Clinical predictors of repeat intramuscular medication administration for acute agitation: A retrospective cohort study

Abstract

Background:

Intramuscular (IM) antipsychotics are frequently used for acute agitation in psychiatric settings. While prior studies have focused on onset of sedation and adverse effects, the need for repeat IM medication, an important real-world marker of treatment effectiveness, has not been well characterized.

Methods:

This retrospective study included adult patients who received IM haloperidol lactate or IM olanzapine for acute agitation between January 2020 and November 2022 at an academic medical center. The primary outcome was repeat IM medication administration within 24 h. Secondary analyses examined time to repeat administration. Sociodemographic, clinical, and treatment-related variables were compared. Multivariable logistic regression and Cox proportional hazards models were used to identify independent associations.

Results:

Among 342 patients, 81 (23.7%) required repeat IM medication. Antipsychotic agent was not associated with repeat administration. In adjusted analyses, repeat medication was independently associated with co-administration of diphenhydramine (OR = 2.18), history of violence (OR = 1.88), and psychotic disorders (OR = 2.24), while substance use history (OR = 0.48) and inpatient setting (OR = 0.50) were associated with lower odds. Similar findings were observed in time-to-event analyses. Greater prior hospitalizations were associated with shorter time to repeat medication. Benzodiazepine co-administration was not independently associated with outcomes.

Conclusions:

Repeat IM medication for acute agitation is common and appears to be driven primarily by patient-level factors rather than choice of antipsychotic agent. These findings highlight the importance of underlying illness severity and agitation phenotype in determining treatment response. Identifying patients at higher risk for escalation may inform targeted management strategies and resource allocation in acute psychiatric care settings.

1. Introduction

Acute agitation is highly prevalent across acute care psychiatric settings, occurring in approximately 25% of psychiatric admissions [1]. In the emergency department setting, iťs estimated there are 1.7 million episodes of agitation annually in the United States [2]. Agitation is a transdiagnostic phenomenon, commonly observed in schizophrenia, bipolar disorder, dementia, and neurodevelopmental disorders [3]. It can also present in the course of acute substance intoxication or withdrawal [4].

Untreated agitation poses significant risks to both patients and healthcare staff [5]. Project BETA (Best Practices in the Evaluation and Treatment of Agitation), developed by the American Association for Emergency Psychiatry, provides guidance for the management of acute agitation in emergency settings, emphasizing that pharmacotherapy should be based on the assessment of the most likely etiology of the agitation, and that verbal de-escalation should be first-line treatment [6]. When verbal de-escalation, environmental modification, and oral medications are not effective or feasible, intramuscular (IM) pharmacologic management may be indicated, frequently involving antipsychotics (APs) and benzodiazepines, with selection guided by the underlying etiology of agitation [6].

Both first-generation and second-generation antipsychotics (FGAs; SGAs) are commonly utilized for this purpose, and the selection of the agent often depends on availability, institutional protocols, clinician preferences, and consideration of adverse effects [7,8]. Historically, haloperidol lactate (HL) has been the standard agent in the management of agitation [9]. However, when used alone, iťs associated with significant adverse effects including acute dystonia, which occurred so frequently in one trial that it was terminated early [10]. Due to this, current guidance does not support the use of HL monotherapy when other agents are available [9]. HL is frequently combined with anticholinergic medications, such as diphenhydramine, benztropine, or promethazine, which in addition to preventing acute extrapyramidal symptoms have been shown to increase the efficacy in reducing agitation [8,10]. HL is also often combined with benzodiazepines, such as lorazepam and midazolam, though there is mixed evidence supporting the efficacy of this combination [9].

The most frequently utilized SGA for this purpose is IM olanzapine, which demonstrates a more favorable side effect profile compared to HL [11]. Some studies have additionally found that olanzapine outperforms HL in efficacy, in addition to having fewer adverse effects and significantly less anticholinergic use [11–13]. However, the literature on the comparative effectiveness of different APs on acute agitation are considerably heterogeneous, often underpowered, and difficult to compare [8].

