Demographic Patterns of Age- and Size-Appropriate Child Restraint Use in Chicago

Abstract

Objective:

Age alone is often used to define appropriate child restraint system (CRS) use. Children who lack access to primary care may rely on acute care settings and miss opportunities for anticipatory guidance on child passenger safety. We sought to characterize age- and size-appropriate CRS use among emergency department (ED) and urgent care patients in metropolitan Chicago to inform targets for intervention.

Methods:

We conducted a secondary cross-sectional analysis of clinical trial screening data collected from English and Spanish-speaking caregivers of 6-month through 10-year-old children, 1/2021–8/2022. Caregiver-reported transportation behaviors were linked to electronic health record (EHR) data. CRS use was categorized as appropriate or prematurely transitioned for age and size. Descriptive statistics were calculated. Chi-square and logistic regression analyses were used to test for child, family, and visit characteristics associated with premature transitions.

Results:

Of the 4,269 caregivers with matched screening and EHR data, 4,045 (94.8%) were included and 933 (23.1%) had prematurely transitioned their child. Premature transitions were most associated with child age 2–4 or 8+, Black race or Hispanic/Latine ethnicity, and being seen in the ED. Premature transitions differed across child age groups and were most common among Hispanic/Latine and Black children (age <8), families with larger household size (age 8+), and children seen in the ED (age 2–4).

Conclusion:

Premature transitions are common among children who received acute care in metropolitan Chicago. This population is at risk for crash-related injuries. Child passenger safety interventions targeted to families who seek acute care may benefit specific demographic groups.

Introduction

The 2011 update of the American Academy of Pediatrics’ (AAP) Child Passenger Safety Policy Statement shifted away from age-based transition points to recommend children use car seats and booster seats as long as possible and transition when they reach the manufacturer-specified height or weight limit of the current car seat or booster seat.¹ Extended use of rear-facing and forward-facing car seats has been made possible because of engineering advances in child restraint system (CRS) design.^1,2 Correct and consistent CRS use is promoted through anticipatory guidance at well child visits^3,4 and enforcement of state laws that require restraints specific to a child’s age, and increasingly child size.^5,6

Increased CRS use^7,8 coupled with improvements in vehicle and trafficway engineering, has reduced the number of fatalities and injuries among children involved in motor vehicle collisions (MVCs).⁹ Despite this progress, MVCs remain a leading cause of death among children younger than 10 in the United States (US).^10,11 Suboptimal CRS use remains common^12,13 and is a major contributor to hospitalization and mortality after an MVC.^14,15 Factors associated with suboptimal CRS use, such as a premature transition to a less protective restraint, include socioeconomic status (SES), race and ethnicity, child’s age, and driver’s sex.^13,16 Premature transitions, frequently to a booster seat or seat belt, are associated with older child age, other children being in the vehicle, driver/parent being younger, driver/parent themselves not restrained, and Black race.^17,18 Lower SES additionally correlates with restraint nonuse and premature transition.¹⁹

Much prior research has categorized CRS appropriateness based on child age. Few studies have used weight and height parameters, often using older CRS size limits (e.g., up to 20 pounds for rear-facing). In this secondary cross-sectional analysis of data collected from caregivers of children who received emergency and urgent care in metropolitan Chicago and their child’s size data from the electronic health record (EHR), we sought to describe child passenger safety behaviors and examine child, family, and visit characteristics associated with premature transitions. Based on prior literature, we hypothesized that premature transitions would be associated with lower family income, lower educational attainment, minoritized racial and ethnic groups, and older child age.

