Global trigger tool in child psychiatry: Treatment safety evaluation in adolescents with an acute psychotic episode

1 Introduction

Safety in paediatrics has always been a priority due to the characteristics of the child body. On the one hand, children are characterized by high adaptive potential, on the other hand by increased reactivity to various stimuli. Child psychiatry is therefore an area of special attention, as psychopharmacotherapy is used for months and even years [1]. Complications of psychopharmacotherapy are observed in more than 80% of patients [2, 3], so monitoring of adverse drug effects (ADEs) is especially relevant in a child psychiatric hospital. According to Ayani et al. (2016), 1.4% of antipsychotic-induced ADE were life-threatening and 28% were serious. The right choice of therapy ensures good compliance and, as a consequence, successful remission maintenance [5]. The frequency of re-hospitalization in patients with mental disorders due to ineffective or interrupted treatment due to poor tolerance is quite high, ranging from 30 to 74% [6]. Predicting patient’s ADE when selecting pharmacotherapy is an important task for psychiatric practitioners. But this is not always possible in real practice. In this regard, continuous monitoring of the ADE and elimination of risk factors is a priority task for any hospital. This process not only allows us to identify the most frequent predictors of ADEs but also allows us to optimize the process of treatment and make it safer.

Up to date, the most effective method of retrospective ADEs monitoring has been the Global Trigger Tool [7]. The Global Trigger Tool consists of the following algorithm: the “triggers” are searched for in completed case histories. The “triggers” are special signals about the potential ADE in the patient: prescription of a drug, prescription of a diagnostic study, increase in the time of surgical intervention, and others. For example, prescribing flumazenil, an antidote for poisoning with benzodiazepine tranquilizers, may indicate ADE caused by benzodiazepines. X-ray appointment indicates the risk that a fall in the hospital may have injured the patient. Trigger lists have been developed for many medical specialties, and triggers are searched for in this list [8]. According to the Global Trigger Tool algorithm, the analysis of a single case should not exceed 20 minutes. Hospital discharge summary, prescription lists, results of analyses, protocols of surgical interventions are subject for analysis. When a trigger is detected, the search for the ADE corresponding to it is performed: regular records and diagnostic examination results are examined in more detail. Often enough, there is no indication of ADE in the presence of a trigger. ADE search, with the help of the Global Trigger Tool, allows to reveal them 10 times more often in comparison to the method of spontaneous messages [7]. It should be noted that the Global Trigger Tool does not offer fundamentally new ADEs; it is only a more effective method of detecting already known undesirable phenomena. Currently, the Global Trigger Tool is used for regular monitoring of patient safety in clinics around the world. In addition, this tool is actively used in scientific research, and it is eligible for multisite studies [7, 8]. However, it should be noted that not all areas have trigger sheets, which requires modification and improvement of the basic algorithm.

In addition to the basic version of the Global Trigger Tool, its modifications for different specialties, including psychiatry, have been developed [8, 9]. Development of new triggers for ADE monitoring in adult psychiatry by Sajith et al. (2019) confirms the relevance of the Global Trigger Tool algorithm to mental health [9]. However, paediatric patients remained out of scope for a long time, and the Global Trigger Tool algorithm has not been designed to be used for children [8]. But over the last 5 years, there has been a significant increase in interest in using the Global Trigger Tool in paediatrics.

The Paediatric All-Cause Harm Measurement Tool (PACHMT) and Global Assessment of Paediatric Patient Safety Tool (GAPPS) were developed on the basis of the Global Trigger Tool and were recommended for use in paediatrics [10–12]. The first studies were done by Sharek PJ et al. (2006) [13] and Takata GS et al. (2008) [14], who made the first attempts to identify triggers significant for paediatrics. Matlow A.G. et al. (2012) [15] and Kirkendall et al. (2012) contributed to further development of this approach [16]. Unwanted events in paediatrics have been found to be very common, accounting for 36.7% of hospital admissions [17]. These and other studies confirmed that the global trigger method is an effective tool for detecting ADEs and developing prevention measures for them in paediatrics [17–22]. But to date, no research has been published on the use of global trigger methods for paediatric patients with mental disorders.

2 Material and methods

3 Results

We identified a total of 261 triggers among 151 analysed cases, 51 of which were accompanied by the presence of a confirmed ADE (overall positive predictive value = 19.54% for our trigger list). The value of ADE per 1000 bed days was 4.73, ADE per 100 admissions - 33.77%. We did not identify many of the triggers from the list in the analysed case histories. Several triggers were never accompanied by an ADE: “Prescribing of antihistamine”, “Abrupt reduction of the dose of medication”, “Drug combinations not normally recommended”. At the same time, certain triggers demonstrated 100% power: “Prescribing of an antiemetic”, as well as “Over-sedation, Lethargy, Falls”, which were always a consequence of taking antipsychotics. Trigger “Abrupt medication stop” had positive predictive value of 54.17%.

Other positive triggers were accompanied by confirmed ADE less often: “Ongoing or intermittent laxative use” (positive prediction value = 9.09%), “Hospital readmission within 30 days” (positive prediction value = 28.57%), “Prescribing of extrapyramidal symptoms corrector” (positive prediction value = 25%).

The frequency analysis of the triggers and the corresponding ADEs is shown in Table 2.

