Abstract
Objective:
Delirium in the setting of critical illness occurs in 15%–60% of pediatric patients and is associated with increased duration of hospital admission. Strategies for symptom management include minimizing exposure to deliriogenic medications, optimizing day–night cycles, and in some cases prescribing antipsychotic medications. Symptom management occurs while simultaneously identifying and treating the underlying etiology. Over a 5-year period, our pediatric intensive care units implemented several unit-based initiatives to prevent and minimize the severity of delirium in critically ill children. The objective of this study is to retrospectively characterize antipsychotic prescribing practices over the study period.
Methods:
Over a 5-year timeframe in our children’s hospital, we extracted dispensing data for risperidone, olanzapine, quetiapine, and haloperidol. We used descriptive statistics to characterize patients initiated on an antipsychotic in an ICU area, dosing information, and quantitative change in annual orders over the study period.
Results:
During the study period, 533 antipsychotic orders were placed for 165 patients. Most of the identified patients were male (60.6%) and Black (44.2%). Ninety (16.9%) of the medication orders remained active after patient transfer to a general care floor. Risperidone was the most prescribed antipsychotic (54%), with a 413% increase observed between 2016 and 2018, followed by a decline of 89% between 2018 and 2021. Use of other antipsychotics was less than risperidone but consistent during the study period. Risperidone was more commonly prescribed for younger children, and prescribing patterns suggest a preference for nighttime administration.
Conclusions:
Implementation of unit-based initiatives to promote delirium screening and awareness led to an initial increase in antipsychotic medication prescribing for critically ill children with delirium. This was followed by a significant decline in antipsychotic prescribing with the inclusion of the child and adolescent psychiatry team in our delirium management and prevention algorithm. This study highlights the importance of collaboration with the child and adolescent psychiatry team in the management of pediatric delirium across the hospital system. Additional research on antipsychotic prescribing practices and the associated impact of nonpharmacologic interventions and multidisciplinary collaboration for delirium treatment is warranted.
Introduction
Delirium is common in the setting of critical illness, need for mechanical ventilation, neurodevelopmental disability, and prolonged hospitalization (Malas et al., 2017; Schieveld et al., 2014; Traube et al., 2017). Delirium is defined as an alteration in consciousness and attention due to underlying medical illness and/or exposure to medications that impact cognition (American Psychiatric Association, eds, 2013). Epidemiological studies suggest a prevalence of 15%–60% in pediatric patients with heterogeneity owing to age, underlying illnesses, and differences in diagnostic practices (Ricardo Ramirez et al., 2019; Siegel and Traube, 2020; Traube, 2018).
Management of delirium among pediatric patients can be broadly categorized as physiological, environmental, and pharmacologic (Silver et al., 2019). Physiological strategies include identifying and treating medical illness, optimizing nutrition and bowel/bladder function, promoting physical activity and mobilization in daylight hours, and mitigating nighttime interruptions. Environmental strategies include reorientation of the patient, increasing exposure to natural sunlight, minimizing excessive noise, moderating temperature, and educating the patient and family about delirium. Pharmacologic management includes minimizing deliriogenic medication exposure, managing pain, and use of antipsychotic agents when delirium becomes distressing to the patient or family or limits care (Silver et al., 2019). Quetiapine, olanzapine, risperidone, and haloperidol have all been safely prescribed in the pediatric population; however, second-generation antipsychotics are associated with fewer neuromuscular and prolactin-elevating adverse effects (Capino et al., 2020; Joyce et al., 2015; Kishk et al., 2019; Sassano-Higgins et al., 2013).
There remains a paucity of evidence outlining best practice for use, particularly in conjunction with nonpharmacologic treatment modalities. For adult patients, a meta-analysis considering the use of antipsychotics for the management of delirium indicated that the risks of antipsychotic use may outweigh benefits (Oh et al., 2019). For children, in addition to known cardiac, metabolic, and neuromuscular adverse effects, we do not understand the impact of antipsychotic use on developing brains (Cavagnero et al., 2025). Thus, it is important to prioritize the use of nonpharmacologic strategies and consultation with a child and adolescent psychiatry team for patients with delirium.