Previous studies have predominantly reported on the speed of onset of sedation and the prevalence of adverse effects as primary outcomes [7,9,14]. Few studies have systematically examined the need for additional medication administrations after the initial treatment, despite the clinical relevance of this outcome [9,15,16]. The need for repeat administration serves as a clinically relevant measure of treatment effectiveness, as it reflects whether or not the initial intervention was successful in managing the acute agitation. Additionally, the need for additional doses increases risks to the patient, as well as implications for nursing time, medication costs, and repeat monitoring [15]. This is also a pragmatic real-world outcome, as it does not require the administration of a clinician-rated scale or monitoring at specific time points.

Despite evolving guidance, haloperidol remains widely used in real-world settings, often in combination with other agents. Understanding its effectiveness in pragmatic outcomes such as need for repeat medication remains clinically relevant. The present study sought to examine factors associated with repeat intramuscular medication administration following initial pharmacologic treatment of acute agitation in acute psychiatric settings, including both psychiatric emergency and inpatient units. We also examined time to repeat medication administration.

2. Methods

2.1. Participants

This retrospective chart review was approved by the Institutional Review Board at the Icahn School of Medicine at Mount Sinai (Protocol 22–01627), and a waiver of informed consent was obtained. The sample included all adult patients who received intramuscular (IM) medication for the acute management of agitation at Mount Sinai Beth Israel Hospital between January 1, 2020 and November 30, 2022. Patients were included if they received either IM HL or IM olanzapine in the Comprehensive Psychiatric Emergency Program (CPEP) or on inpatient psychiatric units.

Typical agitation protocols at this institution utilize IM HL (2.5–5 mg) or IM olanzapine (5–10 mg). Patients who received other IM medications for agitation (e.g., IM chlorpromazine or IM fluphenazine) were excluded due to small sample sizes (fewer than 10 cases). Encounters were included regardless of urine drug screening status.

2.2. Data source and variables

Data were extracted from the electronic medical record by members of the study team (SL, MM, VP, CE). Interrater reliability was assessed on a subset of included patients and was high (>95%). Any discrepancies or uncertainties in data abstraction were reviewed by the study team and resolved by consensus. Extracted variables included sociodemographic characteristics (age, race, gender identity) and clinical history (number of prior psychiatric hospitalizations, history of violence, and history of substance use).

Data from the index encounter were also collected, including clinical setting (CPEP vs. inpatient psychiatry), primary billing diagnosis (categorized as mood disorder, psychotic disorder, or other), and the antipsychotic agent administered (HL or olanzapine). Medication-related variables included haloperidol-equivalent dosing and co-administration of benzodiazepines or diphenhydramine at the time of index IM administration. For the purposes of this study, we considered 10 mg of olanzapine as equivalent to 5 mg of HL, consistent with prior reports [12,13].

2.3. Outcomes

The primary outcome was receipt of an additional IM medication for agitation within 24 h of the index IM administration. A secondary outcome was time to repeat IM medication administration. Time was calculated from the initial IM administration to the first subsequent IM dose or censoring at 24 h if no repeat medication was administered.

2.4. Data analysis

Statistical analyses were conducted using IBM SPSS Statistics for Windows, Version 31.0 (IBM Corp., Armonk, NY). Descriptive statistics (means, standard deviations, and frequencies) were used to characterize the study sample. Patients who received repeat IM medication were compared to those who did not using chi-square tests for categorical variables and independent samples t-tests for continuous variables. Effect sizes were calculated for statistically significant comparisons using Cohen's d for continuous variables and Cramér's V for categorical variables. Kaplan–Meier survival analysis was also performed to visualize time to repeat IM medication administration by antipsychotic agent. For time-to-event analyses, only the first repeat IM medication was considered.