Methods

Data collection and extraction

Data were collected as part of the screening process for recruitment into the Tiny Cargo, Big Deal! clinical trial (NCT04238247).²⁰ In brief, caregivers were invited to complete a screening survey during or after their child’s evaluation in the emergency department (ED) or two urgent care centers that are part of Ann & Robert H. Lurie Children’s Hospital of Chicago. The screening survey was built on the Qualtrics^XM platform and distributed using QR codes posted in examination rooms and personalized invitations sent via email, text messages, and patient portal messages between June 2020 and August 2022. Patient lists from ED and urgent care encounters were extracted monthly from the EHR. This study focuses on screening surveys completed January 2021 through August 2022, the period during which demographic questions were asked in the screener. Caregivers reviewed an electronic information sheet and agreed to participate before entering the survey. English or Spanish-speaking caregivers were eligible for the screening if they were at least 18 years old and an Illinois resident, with a child 6 months through 10 years old and shorter than 55 inches tall. We did not reach out to families of children who were critically ill or injured (e.g., Emergency Severity Index triage category 1), evaluated after an MVC, or evaluated for suspected child abuse. Caregivers screened at Central DuPage Hospital, an affiliated suburban, community ED were excluded from these analyses because we did not have access to EHR data from that location. There was no incentive for completing the screening survey. This study was approved by the Institutional Review Board at Lurie Children’s (IRB# 2020–3274).

The screening survey collected information about the child’s travel frequency, seating location (front or back seat), usual CRS, if any, and other CRSs used by the child, with sketched images to assist caregivers in selecting a response that represents their child’s CRS (Appendix Figure 1). Additionally, caregivers provided their age (18–24, 25–34, 35–44, 45+ years), highest level of education (less than high school, high school, Associate’s/some college, Bachelor’s or above), household size (2–4, 5–7, 8+ people), annual household income (below $50,000, $50,000-$99,000, $100,000+), whether any household members were looking for work, and their best estimate of the child’s weight and height. Screening survey data were linked to clinical data from the EHR using a unique study identifier (first three letters of child’s first name, first three letters of child’s last name, and 4-digit year of birth), caregiver contact information, and date of the child’s acute care visit or invitation to participate. Variables extracted from the EHR included child age, weight, height (if available), gender, race, ethnicity and preferred language for the family (as entered in the EHR by hospital staff), and visit characteristics including visit location (ED or urgent care), ESI triage level (2–5), and disposition (discharged home, left not seen, admitted [including inpatient, operating room, psychiatry], or not applicable due to being seen by urgent care). Size information was obtained from the child’s most recent clinical encounter in the Lurie Children’s EHR, which may have been a primary or specialty care visit. When height was not recorded in the EHR, we used caregiver-reported height. We excluded cases where the weight or height was outside of anthropometric norms (<3^rd percentile or >97^th percentile) based on Centers for Disease Control and Prevention’s length-for-age and weight-for-age growth curves.²¹

Identification of premature CRS transition

Table 1 provides parameters by which the caregiver-reported CRS was classified as appropriate, delayed transition, or prematurely transitioned. Classification was based on the AAP Guidelines for Child Passenger Safety Technical Report,¹ child age, weight, height, and the minimum and maximum weight and height limits for typical CRSs on the US market, as listed on the 2024 AAP healthychildren.org website.²² In cases where weight and/or height were unknown, CRS use was categorized using the available size data plus age or age alone (i.e., <2 years old in rear-facing CRS, 2–4 years old in forward-facing CRS, and 5 years and older in booster seats).

Table 1.

Parameters for Classification of Child Restraint System Use

Child Restraint System	Classification*	Age	Weight	Condition	Height
Infant Carrier Rear-Facing Only	Premature Transition		N/A		N/A
	Appropriate		<35 pounds	AND	<32 inches
	Delayed Transition		>=35 pounds	OR	>=32 inches
Rear-Facing Convertible or All in One	Premature Transition		N/A		N/A
	Appropriate		<40 pounds	AND	<40 inches
	Delayed Transition		>=40 pounds	OR	>=40 inches
Forward-Facing with Harness	Premature Transition	Any age	<22 pounds	OR	<28 inches
		<3 years	<40 pounds	AND	<40 inches
	Appropriate	3 years	>=22 pounds and <65 pounds	AND	>=28 inches and <49 inches
	Delayed Transition	Any age	>=65 pounds		>=49 inches
High Back Booster	Premature Transition	Any age	<40 pounds	OR	<38 inches
		<5 years	<65 pounds	AND	<49 inches
	Appropriate	5 years	>=40 pounds and <100 pounds	AND	>=38 inches and <57 inches
	Delayed Transition	Any age	>=100 pounds	OR	>=57 inches
Backless Booster	Premature Transition	Any age	<40 pounds	OR	<43 inches
		<5 years	<65 pounds	AND	<49 inches
	Appropriate	5 years	>=40 pounds and <120 pounds	AND	>=43 inches and <57 inches
	Delayed Transition	Any age	>=120 pounds	OR	>=57 inches
Seat Belt	Premature Transition	Any age	<100 pounds	AND	<57 inches
	Appropriate		>=120 pounds	OR	>=57 inches