In this study, we used the PACHMT and GAPPS algorithms to detect ADEs in adolescents with an acute psychotic episode treated in a psychiatric hospital for the first time. The overall positive predictive value of the trigger list in this study was 19.54%, which is comparable to previous studies in other samples. In particular, Ji HH et al. (2018) reported overall positive predictive value of the trigger list of 13.3% [17], while in other studies it ranged from 3.8 to 38% [16, 18, 19]. The ADE per 1000 patient-bed days established in our sample (4.73 per 1000 patient-bed days) was lower than in the previous studies: 16.9 per 1000 patient-bed days [20], 17.4 per 1000 patient-bed days [17], 15.7 per 1000 patient-bed days [14], 42 per 1000 patient-bed days [4]. This was most likely due to the length of time a person was hospitalized in psychiatry: the period of hospitalization often exceeded a month or more. However, the frequency of detected ADEs was 33.77% per 100 patients, which was higher than in similar studies: 13.7% by Ji HH et al. (2018) [17]; interestingly, our results are close to those of Kirkendall et al. [16] - 36.7 ADE per 100 admissions [16], but the value of ADE per 1000 patient-bed days reported by Kirkendall et al. [16] was much higher than ours - 76.3 [16]. It should be taken into account that those studies considered not only the ADEs but also unwanted events not related to pharmacotherapy. In our study, we focused only on drug-induced events, and on a fairly specific population: patients with an acute psychotic episode. This may be the main reason for the differences in the frequency of the detected ADEs. This was our first experience to use PACHMT and GAPPS among this category of patients. Besides, it is necessary to take into account the length of hospital stay of patients with mental disorders, often exceeding one or even two months, which also might have affected the results of calculation of pharmacoepidemiological indicators.

The most frequently detected trigger in our study was the “Prescribing of extrapyramidal symptoms corrector”. The positive predictive value showed an unexpectedly low value: only in a quarter of the cases when the extrapyramidal corrector (trihexyphenidyl) was prescribed, the extrapyramidal symptoms indeed occurred. This means that in 75% of cases trihexyphenidyl was administered for prophylaxis, simply because a patient was taking a first generation antipsychotic. Poor registration of adverse drug events in patient medical charts could also contribute to this low PPV of the trigger. This ADE (extrapyramidal symptoms) was more common in older children, with extended hospital stay, as well in cases of use of inappropriate drug combinations according to the MAI score. The risk of extrapyramidal symptoms development was higher in patients with a history of fewer admissions. This was probably due to the fact that administering an antipsychotic to a patient for the first time in life bears higher chances of errors in determining the optimal dosing regimen. Readmissions decreased the risk for this ADE due to development of tolerance. Age was a risk factor for antipsychotic-induced extrapyramidal symptoms, most likely because psychiatrists prescribe lower initial antipsychotic doses to younger patients, whereas for older adolescents they starts with doses comparable to adult ones.

Another trigger with a relatively high number of confirmed ADE was “Abrupt medication stop”. This trigger pointed to the different nature of the ADE that led to the withdrawal of one or more drugs. It is worth noting that the ADE was not always specified in the medical records, and in half of the cases (n = 6) the justification for drug withdrawal was “due to poor tolerability, the drug was stopped”. We noted that we identified a trigger without confirmed ADE in cases, in which patients were admitted to the hospital for significantly longer periods of hospital stay, had more irrational pharmacotherapy (by MAI score). This may indicate that the fact of an ADE was often not reflected in patient medical records, allowing an assumption that triggers would have had higher predictive value if all ADEs were accurately described by physicians.

The trigger with the highest positive predictive value was the “Hospital readmission within 30 days”. Only 2 cases of readmission were due to inefficacy of pharmacotherapy. This phenomenon presents a serious problem in psychiatry, since patients with carefully chosen and tapered treatment regimen may suddenly experience lack of effect even being fully compliant. We did not consider incidents of incompliance as ADEs. Larger sample size for data collection is needed to enable detailed analysis of the risk factors for this trigger and related ADE.

An interesting aspect of this study was the detection of the “Drug combinations not normally recommended” trigger. These combinations were determined by assessing the duplication of prescriptions and drug-drug interactions of “Major” category. The trigger did not allow to detect any ADE in a patient directly: usually, either ADE was not recorded in the patient’s medical chart or was confirmed in relation to another trigger. According to Global Trigger Tool Manual [8], this trigger “does not mean that harm has occurred but makes an ADE more likely”. At the same time, a higher MAI score had significant impact on the risk of extrapyramidal symptoms development. Thus, in our sample population the trigger “Drug combinations not normally recommended” did not show its significance as a way to detect ADEs, despite the well-known harmful consequences of irrational drug combinations in themselves [25], in our study, extrapyramidal effects.

At the same time, most triggers have been found to be unsuitable for child psychiatry. Thus, triggers should be selected more strictly when creating a new algorithm, and specific triggers should be developed for use in this population of patients.

5 Conclusion

Our study was the first one to assess ADEs of paediatric in-patients with acute psychotic episode using the Global Trigger Tool at a psychiatric hospital in Russia. More than one third of the psychotic paediatric in-patients experienced at least one ADEs. We identified three triggers, “Abrupt medication stop”, “Prescribing of extrapyramidal symptoms corrector”, and “Hospital readmission within 30 days”, to have reasonable predictive value for ADEs in adolescents with acute psychotic episode. Antipsychotic-induced extrapyramidal symptoms were more prevalent in older adolescents and patients with fewer lifetime hospital admissions. These patients need to be carefully monitored to ensure patient safety. We recommend including the identified triggers into routine systems for patient safety monitoring and emphasize the need for additional training of psychiatrists in clinical pharmacology.

Conflict of interest

The authors report no conflict of interest in this work.