A multivariable logistic regression model was developed to examine factors independently associated with repeat IM medication administration. Variables were selected a priori based on clinical relevance and included antipsychotic agent, co-administered medications, clinical history variables, diagnosis, gender, and clinical setting. Model fit statistics including − 2 log likelihood, chi-square, and Nagelkerke R² were examined. A Cox proportional hazards model was constructed to examine factors associated with time to repeat IM medication administration.

3. Results

3.1. Study sample

A total of 342 patients who received intramuscular (IM) medication for acute agitation were included in the analysis, of whom 81 (23.7%) required an additional IM medication during the index encounter. The mean age of the sample was 40.4 years (SD = 15.2), and 60.8% identified as cisgender male. The majority of encounters occurred in the Comprehensive Psychiatric Emergency Program (CPEP) (71.9%), with the remainder on inpatient psychiatric units (28.1%). HL was administered in 70.8% of cases and olanzapine in 29.2%. Co-administration with benzodiazepines occurred in 67.8% of encounters, and with diphenhydramine in 36.5%.

3.2. Comparison between groups

Patients who required repeat IM medication differed from those who did not on several clinical and treatment-related variables (Table 1). There were no significant differences in age or race between groups. Gender distribution differed significantly, with a higher proportion of cisgender females in the repeat IM group (49.4% vs. 34.5%, p = 0.02). There was a non-significant trend toward males having higher initial AP doses, expressed as haloperidol equivalents (4.6 [SD 1.2] vs 4.3 [SD 1.3], p = 0.06).

Table 1.

Baseline characteristics and current clinical encounter compared between groups.

Variable	Total Sample (N = 342)	Received Additional IM PRN (N = 81)	No Additional PRN (N = 261)	Statistics (test statistic, df, significance, effect size)
Baseline Sociodemographic Characteristics
Age	40.4 (15.2)	39.7 (15.7)	40.6 (15.1)	t = 0.45 df = 340 p = 0.66
Gender				χ2 = 7.95
Cisgender female	130 (38.0%)	40 (49.4%)	90 (34.5%)	df = 2
Cisgender male	208 (60.8%)	39 (48.1%)	169 (64.8%)	p = 0.02*
Transgender or gender non-conforming	4 (1.2%)	2 (2.5%)	2 (0.8%)	V = 0.15
Race				χ2 = 1.82
Asian or South Asian	21 (6.1%)	6 (7.4%)	15 (5.7%)	df = 4
Black	132 (38.6%)	35 (43.2%)	97 (37.2%)	p = 0.77
Latinx	66 (19.3%)	14 (17.3%)	52 (19.9%)
White	86 (25.1%)	17 (21.0%)	69 (26.4%)
Other or Unknown	37 (10.8%)	9 (11.1%)	28 (10.7%)
Clinical History
Number of prior hospitalizations	2.3 (2.9)	3.1 (3.8)	2.0 (2.5)	t = −2.35 df = 340 p = 0.02* d = −0.37
History of violence	118 (34.5%)	35 (43.2%)	83 (31.8%)	χ2 = 3.56 df = 1 p = 0.06*
History of substance use	249 (72.8%)	52 (64.2%)	197 (75.5%)	χ2 = 3.97 df = 1 p = 0.05* V = 0.11
Current Encounter
Clinical Context				χ2 = 0.24
CPEP	246 (71.9%)	60 (74.1%)	186 (71.3%)	df = 1
Inpatient Psychiatry	96 (28.1%)	21 (25.9%)	75 (28.7%)	p = 0.62
Antipsychotic prescribed				χ2 = 3.12
Haloperidol lactate	242 (70.8%)	51 (63.0%)	191 (73.2%)	df = 1
Olanzapine	100 (29.2%)	30 (37.0%)	70 (26.8%)	p = 0.08*
Haloperidol equivalents (milligrams)	4.5 (1.2)	4.4 (1.2)	4.5 (1.2)	t = 0.75 df = 340 p = 0.46
Co-administered with diphenhydramine	125 (36.5%)	42 (51.9%)	83 (31.8%)	χ2 = 10.72 df = 1 p = 0.001* V = 0.18
Co-administered benzodiazepine	232 (67.8%)	48 (59.3%)	184 (70.5%)	χ2 = 3.58 df = 1 p = 0.06*
Billing Diagnosis				χ2 = 10.53
Mood disorder	159 (46.5%)	25 (30.9%)	134 (51.3%)	df = 2
Psychotic Disorder	159 (46.5%)	48 (59.3%)	111 (42.5%)	p = 0.005*
Other	24 (7.0%)	8 (9.9%)	16 (6.1%)	V = 0.18