^*NOTE: Each child’s CRS was classified into one of three groups according to the child’s age, weight, height, and the most common manufacturer-specified minimum and maximum weight and height limits for CRS types available on the US market. Premature transition: the child does not meet the minimum size requirements for the CRS type (top row). Delayed transition: the child is still under the weight OR height maximum for a more protective CRS type (bottom row).

Variables and analyses conducted

For children using multiple CRS, we used the caregiver-designated “usual” CRS for the main analysis, with their least protective restraint used in a sensitivity analysis. Only 7% of children were reported to be using multiple types of restraints. The child’s race, ethnicity, and caregiver’s preferred language were obtained from the EHR and not collected with screening survey responses to minimize response burden. Caregivers could toggle between English and Spanish versions of the survey based on their language preference. We examined associations between premature transitions and child race and ethnicity to place our findings in the context of prior literature.^23,24 The “other” race category contained American Indian/Alaska Native children and those with other or no race indicated in the EHR (Appendix Table 1). Those whose language preference was not listed in the EHR were considered to prefer English.

Child age was categorized in the following groups: <2 years, 2-<5 years, 5-<8 years, 8+ years. The caregiver education levels, less than high school and high school, and household size levels, 5–7 and 8+, were collapsed due to small sample sizes; similarly, the child’s visit disposition was collapsed into two groups: admitted or discharged/ left not seen/not applicable.

The primary outcome for analysis was premature transition, defined as a child smaller or younger than appropriate for the CRS they were using. All statistical analyses were conducted in RStudio version 4.3.1. Figures were generated to display appropriate CRS use across child height, weight, and age. Chi-square tests and multivariable logistic regression analyses, stratified by age group, were conducted to assess factors associated with premature transition. Due to the primary focus of this study being premature transitions, children with delayed transitions and those unrestrained were excluded from statistical analysis. Results are reported in accordance with the STROBE checklist for observational studies.

Results

Overall demographics

Participant flow is presented in Figure 1. Screening surveys were distributed after 42,240 ED and urgent care visits, to caregivers of children 6 months to <2 years (n=13,460; 31.9%), 2 to <5 years (n=15,117; 35.8%), 5 to <8 years (n=9,233; 21.9%), and 8 years and older (n=4,430; 10.5%) (Table 2). In the population invited to screen, there were more boys (n=22,809; 54.0%) than girls (n=19,429; 46.0%). Children were Asian Non-Hispanic/Latine (n=1,748; 4.1%), Black Non-Hispanic/Latine (n=5,671; 13.4%), Hispanic/Latine (n=18,902; 44.7%), white Non-Hispanic/Latine (n=11,190; 26.5%), multiple races (n=347; 0.8%), and other racial groups (n=4,382; 10.4%). Spanish language preference was documented in the EHR for 5,259 (12.5%).

Figure 1.

Flow of Observations for Analysis of Premature Transition per Usual Child Restraint System

Table 2.

