailed review.
For example, the administration of naloxone (a drug used to reverse the effect of opioid medications) to a hospitalized patient would be a reasonable trigger that could help identify instances where a patient was given a harmful dose of an opioid drug. When naloxone is administered in an inpatient ward, it may be because the patient received an excessive dose of morphine or another opioid medication. Therefore, pharmacists and patient safety personnel could use that trigger to identify cases that may represent problems with the ordering or administration of opioid medications. This method would miss many less severe cases (ones that weren't severe enough to merit naloxone administration), but the cases it did identify would very likely represent preventable adverse events. Well-defined, specific triggers like this also lend themselves to automated electronic detection, making them particularly efficient for ongoing monitoring activities. When the trigger correctly identifies an adverse event, causative factors can be determined and interventions developed to reduce the frequency of such events. Triggers can also be used to track rates of adverse events over time.
The Harvard Medical Practice Study and other classic studies used fairly blunt triggers, general indications that harm may have occurred such as death, readmission, or unexpected return to the operating room. There was no expectation that most cases would turn out to be adverse events. Refinement of this methodology has led to the development of more specific triggers.
The Institute for Healthcare Improvement Global Trigger Tool (GTT) has become one of the most widely used trigger tools for detecting harm in hospitalized patients. It combines blunt triggers (such as rapid response team activation) with more specific but relatively insensitive ones (such as an abnormally low blood glucose measurement). The GTT includes 53 different triggers, some applicable to all patients and some inappropriate for certain patient populations or settings of care. The GTT is practical for routine improvement efforts and not just research studies, and it includes detailed instructions for training reviewers and interpreting results. Conceptually, though, the method resembles that of major adverse event studies such as the Harvard Medical Practice Study. Use of the trigger tool involves screening a defined sample of medical records by two independent clinicians for presence of one or more triggers. After a trigger is identified, the entire chart is reviewed to determine whether an adverse event took place, and if so, to grade the level of harm experienced by the patient.
Various studies have assessed the reliability of judgments using the GTT and also modified it for application in different clinical settings, including pediatric patients and patients with cancer. Versions of the tool appropriate for prospective detection in real-time have also been developed.
The IHI cautions that the GTT (or any trigger tool method) cannot identify all sources of patient harm or the cause of harm, a point emphasized in an influential commentary. Also, trigger tools are designed to detect all adverse events; reviewers are explicitly instructed to avoid making judgments about preventability of these events during the initial review process. Nonetheless, many studies have used the GTT or other similar tools to estimate the frequency of preventable adverse events in a variety of clinical settings. This is not inappropriate per se, but readers of such studies should be aware that inter-rater agreement around preventability is generally only moderate, as discussed in more detail in the Measurement of Patient Safety primer.
Concerns have also been raised regarding the reliability of trigger tools, both for detection of adverse events and for rating the severity of harm experienced by patients. One Swedish study used 5 teams of reviewers (each of whom had at least 3 years of GTT experience) to review a random sample of hospitalizations and found that agreement between teams on the presence of an adverse event was only slightly better than chance. Another influential study of temporal trends in adverse events also found markedly different rates when the GTT was used by personnel internal or external to the hospitals being studied. It is likely that reliability of trigger tools is significantly influenced by the level of training and experience of the reviewers and their familiarity with the clinical setting being evaluated.
Finally, most existing trigger tools have been used to identify adverse events in the inpatient setting. Although some studies have sought to develop trigger tools for ambulatory care, there is relatively little data on the accuracy and reliability of these tools.