Patients who required repeat IM medication had a greater number of prior psychiatric hospitalizations (mean 3.1 vs. 2.0, p = 0.02). A history of substance use was less common among those requiring repeat IM medication (64.2% vs. 75.5%, p = 0.05). A history of violence was more common in the repeat IM group, although this did not reach statistical significance (43.2% vs. 31.8%, p = 0.06).

With respect to the index encounter, there was no significant difference in clinical setting (CPEP vs. inpatient, p = 0.623) or IM antipsychotic administered (p = 0.08). Kaplan–Meier survival analysis demonstrated no significant difference between antipsychotic agents (Fig. 1). Haloperidol equivalents did not differ between groups (p = 0.46). However, co-administration with diphenhydramine was significantly more common among patients requiring repeat IM medication (51.9% vs. 31.8%, p = 0.001), while benzodiazepine co-administration showed a non-significant trend toward being less common (59.3% vs. 70.5%, p = 0.06).

Fig. 1.

Kaplan-Meier survival analysis of time to next administration between antipsychotic agents.

χ² = 3.37, df = 1, p = 0.07.

Haloperidol lactate: 242 patients, 51 events; Olanzapine: 100 patients, 30 events.

Exploratory analyses were conducted to examine factors associated with diphenhydramine co-administration. Patients who received diphenhydramine were administered lower initial antipsychotic doses compared to those who did not (mean haloperidol equivalents 4.2 vs 4.6, p = 0.005). There were no significant differences in diphenhydramine use by race or diagnosis, though there was a non-significant trend toward greater use among patients with a history of violence (p = 0.06).

Diagnostic distribution differed significantly between groups (p = 0.005), with a higher proportion of psychotic disorders among those requiring repeat IM medication (59.3% vs. 42.5%) and a lower proportion of mood disorders (30.9% vs. 51.3%).

3.3. Multivariable analyses

A multivariable logistic regression model was developed to examine factors associated with repeat IM medication administration (Table 2). The overall model was statistically significant (χ² = 43.45, df = 11, p < 0.001), with a − 2 log likelihood of 330.98 and Nagelkerke R² of 0.18.

Table 2.

Logistic regression model predicting repeat medication administration.

Variable	B	Wald	Df	Sig.	Odds Ratio	95% CI
Olanzapine (compared to haloperidol)	−0.50	0.56	1	0.45	0.60	0.16–2.25
Co-administered diphenhydramine	0.78	6.08	1	0.01	2.18	1.17–4.05
Co-administered benzodiazepine	−0.61	0.84	1	0.36	0.54	0.15–2.00
Number of prior admissions	0.08	2.78	1	0.10	1.08	0.99–1.18
History of violence	0.63	4.42	1	0.04	1.88	1.04–3.40
History of substance use	−0.74	5.58	1	0.02	0.48	0.25–0.88
Billing diagnosis (mood disorder as reference)	–	7.41	2	0.03	–	–
Psychotic disorder	0.81	6.73	1	0.009	2.24	1.22–4.11
Other disorder	0.84	2.74	1	0.10	2.32	0.86–6.27
Gender (cisgender female as reference)	–	9.15	2	0.01	–	–
Cisgender male	−0.74	6.72	1	0.01	0.48	0.27–0.83
Transgender or gender non-conforming	1.31	1.43	1	0.23	3.70	0.43–31.55
Clinical context (CPEP as reference)	−0.70	4.17	1	0.04	0.50	0.26–0.97

χ² = 43.45, df = 11, p < 0.001, – 2 log likelihood = 330.98, Nagelkerke R² = 0.18.