Analytic Sample and Invited Population Characteristics

Variable	Sample N (%) Total N: 4,045	Invited N (%) Total N: 42,240
Child Sex (as recorded in the EHR)
Female	1,802 (44.5%)	19,429 (46.0%)
Male	2,243 (55.5%)	22,809 (54.0%)
Caregiver Language (as recorded in the EHR)
English-Language preference	3,906 (96.6%)	36,960 (87.5%)
Spanish-Language preference	248 (6.1%)	5,259 (12.5%)
Unknown	109 (2.7%)	121 (0.3%)
Child Race/Ethnicity (as recorded in the EHR)
Asian, Non-Hispanic/Latine	157 (3.9%)	1,748 (4.1%)
Black, Non-Hispanic/Latine	357 (8.8%)	5,671 (13.4%)
Hispanic/Latine, any racial category	1,138 (28.1%)	18,902 (44.7%)
Multiple races, Non-Hispanic/Latine	58 (1.4%)	347 (0.8%)
Other* or Unknown, Non-Hispanic/Latine	516 (12.8%)	4,382 (10.4%)
White, Non-Hispanic/Latine	1,891 (46.7%)	11,190 (26.5%)
Child Age
<2 yr	1,428 (35.3%)	13,460 (31.9%)
2 to <5 yr	1,509 (37.3%)	15,117 (35.8%)
5 to <8 yr	786 (19.4%)	9,233 (21.9%)
8+ yr	322 (8.0%)	4,430 (10.5%)
Child Weight (pounds)** from their most recent ED or other clinical encounter
Median (Interquartile Range)	33.1 (24.3–44.1)	34.5 (25.7–48.4)
Not available	57 (1.2%)	3,288 (7.8%)
Child Height (inches)** from their most recent primary or specialty care visit or caregiver report
Median (IQR)	37.0 (32.0–60.0)	39.0 (32.3–45.7)
Not available	1,461 (36.1%)	36,602 (86.7%)
CRS (usual)		Not available
Rear-Facing Only Infant Carrier	429 (10.6%)
Rear-Facing CRS	1,038 (25.7%)
Forward-Facing CRS with 5-point harness	1,570 (38.8%)
High Back Booster Seat	492 (12.2%)
Backless Booster Seat	379 (9.4%)
Seat Belt Only	137 (3.4%)
CRS (least protective used)		Not available
Rear-Facing Only Infant Carrier	367 (9.1%)
Rear-Facing CRS	1,056 (26.1%)
Forward-Facing CRS with 5-point harness	1,532 (37.9%)
High Back Booster Seat	471 (11.6%)
Backless Booster Seat	443 (11.0%)
Seat Belt Only	176 (4.4%)
Caregiver Age		Not available
18–24 yr	146 (3.6%)
25–34 yr	1,358 (33.6%)
35–44 yr	2,243 (55.5%)
45+ yr	298 (7.4%)
Household Size		Not available
2–4 people	3,026 (74.8%)
5+ people	1,019 (25.2%)
Caregiver Education		Not available
High school graduate or below	626 (15.5%)
Associate’s degree or some college	622 (15.4%)
Bachelor’s, graduate, or professional school	2,797 (69.1%)
Annual Household Income		Not available
<$50k	1,082 (26.7%)
$50–99k	589 (14.6%)
$100k+	2,374 (58.7%)
Someone in Household Looking for Work		Not available
No	3,262 (80.6%)
Yes	783 (19.4%)

^*Other includes: American Indian/Alaska Native, and other or unknown race or ethnicity.

^**57 (1.2%) missing weight, 1,461 (36.1%) missing height

Of 5,621 screening surveys completed with 90% or more questions answered, 4,951 screening responses could be matched to EHR data, with 4,269 screening surveys completed after the addition of family demographic questions in January 2021. We excluded 224 (5.2%) from analyses (Figure 1). Within the analytic sample, 2,659 (65.7%) children had complete data for age, height, and weight; 1,437 (35.5%) had age and weight data; 26 (0.6%) had age and height data; and 33 (0.8%) were evaluated by age alone (Figure 1).

