Disparities in Pharmacologic Restraint Use in Pediatric Emergency Departments

We conducted a cross-sectional study of ED visits to pediatric hospitals in the Pediatric Health Information System (PHIS) by patients aged 3 to 21 years with MBH conditions between January 1, 2010, and December 31, 2020. The PHIS is an administrative database containing inpatient, ED, ambulatory surgery, and observation encounter-level data from more than 49 not-for-profit tertiary care pediatric hospitals in the United States. These hospitals are affiliated with the Children’s Hospital Association (Lenexa, KS). We have previously used the PHIS database to describe pharmacologic restraint use in pediatric patients with MBH conditions in the ED.²¹  Children younger than 10 years of age were included in the study based on recent literature reporting an increase in young children between 3 and 10 years old seeking care in the ED for MBH complaints.^26,27  Only ED visits, as designated specifically in the PHIS database, were included for this study. Data quality and reliability are assured through a joint effort between the Children’s Hospital Association and participating hospitals. Portions of the data submission and data quality processes for the PHIS database are managed by Truven Health Analytics (Ann Arbor, MI). For the purposes of external benchmarking, participating hospitals provide discharge/encounter data including demographics, diagnoses, and procedures. Nearly all these hospitals also submit resource utilization data (eg, pharmaceuticals, imaging, laboratory) into PHIS. Data are deidentified at the time of data submission and are subjected to multiple reliability and validity checks before inclusion in the database. For this study, only data from the 30 hospitals with complete discharge and billing data available for the entire study period were included (Supplemental Fig 2).

MBH visits were identified using International Classification of Disease, Ninth Revision (ICD-9) and International Classification of Disease, Tenth Revision (ICD-10) codes. We used a previously validated classification system that organizes MBH diagnoses into 30 groupings by ICD-9 and ICD-10 diagnostic codes (Supplemental Table 3).^28–30  We included ED visits with a primary diagnosis within 1 of the 30 MBH diagnosis groupings.²⁵ 

For each visit, we collected demographic data including patient age in years (young children 3-7, school-age children 8-12, adolescents 13-17, young adults 18-21), sex, race, ethnicity, insurance type, ED disposition (admission to the hospital, transfer to another medical facility, discharge from the ED, leaving against medical advice, or death), time of day of ED arrival (day [8 am-4 pm], evening [4 pm-12 am], overnight [12 am- 8 am]), ED arrival day of week (weekend versus weekday), and whether the visit was an initial or repeat visit with a MBH diagnosis during the study period.^19,25–27  Repeat visits with a primary MBH condition were counted independently and categorized as a repeat visit for analysis. Hospital-level characteristics collected were the average annual volume of MBH visits and ED visits.

Pharmacologic restraint use was defined as the presence of a billing code for any first- or second-generation antipsychotic medication or medication previously cited in the literature as being used for pediatric pharmacologic restraint administered via intramuscular or intravenous routes during the visit (Supplemental Table 4).^31–33  Among ED encounters resulting in hospital admission, pharmacological restraint was defined as positive only when administered on day 1 or 2 of the encounter because medication timing was specified in the dataset only by calendar day and not by hour or setting (ED versus inpatient).²⁰  Visits with pharmacologic restraint were included within the analysis but the total number of pharmacologic restraint administrations within each visit was not quantified.

To describe visit level factors, we calculated frequencies of ED visits with pharmacologic restraints use overall and stratified by patient demographic and hospital characteristics. We used the χ² test to compare pharmacologic restraint use by age, sex, race, ethnicity, payer, disposition, time of day, and visit frequency. We accounted for the inflated possibility of type I errors because of multiple statistical tests with the Benjamini-Hochberg procedure of the false discovery rate.³⁴  To examine hospital-level use of pharmacologic restraints, hospitals were categorized by average annual MBH visits and average annual ED volume. We estimated a linear regression model with the hospital-level proportion of number of visits with pharmacologic restraint use as the dependent variable and the average annual ED MBH volume adjusted per 100 visits and average annual ED volume adjusted per 10 000 visits as the independent variables.