In the adjusted model, repeat IM medication administration was independently associated with co-administration of diphenhydramine (OR = 2.18, 95% CI 1.17–4.05, p = 0.01), history of violence (OR = 1.88, 95% CI 1.04–3.40, p = 0.04), and a diagnosis of psychotic disorder compared to mood disorder (OR = 2.24, 95% CI 1.22–4.11, p = 0.009). History of substance use was associated with lower odds of repeat IM medication (OR = 0.48, 95% CI 0.26–0.88, p = 0.02), and inpatient setting was associated with lower odds compared to CPEP (OR = 0.50, 95% CI 0.26–0.97, p = 0.04). Gender was also associated with repeat IM medication, with cisgender male patients having lower odds compared to cisgender female patients (OR = 0.48, 95% CI 0.27–0.83, p = 0.01). Number of prior psychiatric hospitalizations showed a trend toward increased odds of repeat IM medication (OR = 1.08 per admission, p = 0.10). Antipsychotic agent (olanzapine vs. HL) and benzodiazepine co-administration were not significantly associated with repeat IM medication (Table 2). Interaction analyses examining whether race moderated the association between prior hospitalizations or history of violence and repeat IM medication were not statistically significant (p = 0.36 and p = 0.73, respectively), suggesting that these relationships did not differ by race.

A Cox proportional hazards model was then constructed to examine time to repeat IM medication administration. The overall model was statistically significant (χ² = 46.72, df = 11, p < 0.001) (Table 3). Shorter time to repeat IM medication was associated with co- administration of diphenhydramine (HR = 1.95, 95% CI 1.18–3.24, p = 0.010), history of violence (HR = 1.81, 95% CI 1.11–2.97, p = 0.02), greater number of prior psychiatric hospitalizations (HR = 1.07 per admission, p = 0.028), and a diagnosis of psychotic disorder (HR = 1.98, 95% CI 1.19–3.29, p = 0.009). History of substance use (HR = 0.51, 95% CI 0.31–0.84, p = 0.008) and inpatient setting (HR = 0.53, 95% CI 0.31–0.93, p = 0.03) were associated with longer time to repeat IM medication. Antipsychotic agent and benzodiazepine co-administration were not significantly associated with time to repeat IM medication (Table 3).

Table 3.

Cox Regression Predicting Hazard of Earlier Administration.

Variable	B	Wald	Df	Sig.	Hazard Ratio	95% CI
Olanzapine (compared to haloperidol)	−0.44	0.61	1	0.44	0.65	0.22–1.94
Co-administered diphenhydramine	0.67	6.67	1	0.01	1.95	1.18–3.24
Co-administered benzodiazepine	−0.55	0.96	1	0.33	0.58	0.19–1.74
Number of prior admissions	0.07	4.83	1	0.03	1.07	1.01–1.14
History of violence	0.59	5.58	1	0.02	1.81	1.11–2.97
History of substance use	−0.68	7.00	1	0.008	0.51	0.31–0.84
Billing diagnosis (mood disorder as reference)	–	7.41	2	0.03	–	–
Psychotic disorder	0.68	6.83	1	0.009	1.98	1.19–3.29
Other disorder	0.72	2.97	1	0.09	2.05	0.91–4.63
Gender (cisgender female as reference)	–	10.47	2	0.005	–	–
Cisgender male	−0.65	7.60	1	0.006	0.52	0.33–0.83
Transgender or gender non-conforming	0.93	1.54	1	0.22	2.54	0.58–11.09
Clinical context (CPEP as reference)	−0.63	4.96	1	0.03	0.53	0.31–0.93

χ² = 46.72, df = 11, p < 0.001, – 2 log likelihood = 879.69.