The characteristics of children in the study sample were similar to the population invited to screen in terms of child sex (55.5% male, n=2,243) and age (35.3% <2 years, n=1,428, and 37.3% 2 to <5 years, n=1,509) and differed on caregiver preferred language, race, and ethnicity (Table 2). Most children were listed as being non-Hispanic/Latine, white (96.6%, n=3,906), and had English as the caregiver’s preferred language (46.7%, n=1,891) in the EHR. Most screening respondents were 25–34 years (n=1,358; 33.6%) or 35–44 years (n=2,243; 55.5%) with a Bachelor’s degree or graduate/professional education (n=2,797; 69.1%), in households of 2–4 people (n=3,026; 74.8%) where no one was looking for work (n=3,262; 80.6%).

Premature CRS transitions

The usual CRSs represented in the sample were 429 (10.6%) infant carriers, 1,038 (25.7%) rear-facing convertible CRS, 1,570 (38.8%) forward-facing CRS with a 5-point harness, 492 (12.2%) high-back booster, 379 (9.4%) backless booster, and 137 (3.4%) seat belt alone (Table 2). Overall, 933 (23.1%) were categorized as prematurely transitioned (Figure 2), with variation across CRS mode: 39.2% (n=615) forward-facing car seat, 27.0% (n=133) high-back booster, 13.2% (n=50) backless booster, and 98.5% (n=135) seat belt. Premature transition was not relevant to rear-facing seats. We found slightly more children (n=1,043; 25.8%) were categorized as prematurely transitioned in analyses of the caregiver-reported least protective CRS instead of the usual CRS.

Figure 2.

Age and Weight Distribution by Appropriateness versus Premature Transition across Child Restraint Systems*

*Black box and whiskers show the ages and weights of children considered appropriate within each child restraint system. Red box and whiskers show the ages and weights of children considered prematurely transitioned. The central vertical line indicates the median, upper and lower bounds of the boxes indicate interquartile range, the whiskers extend to 1.5 times the interquartile range in either direction from the median, and the dots indicate outlier observations (i.e., more than 1.5 times outside of the median).

Abbreviations: HB: high back; NB: no back/backless

Factors associated with premature transitions

Child race/ethnicity, caregiver age, household size, caregiver education, annual household income, caregiver employment status, child age, and usual CRS were all associated with premature transitions (Table 3). Children aged 2–4 or 8+, who were Hispanic/Latine or Black, non-Hispanic/Latine were most likely to be prematurely transitioned (Table 4). Presenting to the ED vs. a satellite urgent care location (OR 1.23, 95% CI 1.00–1.52) was associated with premature transition. Caregiver age 35–44 was protective (OR 0.65, 95% CI 0.44–0.98) against premature transition. In chi-square analysis, caregiver language preference was additionally associated with premature transitions (Table 3, p=0.0052).

Table 3.

Child, Clinical, and Family Characteristics Associated with Premature Transitions

	Appropriate Total N: 3,112		Premature Total N: 933
Variable	n	Row %	n	Row %	p-value
Child Sex (as recorded in the EHR)					0.36
Female	1,399	77.6	403	22.4
Male	1,713	76.4	530	23.6
Caregiver Language (as recorded in the EHR)					0.0052
English-Language preference	2,946	77.4	860	22.6
Spanish-Language preference	165	69.3	73	30.7
Child Race/Ethnicity (as recorded in the EHR)					<0.001
Asian, Non-Hispanic/Latine	120	78.9	32	21.1
Black, Non-Hispanic/Latine	216	61.7	134	38.3
Hispanic/Latine, any racial category	805	70.7	333	29.3
Multiple races, Non-Hispanic/Latine	51	87.9	7	12.1
Other* or Unknown, Non-Hispanic/Latine	1,521	93.3	109	6.7
White, Non-Hispanic/Latine	1,517	82.7	318	17.3
Child Age					<0.001
<2 yr	1,232	86.3	196	13.7
2 to <5 yr	959	63.6	550	36.4
5 to <8 yr	695	88.4	91	11.6
8+ yr	226	70.2	96	29.8
CRS (usual)					<0.001
Rear-Facing Only Infant Carrier	429	100	N/A	0
Rear-Facing CRS	1,038	100	N/A	0
Forward-Facing CRS with 5-point harness	955	60.8	615	39.2
High Back Booster Seat	359	73.0	133	27.0
Backless Booster Seat	329	86.8	50	13.2
Seat Belt Only	2	1.5	135	98.5
Disposition					<0.001
Home	1,620	75.5	526	24.5
Admitted	358	74.0	126	26.0
Not applicable after Urgent Care	1,134	80.1	281	19.9
Location					0.004
Urgent Care	690	81.3	159	18.7
Emergency Department	2,422	75.8	774	24.2
Caregiver Age					<0.001
18–24 yr	89	61.0	57	39.0
25–34 yr	1,009	74.3	349	25.7
35–44 yr	1,791	79.8	452	20.2
45+ yr	223	74.8	75	25.2
Household Size					0.003
2–4 people	2,363	78.1	663	21.9
5+ people	749	73.5	270	26.5
Caregiver Education					<0.001
High school graduate or below	423	67.6	203	32.4
Associate’s degree or some college	427	68.6	195	31.4
Bachelor’s, graduate, or professional school	2,262	69.3	535	30.7
Annual Household Income					<0.001
<$50k	728	67.3	354	32.7
$50–99k	446	75.7	143	24.3
$100k+	1,938	81.6	436	18.4
Someone in Household Looking for Work					<0.001
No	2,555	78.3	707	21.7
Yes	557	71.1	226	28.9