A previous chart review from our institution indicated that we under-recognized delirium on our inpatient services; however, in the intervening years, the multifaceted PICU Up! early mobility initiative has substantially increased delirium screening and diagnosis (Kelly and Frosch, 2012; Wieczorek et al., 2016). Incorporating child and adolescent psychiatry, an invaluable resource for the care of critically ill children with delirium, into our delirium management and prevention algorithm allowed for early involvement from that team. With increasing recognition and collaboration between medical and psychiatric teams, there has been an increase in awareness, screening, and interdisciplinary treatment (Barnes et al., 2018). The impact of these initiatives on antipsychotic prescribing remains unknown. The primary aim of this study was to evaluate antipsychotic prescribing practices in the intensive-care units at our tertiary-care children’s hospital over time.
We hypothesized that with increasing national and institutional awareness of pediatric delirium, we would see an increase in recognition of delirium in ICU settings with a subsequent increase in prescribing of antipsychotics for symptom management (Smith et al., 2011; 2016). Then, with the publication of a delirium clinical pathway and the shift in our hospital ICU’s approach to encourage early child and adolescent psychiatry consultation, we may see a decrease in antipsychotic prescribing with a goal of antipsychotic stewardship and multimodal delirium management strategies (Silver et al., 2019). This study is a retrospective data pull considering these hypotheses.
Materials and Methods
This project was acknowledged by our hospital’s Institutional Review Board as quality improvement/non-human subject research. Informed consent was waived as data were collected retrospectively and were a review of routine clinical care. The institution is a 196-bed tertiary care children’s hospital with 28 pediatric intensive care (PICU) beds, 12 pediatric cardiac intensive care (PCICU) beds, and 49 neonatal intensive care beds, serving patients from birth through age 25. The study population includes all children in an ICU setting who were prescribed an antipsychotic commonly used for delirium management (risperidone, olanzapine, quetiapine, and haloperidol) from June 2016 through April 2021. Dispensing data for this population were extracted retrospectively from the medical record. June 2016 was chosen as a starting point because this is when our institution implemented the use of Epic® for medical record keeping. The data were filtered to include all antipsychotics initiated in an ICU setting.
At the time of this study, there was no child and adolescent psychiatry engagement in the Neonatal ICU. The PICU and PCICU units admit patients from infancy (who had previously been discharged home from the hospital) through age 22, and occasionally up to age 26 for select patient populations. The Neonatal ICU admits patients in need of immediate intensive care after birth, other than those patients in need of Extracorporeal Membrane Oxygenation (ECMO) or cardiac surgery who were transferred to the PICU and PCICU, units respectively. There are no institutional guidelines surrounding the age at which antipsychotics can be administered.
Beginning in 2016, our PICU and PCICU nurses screened for delirium at bedside once per shift using the pediatric Confusion Assessment Method (pCAM) or preschool Confusion Assessment Method (psCAM). Prior to 2021, if an infant was under 6 months of age, the psCAM was inconsistently used, as it had not yet been validated for use in infants under 6 months. If a patient screened positive for delirium with use of this tool, the primary team was notified and confirmed the diagnosis based on their own clinical exam. Child and adolescent psychiatry was consulted for patients screening positive for delirium based on the comfort of the primary team, with increased awareness of and more consistent consultation with the child and adolescent psychiatry team after 2018.
Administration data were pulled from MAR activity (meds listed as administered) for ICU patients, for the select medications, during the study time frame. Order data were pulled from pharmacy analytics data, specifically orders that migrated to the pharmacy queue for verification for the select medication during the study period. All orders were initiated by ICU prescribers (rather than by psychiatry) during the admission, confirmed by the hospital admission and discharge dates for each patient included.