To investigate patient-level factors associated with pharmacologic restraint, we estimated univariate logistic regression models with pharmacologic restraint use as the dependent variable. Any visit-level characteristic that had a statistically significant association with pharmacologic restraint use was included as an independent variable in a multivariable model. Based on the multivariable model results, we conducted a secondary analysis for statistically significant race variables because race has been previously associated with disparities in physical restraint use for MBH conditions in the ED.^23,24,35  We also conducted a secondary analysis for serious mental illness (SMI) diagnosis groupings (schizophrenia spectrum and other psychotic disorders, depressive disorders, bipolar and related disorders) based on prior literature reporting associations with physical restraint use and psychosis or bipolar disorder.^22,36  Logistic regression models used robust standard errors clustered on hospital to account for intrahospital correlation. All effect estimates were reported as odds ratios (ORs) with 95% confidence intervals (CIs). All statistical tests were 2-tailed, and α was set at 0.05. Data analyses were performed using Stata version 16.0 (Stata Corp, College Station, TX). This study was declared exempt by the institutional review board at the study institution.

Among the 30 study hospitals, there were 545 800 visits to pediatric EDs with primary MBH diagnoses. Of these visits, 22 194 (4.1%) involved pharmacologic restraint use. A higher proportion of adolescents (13-17 years old) received pharmacologic restraint (60.8%, P < .001) compared with other age categories; however, this age group also comprises the largest proportion of MBH ED visits (59.6%; Table 1). Pharmacologic restraint use was more common in patients with public (55.6%) compared with private (37.9%) insurance (P < .001). ED visits with an evening arrival time and weekend visit experienced more frequent pharmacologic restraint (46.1%, 23.1%, P < .001), as did visits resulting in hospital admission (62.3%, P < .001).

Univariable logistic regression analyses identified statistically significant associations of patient-level factors and pharmacologic restraint use (Table 2). In the multivariable logistic regression analysis, after adjusting for patient-level characteristics, pharmacologic restraint was associated with young adult age (OR, 1.88; 95% CI, 1.59–2.22), male sex (OR, 1.25; 95% CI, 1.16–1.34), and Black race (OR, 1.22; 95% CI, 1.09–1.35). Furthermore, in this adjusted analysis, visits originating during overnight hours (OR, 1.68; 95% CI, 1.44–1.96), visits originating on the weekend (OR, 1.26; 95% CI, 1.22–1.30), and repeat ED visits (OR, 1.31; 95% CI, 1.17–1.47) were associated with pharmacologic restraint. Because Black race was the only statistically significant race associated with pharmacologic restraint use, we conducted a secondary analysis among Black patients. This multivariable model identified associations with young adult age (OR, 1.86; 95% CI, 1.47–2.36), male sex (OR, 1.34; 95% CI, 1.22–1.48), visits originating during overnight hours (OR, 1.34; 95% CI, 1.18–1.53), weekend visits (OR, 1.22; 95% CI, 1.14–1.3), and repeat ED visits (OR, 1.54; 95% CI, 1.35–1.75) (Supplemental Table 5). A secondary analysis among visits within diagnosis groupings of SMI was performed and revealed associations with young adult age (OR, 1.5 2; 95% CI, 1.09–2.11), male sex (OR, 1.80; 95% CI, 1.61–2.02), Black race (OR, 2.13; 95% CI, 1.69–2.70), Asian race (OR, 1.61; 95% CI, 1.08–2.41), public insurance (OR, 1.33; 95% CI, 1.14–1.54), visits originating during overnight hours (OR, 1.58; 95% CI, 1.24–2.01), weekend visits (OR, 1.40; 95% CI, 1.29–1.53), and repeat ED visits (OR, 1.56; 95% CI, 1.30–1.88) (Supplemental Table 6).