4. Discussion

In this retrospective cohort of patients receiving intramuscular medication for acute agitation, approximately one quarter required repeat medication administration. The choice of antipsychotic agent was not independently associated with repeat medication use. However, repeat medication administration was associated with clinical characteristics including psychotic disorders, history of violence, and prior psychiatric hospitalizations. A key consideration in interpreting these findings is the high rate of co-administration of adjunctive medications, particularly benzodiazepines and diphenhydramine. As a result, this study does not represent a pure comparison of haloperidol versus olanzapine monotherapy, but rather reflects real-world treatment patterns in which combination pharmacotherapy is common. Co-administration may influence both initial response and the need for repeat medication, and likely reflects clinician perception of agitation severity or etiology.

The results of this study suggest that the selection of antipsychotic may be less important than underlying patient factors, as both HL and olanzapine demonstrated comparable rates of repeat medication administration in this study. This is consistent with prior literature showing comparable efficacy [12,13,17]. These findings support existing recommendations to individualize treatment based on clinical presentation rather than relying on a single preferred agent, and reinforce the importance of considering both risks and real-world effectiveness when selecting pharmacologic strategies. Although this study focused on HL and olanzapine, other agents are also used in the pharmacologic management of acute agitation. For example, prior studies have suggested that agents such as ketamine have been associated in some studies with more rapid sedation [18]. These comparisons were outside the scope of the present study, which was designed to examine real-world predictors of repeat intramuscular medication among the two most commonly used antipsychotic agents at our institution.

Independent predictors in this study included psychosis, history of violence, and prior admissions. These variables may reflect more refractory agitation or be associated with a more chronic and severe illness necessitating more intensive management. This is consistent with prior studies which have found that psychotic disorders and greater illness chronicity may be associated with risk for violence and agitation [2,19]. These findings suggest that the underlying phenotype of agitation may influence treatment response and the need for escalation.

In this sample, patients with co-occurring substance use were less likely to have repeat medication administrations, both in univariate and multivariate analyses. This finding was somewhat counterintuitive, as some studies have suggested that substance-related agitation may require more frequent and ongoing intervention [16]. However, others have suggested that patients with substance use may present with more transient agitation, related to acute intoxication or withdrawal [20]. Alternatively, agitation secondary to substance use may respond more easily to pharmacologic intervention, consistent with prior findings from the emergency setting [21]. We were unable to differentiate between specific substances (e.g., stimulants vs opioids), which may have distinct effects on agitation phenotype and treatment response.

Medication-related factors also played an important role in repeat medication use. Co-administration with diphenhydramine was associated with both increased likelihood of repeat medication administration as well as earlier administration. One possible explanation is confounding by indication, as diphenhydramine may be more likely to be used in cases of more severe agitation [16,22]. Supporting this, there was a non-significant trend toward greater diphenhydramine use among patients with a history of violence.

Additionally, exploratory analyses demonstrated that patients receiving diphenhydramine were administered lower initial antipsychotic doses. This raises the possibility that clinicians may, in some cases, use diphenhydramine as an adjunct in place of higher antipsychotic dosing, which could contribute to insufficient initial treatment and a greater likelihood of requiring repeat medication. Alternatively, diphenhydramine may be preferentially used in anticipation of extrapyramidal side effects, particularly in combination with haloperidol, reflecting clinician concern rather than a direct causal effect on treatment response [16,22,23]. Interestingly, benzodiazepine co-administration was not associated with repeat medication, in contrast to diphenhydramine. This may reflect differences in clinical use, as benzodiazepines are often selected based on suspected etiology (e.g., substance-related agitation), whereas diphenhydramine may be used more reflexively alongside haloperidol or in cases perceived as more severe.