Demographic data, hospital length of stay, continuation of therapy upon transfer out of the ICU and upon hospital discharge, all prescribing data, and annual dispensing of each antipsychotic agent during the study period were characterized. Medication initiations, defined as initial medication order or reinstatement of previously discontinued medication order, per patient, were also quantified and characterized. It is possible that initial medication orders could reflect continuation of home therapy; as a proxy to determine prescribing for delirium, a manual search for the term “delirium” in each of the 165 medical records was completed. This search queried the term “delirium” mentioned in the medical record, though did not confirm a diagnosis of delirium in the problem list for each patient. Descriptive statistics were utilized to characterize the prescribing practices. Means (standard deviation [SD]) were used for normally distributed data and median (interquartile range [IQR]) for nonnormally distributed data.
Results
A total of 533 antipsychotic orders were identified for 165 unique patients treated in the ICU setting between June 2016 and April 2021. Most patients (143 [86.7%]) had one admission during the study period (Table 1), though several patients had between two and five readmissions (Table 1). The median (IQR) hospital length of stay was 15.5 (4.6–38.6) days. The majority of the identified patients were male (60.6%) and Black (44.2%) (Table 2).
Admissions by Frequency
Admissions by Race
Most antipsychotic orders were for risperidone (288 [54%]) (Fig. 1), for which a 413% increase in prescribing was observed between 2016 and 2018, followed by an 89% decrease in prescribing through 2021.
Total antipsychotic medication orders from June 2016 to April 2021. Hospital formulary included the following formulations of each medication: risperidone: liquid, m-tab, and tablet; quetiapine: tablet; olanzapine: intramuscular, tablet, and oral disintegrating tablet; haloperidol: intravenous, tablet and liquid.
Use of the other antipsychotics was less common but consistent during the study period (Fig. 2). Table 3 shows patient and prescribing information for each antipsychotic order. Orders for olanzapine, quetiapine, and haloperidol were more commonly written for adolescents. Risperidone was more commonly prescribed for children with a mean age of 10 years, and prescribing patterns suggest a preference for nighttime administration. Haloperidol was used proportionally more often as a pro re nata, or as-needed medication, whereas olanzapine, quetiapine, and risperidone were typically scheduled once or twice daily.
Antipsychotic medication orders by medication by year from 2016 to 2021.
Antipsychotic Medication Orders by Age, Weight (with Dosing), and Duration
Ninety (16.9%) of the PICU orders were continued upon transfer to the general pediatric floors, and 28 (5.3%) were continued upon discharge.
Given initiation during the critical illness period and discontinuation of 94.7% of antipsychotic orders by discharge, we inferred that antipsychotic prescriptions were for management of symptoms associated with delirium. A manual chart search found the term “delirium” in 107 out of 165 (65%) of patient records.
Discussion
Our study demonstrates that prescription of antipsychotic medication in our center increased each year from 2016 to 2018, followed by a decrease in subsequent years. The previous study by Wieczorek et al., which focused on early identification of delirium and proactive rehabilitation, increased awareness of delirium (Wieczorek et al., 2016). With the publication of this article, the PICU focused its efforts on consistent screening by bedside RN staff using the pCAM and psCAM (Smith et al., 2011; 2016). It is possible that this increase in recognition led to the rise in prescribing medication for the management of delirium. The peak of prescribing occurred in 2018, followed by a subsequent decline. That year, the PICU modified their delirium management and prevention algorithm to emphasize early child and adolescent psychiatry consultation for all children who screened positive for delirium on the pCAM or psCAM. The process of psychiatric consultation includes a thorough chart review with a focus on illness course, lab work and medical imaging, and current and recent pharmacotherapy. After chart review, the team gathers a comprehensive history and completes a mental status exam with a focus on pre-illness baseline, changes in sleep, behavior, attention, new perceptual disturbances, and other changes in level of functioning. An assessment of the environment in the room, including monitors, availability of natural light, and visual reminders of home and date/time/location is also considered. The child and adolescent psychiatry team considers the potential contributing factors to delirium and limits use of antipsychotic prescribing to those patients for whom delirium is leading to safety concerns, significant distress, inability to wean sedating medications, or inability to receive necessary medical care. We hypothesize that normalizing the interdisciplinary care of critically ill children with delirium improved awareness of modifiable risk factors and re-emphasized prioritization and utilization of non-pharmacological strategies (Barnes et al., 2018).
