Abstract
This cross-sectional study explores the use of physical restraint on youth at risk of harming themselves or others in the emergency department, stratified by race and ethnicity.
Black adult patients are more likely to be physically restrained in the emergency department (ED).1 For pediatric ED visits, physical restraint is an uncommon but serious intervention associated with injury, trauma, and death.2 Physical restraint is indicated for youth at risk of harming themselves or others, often due to acute exacerbation of a neuropsychiatric disorder. Physical restraint should only be used after trialing behavioral deescalation and/or chemical restraint.3 Limited literature has examined physical restraint of children in the ED.4
Our objective was to compare characteristics of pediatric ED visits with and without physical restraint use. We hypothesized that physical restraint use would be more common in ED visits for Black non-Hispanic compared with White non-Hispanic pediatric patients.
Methods
We conducted a cross-sectional study of pediatric visits (children aged 0-16 years) to 11 different EDs within a New England health care system between 2013 and 2020. This study followed STROBE reporting guidelines for observational studies. The Yale University Human Investigation Committee approved this study and waived the need for informed consent.
Our primary outcome was the proportion of ED visits with a physical restraint order. We reported visit characteristics with descriptive statistics. We examined the association between physical restraint and race and ethnicity using a generalized linear multivariable mixed model with a binary logistic link, incorporating nesting by patient, clinician, and site. Race and ethnicity were extracted from the electronic health record. We controlled for the following covariates: age, sex, language spoken, insurance, school month, behavioral health presenting problem, presence of an outpatient psychotropic medication, and history of behavioral health condition. Analyses were performed with R software version 3.6.3.
Results
Our sample included 551 740 pediatric ED visits. Of the 532 visits (0.1%) with an associated physical restraint order, 208 patients (39.1%) were Black, 200 (37.6%) were White, and 103 (19.4%) were Hispanic. A total of 432 (81.2%) were aged 10 to 16 years, 352 (66.2%) were male, and 379 (71.2%) were publicly insured. Of the 532 visits, 430 (80.8%) included a child presenting with a behavioral health problem, 393 (73.9%) had outpatient psychotropic medications, and 233 (43.8%) had a behavioral health diagnosis (Table 1).
Table 1. Demographic and Visit Characteristics of Pediatric ED Visits by Presence of a Physical Restraint Order in the EHR, January 2013-December 2020.
| Physical restraint use, No. (%)a | ||
|---|---|---|
| No (n = 551 208) | Yes (n = 532) | |
| Race and ethnicity | ||
| American Indian or Alaska Native | 1107 (0.2) | 0 |
| Asian | 13 140 (2.4) | 5 (0.9) |
| Black non-Hispanic | 124 509 (22.6) | 208 (39.1) |
| Hispanic or Latinx | 206 806 (37.5) | 103 (19.4) |
| Native Hawaiian or Other Pacific Islander | 2136 (0.4) | 0 |
| White non-Hispanic | 183 413 (33.3) | 200 (37.6) |
| Otherb | 20 097 (3.6) | 16 (3.0) |
| Age, y | ||
| 0-4 | 228 828 (41.5) | 15 (2.8) |
| 5-9 | 139 382 (25.3) | 85 (16.0) |
| 10-16 | 182 998 (33.2) | 432 (81.2) |
| Sex | ||
| Female | 257 563 (46.7) | 180 (33.8) |
| Male | 293 641 (53.3) | 352 (66.2) |
| Preferred language | ||
| English | 472 132 (85.7) | 495 (93.0) |
| Non-English | 79 076 (14.3) | 37 (7.0) |
| Insurance | ||
| Private | 179 275 (32.5) | 150 (28.2) |
| Public | 341 195 (61.9) | 379 (71.2) |
| Other | 30 738 (5.6) | 3 (0.6) |
| School monthc | ||
| No | 127 494 (23.1) | 131(24.6) |
| Yes | 423 714 (76.9) | 401 (75.4) |
| Behavioral health presenting problemd | ||
| No | 533 595 (96.8) | 102 (19.2) |
| Yes | 17 613 (3.2) | 430 (80.8) |
| Behavioral health outpatient medicationse | ||
| No | 502 798 (91.2) | 139 (26.1) |
| Yes | 48 410 (8.8) | 393 (73.9) |
| Behavioral health medical history f | ||
| No | 505 106 (91.6) | 299 (56.2) |
| Yes | 46 102 (8.4) | 233 (43.8) |
Abbreviations: ED, emergency department; EHR, electronic health record.
Percentages may not total 100% due to rounding.
The “other” race and ethnicity group includes those categorized as unknown/other or patient refusal to answer. A total of 1973 patients (0.36%) with missing race and ethnicity data were excluded from the sample; none of these patients had a restraint order.
School months included September through June.
Presenting problems were identified as “behavioral health” problems based on regular expression searches and consensus from manual review by 4 different authors.
Medications were classified as behavioral health medications according to groupings within the Anatomical Therapeutic Chemical (ATC) classification.
Diagnoses in the medical history were categorized as “behavioral health” diagnoses by mapping diagnoses to the AHRQ Clinical Classification System and manually grouping psychiatric and substance use diagnoses.