Across included pediatric EDs, there was a statistically significant inverse relationship between the average annual volume of patients with MBH conditions seen and the rates of overall restraint use. Every 100-visit increase in average annual MBH volume was associated with a 0.09% decrease in restraint use (95% CI, –0.15 to –0.04) (Fig 1A). However, there was no significant relationship between average annual ED volume and pharmacologic restraint use (95% CI, –0.25 to 0.25) (Fig 1B).

In this large, multicenter, 11-year review examining pharmacologic restraint use in children and youth presenting with MBH conditions to U.S. pediatric EDs, higher odds of restraint use were associated with patients who were young adults (18-21 years), males, of Black race, and had repeat ED visits for an MBH condition. Time-specific factors associated with pharmacologic restraint included overnight or weekend arrival times. Finally, we found that pharmacologic restraint decreased as average annual ED MBH volume increased. These findings can help inform targeted interventions in high-risk populations to provide more equitable and safe care for this vulnerable ED population.

Our findings noted disparities in pharmacologic restraint use by race, with Black patients having a 22% increased odds of pharmacologic restraint use compared with white patients. This is consistent with previous studies reporting an association between Black race and physical restraint use in EDs in both pediatric and adult patients.^22–24,35  In addition, among visits with a SMI diagnosis grouping, Black and Asian patients had increased association with pharmacologic restraint when compared with visits with a non-SMI diagnosis grouping. Although disparities in care based on race have been identified in children in the ED with long bone fracture, sepsis, and appendicitis,^37–39  this racial disparity in children receiving pharmacologic restraint has not been well described. Among Black children, patient-level factors associated with pharmacologic restraint use were similar to those in the initial multivariable model including children of all races. This finding has significant implications as the health care system strives to identify and address disparities in health outcomes, which could be a result of explicit or implicit bias. Quality improvement interventions aimed at standardizing use of pharmacologic restraints have the potential to reduce racial disparities.³⁴  For example, using consensus guidelines for the evaluation and treatment of agitation, objective scores such as the Modified Overt Aggression Scale or a restraint checklist could aid in standardizing pharmacologic restraint use, although additional study on these tools in the ED setting is needed.^8,40–42 

Male sex and young adult age were associated with increased odds of pharmacologic restraint use, which is consistent with previous studies in pediatric and adult populations.^21,23,24  Young adults who continue to seek care at specialized children’s EDs rather than transitioning to adult centers may represent a group with severe chronic mental illness or developmental delays, which might contribute to the higher risk of restraint within our study’s population.⁴³  This is an important consideration given the dramatic increase in psychiatric ED visits among adolescents and young adults.⁷  Conversely, this may possibly represent a reallocation of resources toward a continuation of care for young adults within children’s hospitals. As ED utilization for MBH conditions continues to rise, EDs should consider developing restraint reduction and early deescalation strategies targeted at young adult and male patients. Training of staff in pediatric EDs, who may have less experience caring for young adults with agitation or aggressive behavior, will also be important to optimize the care and safety of these patients.

Repeat MBH ED visits were found to be associated with increased odds of pharmacologic restraint use. This is potentially a proxy for severity of psychiatric illness, as supported by previous literature describing an association between disease severity and repeat ED visits.⁴⁴  However, this association between repeat ED visits for MBH emergencies and restraint utilization also represents an important opportunity to improve care. Previous studies have shown that among certain populations, including those with diagnoses of autism spectrum or bipolar disorder, implementation of a daily schedule to encourage activity routines and sleep hygiene may be helpful in setting boundaries and expectations for behavior in the ED and promote mood stability.⁴⁵  Other recommended strategies include screening for agitation risk early within the ED stay, environmental room modification, engagement of psychiatry or behavioral health and child life specialists, and early engagement of caregivers.^8,25  In addition, an individualized deescalation plan established during an initial ED visit could be carried forward between hospital encounters. Individualized care plans using nonpharmacologic and pharmacologic strategies in a tiered or stepwise approach should be developed by the mental health team in collaboration with the patient and caregiver(s). This type of care plan may include factor such as a description of baseline behaviors, cues or warning signs of impending escalation, patient-specific deescalation techniques, favored activities, and recommended personal protective equipment to keep ED staff safe (ie, tie hair back, no dangling earrings).