Though antipsychotics are effective for the management of impulsivity and aggression associated with delirium, they should be limited to the duration of symptoms and weaned as soon as possible upon improvement of distressing symptoms associated with delirium. For our study population, the vast majority of medications were discontinued with transfer from ICU-level care to the general pediatrics unit, but 16.9% of patients remained on antipsychotics at transfer. The child and adolescent psychiatry team follows these patients after discharge from the ICU, providing continuity in their care and guiding in the cessation of antipsychotic medications. Antipsychotic pharmacotherapy was terminated for many patients prior to discharge home, with the exception of 5.3% of the study population. The cause for the continuation of antipsychotic medications in this group is unknown. Continuation could be attributed to prolonged effects of delirium, apprehension to wean medications in conjunction with other therapy modifications, home medication prior to hospitalization, or prescriber error upon discharge.
Patient and prescribing information for each antipsychotic order showed a preference for younger children to be prescribed risperidone (mean age of 10.3 ± 6.5), whereas olanzapine, quetiapine, and haloperidol were more likely to be prescribed for adolescents. These prescribing preferences may reflect an institutional bias towards olanzapine, quetiapine, and haloperidol being used in older patients with unsafe behaviors. This finding may also be due to pharmacy formulary restrictions. Specifically at our institution during the study period, risperidone was the only atypical antipsychotic available in a liquid formulation, allowing for administration of smaller doses for young patients.
Overall, there was a trend towards prescribing antipsychotic medications at night. One explanation for nighttime prescribing is for the possible benefit of sleep promotion. Promoting a consolidated sleep-wake cycle in critically ill patients is an important physiologic strategy for preventing or minimizing delirium. Combined with nonpharmacologic modalities such as early mobilization, it is possible that timing antipsychotic administration for the evening may provide some sedative properties. Thus, while antipsychotics should not be used with a primary aim of promoting sleep, if use is indicated for symptom management of delirium, it can be beneficial to preferentially prescribe at night to promote sleep/wake cycles. Nighttime use can also limit the need for additional pharmacologic sleep aids in patients with disrupted sleep. The duration of prescribing was roughly 2 to 3 days for the 2nd generation antipsychotics, whereas for haloperidol, the duration was shorter. As a 1st-generation antipsychotic, haloperidol is associated with a higher risk of extrapyramidal side effects, and IV haloperidol is associated with a greater risk of QTc prolongation and need for cardiac monitoring. These considerations for increased side effect burden likely explain the shorter duration of haloperidol prescribing when compared with 2nd-generation antipsychotics.
Limitations of this retrospective study include the fact that inferences are based on trends of prescribing data. Additionally, as many as 35% of the medical records for our sample population did not include the term “delirium,” which indicates that there is some heterogeneity in our sample and raises the possibility that delirium is not coded in the medical record even when recognized or treated. At the time of this study, our delirium screening tool was only validated for children 6 months of age and older. There may be some patients for whom antipsychotics were prescribed owing to concern for delirium but who were not screened because of their age. Additionally, some portion of the antipsychotic prescriptions were for patients who were continuing a home medication and not being treated for delirium. We did not have access to data about pre-existing home medications for this study; however, the percentage of patients who were continued on antipsychotics after discharge was negligible and would not skew the data or inferences made.
Conclusion
Implementation of unit-based initiatives to promote delirium screening and awareness was associated with an initial increase in antipsychotic medication prescribing for critically ill children with delirium. That increase was followed by a significant decline in antipsychotic prescribing with the inclusion of child and adolescent psychiatry in delirium management and the prevention algorithm. This study highlights the importance of collaboration with child and adolescent psychiatry in the management of pediatric delirium across the hospital system. Additional research on antipsychotic prescribing practices, the impact of nonpharmacologic interventions and multidisciplinary collaboration for delirium management is warranted.
Footnotes
Author Disclosure Statement
No competing financial interests exist.
Funding Information
No funding was received for this article.