Gender was also associated with repeat IM medication administration, with cisgender male patients demonstrating lower odds and longer time to repeat medication compared to cisgender female patients. The reasons for this finding are not immediately clear. Some studies have found that males are more likely to present with agitation than females, though results overall have been heterogeneous [24–27]. It is possible that differences in the presentation or perceived severity of agitation across genders may influence clinical decision-making, including thresholds for additional medication administration. Alternatively, this finding may reflect unmeasured confounding variables, such as differences in diagnostic distribution, substance use patterns, or behavioral presentation. Although there was a non-significant trend toward higher initial dosing among male patients, this difference did not reach statistical significance, suggesting that dosing differences alone are unlikely to fully explain the observed gender differences in repeat medication. Given the exploratory nature of this finding and the limited prior literature examining gender differences in acute agitation management, it should be interpreted with caution.

This study is not without limitations. As a retrospective chart review conducted at a single academic medical center, the findings may not be generalizable to other clinical settings or patient populations. We did not have access to the overall racial distribution of patients presenting to the psychiatric emergency and inpatient settings during the study period, limiting our ability to determine whether the study sample reflects underlying population demographics or potential differences in treatment patterns. Additionally, the observational design introduces the possibility of confounding by indication, as medication selection and co-administration practices were determined by treating clinicians rather than a standardized protocol. As such, associations between treatment variables (e.g., diphenhydramine co-administration) and repeat IM medication may reflect underlying agitation severity or clinician response rather than causal effects. Importantly, measures of agitation severity were not consistently available, as standardized assessments such as the Brøset Violence Checklist were not documented for many patients. As a result, the analyses relied on proxy measures of illness severity, including diagnosis, history of violence, and prior psychiatric hospitalizations, which may incompletely capture the clinical context of agitation. We did not find evidence that race moderated the association between clinical history variables and repeat medication use, though the study may have been underpowered to detect more subtle interaction effects.

Differences between care settings (emergency vs. inpatient psychiatry) may have influenced treatment practices and thresholds for repeat medication administration, and although clinical context was included in multivariable models, residual confounding related to environmental or staffing factors may remain. Simultaneously, medication-related variables were limited to those documented in the electronic medical record, and the timing and rationale for co-administration of adjunctive agents such as benzodiazepines or diphenhydramine were not always clearly specified. Additionally, substance use was determined based on documented history rather than systematic drug screening results, which may have led to misclassification. Exploratory analyses examining factors associated with diphenhydramine use were not adjusted for all potential confounders and should be interpreted cautiously Finally, the use of multiple analyses on a limited sample size increases risk for Type 1 error, and no corrections were performed. However, effect sizes were calculated to provide an estimate of clinical significance.

Strengths of the study include a relatively large, real-world sample of patients receiving intramuscular medication for acute agitation across both emergency and inpatient psychiatric settings, enhancing the clinical relevance of the findings. The inclusion of both emergency and inpatient contexts allowed for examination of agitation management across different care environments. Additionally, the study focused on pragmatic and clinically meaningful outcomes, including repeat intramuscular medication administration and time to repeat dosing, which reflect real-world treatment effectiveness and resource utilization without requiring standardized rating scales. The use of multivariable logistic and Cox regression models also enabled adjustment for multiple clinically relevant factors, and the consistency of findings across these analytic approaches strengthens confidence in the results.

5. Conclusion

In this retrospective cohort of patients treated with intramuscular medication for acute agitation, repeat medication administration was associated primarily with patient-level clinical characteristics, including psychotic disorders, history of violence, and prior psychiatric hospitalizations, rather than the specific antipsychotic agent administered. These findings suggest that the underlying clinical context of agitation may play a greater role in treatment escalation than initial pharmacologic choice. Identifying patients at higher risk for repeat medication use may help inform clinical decision-making and guide more targeted management strategies in acute psychiatric settings. Future studies should prospectively examine the relationship between clinical presentation, treatment strategies, and escalation outcomes in acute agitation, incorporating standardized measures of agitation severity and more detailed characterization of medication regimens. Additionally, studies evaluating interventions targeted toward patients at higher risk for repeat medication use may help improve outcomes and reduce the need for repeated pharmacologic or behavioral interventions.

Data availability

The data that support the findings of this study are available from the corresponding author, Sean T. Lynch, upon reasonable request.