Overnight and weekend arrival times were also identified as risk factors for pharmacologic restraint. Literature describing variability in ED care quality based on time and day of initial presentation has been mixed. Some studies report increased morbidity and mortality for patients admitted from the ED during weekends or overnight hours, whereas other studies note no difference based on time or day of the week.^46–48  However, these studies did not specifically examine care quality for MBH ED visits. Because of reduced staffing ratios during overnight and weekend hours, staff may be more limited in their ability to spend time engaging patients effectively with nonverbal deescalation techniques. Additionally, there may be fewer specialists available to assist in-house such as child life specialists and psychiatrists. Finally, there may be decreased tolerance for patients making noise and disturbing other patients during sleeping hours. Increasing resources during overnight hours may be considered as 1 strategy to reduce restraint use during this high-risk period.

An inverse relationship was found between average annual ED MBH volume and the proportion of visits with pharmacologic restraint use. EDs with higher MBH volumes may have more experience and personnel resources dedicated to this vulnerable population, including role groups such as child life specialists, behavioral health specialists, or social workers. There also may be more ED-specific physical resources such as “safe rooms” with potentially harmful objects removed, age-appropriate activities, daily schedules, and specific protocols to guide care. Finally, ED staff may have greater comfort and experience with management of children with MBH conditions, such as verbal and behavioral deescalation strategies, thereby minimizing restraint use. Further study of EDs with lower and higher restraint rates is vital to understand which of these factors contribute the most to variation in care and how successful strategies could be adapted for use in EDs with lower MBH volumes.

Our study has several limitations. First, as with any administrative database, diagnoses and demographic characteristics may be misclassified. Because diagnosis codes transitioned in 2015 from ICD-9 to ICD-10, the composition of the study population might have shifted, although in previously published literature using this database and definition criteria, we appreciated continuity of trend lines, which is reassuring.²¹  Additionally, patient disposition cannot be verified, and thus an admission to the hospital could represent medical admission or hospitalization while awaiting acceptance to an inpatient psychiatry unit. We assumed the indication for use of the medications was for restraint, based on the ED primary MBH diagnosis but could not discern if the medication was a previously prescribed (home) medication. Additionally, some of the medications captured by our pharmacological restraint variable may have been administered in the ED as part of a patient’s routine psychiatric care, and not specifically prescribed for a restraint indication. Thus, our estimates of the prevalence of pharmacological restraint may be artificially inflated. We did not include oral medication in our definition of pharmacologic restraint because these may be given as home medications or as treatment of MBH conditions rather than for the indication of restraint, which could have led to an underreporting of restraint use. Furthermore, medications were identified based on billing codes, meaning their use cannot be confirmed if medications were successfully administered to the patient after being ordered. Visit arrival data were used to determine day of week and time of day reported, which may not represent the time of pharmacologic restraint use. For patients who were subsequently admitted to the hospital, intravenous or intramuscular medication ordered in the first 2 days of care was attributed to delivery in the ED, although it is possible the medication was administered when the patient was hospitalized, and possibly for indications other than pharmacological restraint. Finally, because PHIS data are derived from a select group of children’s hospitals, these findings may be less generalizable to pediatric patients with MBH conditions evaluated in other hospitals.

In our study of patients presenting to the pediatric ED with MBH diagnoses, pharmacologic restraint use was associated with young adult age, male sex, Black race, visits beginning during overnight hours, visits originating on the weekend, and repeat MBH ED visits. Additionally, EDs with larger average annual MBH visit volumes had lower frequency of pharmacologic restraint use. Future research is needed to better understand variation and driving factors for disparities in pharmacologic restraint use among pediatric EDs. Implementation of standardized best practices decreases biases in care related to the management of agitation in children with MBH conditions.

CI

confidence interval

ED

emergency department

MBH

mental and behavioral health

OR

odds ratio

PHIS

Pediatric Health Information System

SMI

serious mental illness