In adjusted regression analyses, visits for Black patients were more likely to use physical restraint compared with visits for White patients (adjusted odds ratio [AOR], 1.80 [95% CI, 1.40-2.32]). There was no difference in restraint use between visits for Hispanic and White patients. Visits for male vs female patients had higher odds of restraint (AOR, 1.95 [95% CI, 1.57-2.41]). Visits for patients using public vs private insurance had higher odds of restraint (AOR, 1.28 [95% CI, 1.01-1.62]). Visits for patients presenting with a behavioral health concern, a medical history of a behavioral health condition, and/or a psychotropic medication on their home medication list had increased odds of restraint use (Table 2).
Table 2. Percentage of Visits and Adjusted Odds of Receiving a Physical Restraint Order by Variable .
| Visits with a restraint order, %a | Adjusted odds ratio (95% CI)b | |
|---|---|---|
| Race and ethnicity | ||
| Black non-Hispanic | 0.17 | 1.80 (1.40-2.32) |
| Hispanic | 0.05 | 0.76 (0.56-1.02) |
| White non-Hispanic | 0.11 | 1 [Reference] |
| Otherc | 0.06 | 0.89 (0.53-1.49) |
| Age, y | ||
| 0-4 | 0.01 | 1 [Reference] |
| 5-9 | 0.06 | 3.23 (1.81-5.75) |
| 10-16 | 0.24 | 4.43 (2.52-7.77) |
| Sex | ||
| Female | 0.07 | 1 [Reference] |
| Male | 0.12 | 1.95 (1.57-2.41) |
| Preferred language | ||
| English | 0.10 | 1 [Reference] |
| Non-English | 0.05 | 1.17 (0.76-1.84) |
| Insurance | ||
| Private | 0.08 | 1 [Reference] |
| Public | 0.11 | 1.28 (1.01-1.62) |
| Other | 0.01 | 0.30 (0.09-0.96) |
| School month | ||
| No | 0.10 | 1 [Reference] |
| Yes | 0.09 | 0.84 (0.68-1.04) |
| Behavioral health presenting problem | ||
| No | 0.02 | 1 [Reference] |
| Yes | 2.44 | 24.7 (18.8-32.6) |
| Behavioral health outpatient medications | ||
| No | 0.03 | 1 [Reference] |
| Yes | 0.81 | 2.59 (2.02-3.33) |
| Behavioral health medical history | ||
| No | 0.06 | 1 [Reference] |
| Yes | 0.51 | 1.47 (1.19-1.83) |
Data are presented as percentage of patients. Percentages may not total 100% due to rounding.
P value of <.05 was considered statistically significant.
The other race categories reported in Table 1 were too small to individually analyze.
Discussion
Our study demonstrates that even after adjusting for sociodemographic and clinical characteristics, Black children are more likely to be physically restrained in the ED compared with White children.
Racial and ethnic disparities in restraint use likely reflect racism at multiple levels within and beyond the emergency care continuum (community, emergency medical services, and ED). Potential drivers include interpersonal racism in the perception of threat or aggression,5 institutional racism leading to disparities in access to behavioral health care, and structural racism leading to disproportionate experiences with risk factors for mental illness, including poverty, trauma, and discrimination.6
The use of cross-sectional data from the electronic health record limited our ability to control for clinical variation, such as trials of deescalation prior to physical restraint. The reliability of our race and ethnicity data is unknown.
Documenting racial and ethnic inequities is important but insufficient. Physical restraint is a traumatic experience that can have lasting consequences for a child’s development and well-being. Further research should identify interventions that address inequities in physical restraint use and ensure a health care system that is a therapeutic rather than traumatic environment.
References
- 1.Wong AH, Whitfill T, Ohuabunwa EC, et al. Association of race/ethnicity and other demographic characteristics with use of physical restraints in the emergency department. JAMA Netw Open. 2021;4(1):e2035241-e2035241. doi: 10.1001/jamanetworkopen.2020.35241 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Nunno MA, Holden MJ, Tollar A. Learning from tragedy: a survey of child and adolescent restraint fatalities. Child Abuse Negl. 2006;30(12):1333-1342. doi: 10.1016/j.chiabu.2006.02.015 [DOI] [PubMed] [Google Scholar]
- 3.Gerson R, Malas N, Feuer V, Silver GH, Prasad R, Mroczkowski MM. Best practices for evaluation and treatment of agitated children and adolescents (BETA) in the emergency department: consensus statement of the American Association for Emergency Psychiatry. West J Emerg Med. 2019;20(2):409-418. doi: 10.5811/westjem.2019.1.41344 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dorfman DH, Mehta SD. Restraint use for psychiatric patients in the pediatric emergency department. Pediatr Emerg Care. 2006;22(1):7-12. doi: 10.1097/01.pec.0000195758.12447.69 [DOI] [PubMed] [Google Scholar]
- 5.Todd AR, Thiem KC, Neel R. Does seeing faces of young black boys facilitate the identification of threatening stimuli? Psychol Sci. 2016;27(3):384-393. doi: 10.1177/0956797615624492 [DOI] [PubMed] [Google Scholar]
- 6.Trent M, Dooley DG, Dougé J; Section on Adolescent Health; Council on Community Pediatrics; Committee on Adolescence . The impact of racism on child and adolescent health. Pediatrics. 2019;144(2):e20191765. doi: 10.1542/peds.2019-1765 [DOI] [PubMed] [Google Scholar]