^*Other includes: American Indian/Alaska Native, and other or unknown race or ethnicity.

Table 4.

Predictors of Premature Transitions, Full Sample

	Adjusted OR (95% CI)
Child Sex
Female	Ref
Male	1.10 (0.94–1.28)
Caregiver Language (as recorded in the EHR)
English-preferred	Ref
Spanish-preferred	0.96 (0.68–1.34)
Child Race/Ethnicity (as recorded in the EHR)
Asian, Non-Hispanic/Latine	1.18 (0.76–1.79)
Black, Non-Hispanic/Latine	2.23 (1.66–3.00)
Hispanic/Latine, any racial category	1.46 (1.14–1.86)
Multiple races, Non-Hispanic/Latine	0.57 (0.23–1.21)
Other or Unknown, Non-Hispanic/Latine	1.45 (1.11–1.88)
White, Non-Hispanic/Latine	Ref
Child Age group
<2 yr	Ref
2–<5 yr	3.80 (3.14–4.60)
5–<8 yr	0.80 (0.60–1.05)
8+ yr	2.63 (1.92–3.58)
Caregiver Age (years)
18–24 yr	Ref
25–34 yr	0.78 (0.53–1.25)
35–44 yr	0.65 (0.44–0.98)
45+ yr	0.77 (0.47–1.26)
Household size
2–4 people	Ref
5+ people	1.18 (0.99–1.41)
Household work status
No one looking for work	Ref
Looking for work	1.05 (0.85–1.28)
Annual household income
<$50k	Ref
$50k–99k	0.81 (0.62–1.07)
$100k+	0.80 (0.59–1.09)
Caregiver Education
High School Graduate or below	Ref
Associate’s degree/some college	0.99 (0.76–1.30)
Bachelor’s degree or above	0.83 (0.61–1.12)
Location
Urgent Care	Ref
Emergency Department	1.23 (1.00–1.52)

^*note: bold indicates statistical significance, p<0.05

In analyses stratified by age group, we found race and ethnicity were most predictive of premature transition for children under eight, with higher odds for Black (<2 year olds), Black or other (2 to <5 year olds), and Black or Hispanic/Latine (5 to <8 year olds), respectively (Table 5). Children 2 to <5 years old seen in the ED had higher odds of premature transition than those screened after an urgent care visit. Children 8 years old or older who lived in households of 5 or more individuals had higher odds of premature transition.

Table 5.

Predictors of Premature Transitions, Stratified by Child Age Group

	<2 yr	2–<5 yr	5–<8 yr	8+ yr
	Adjusted OR (95% CI)	Adjusted OR (95% CI)	Adjusted OR (95% CI)	Adjusted OR (95% CI)
Child Sex (as recorded in the EHR)
Female	Ref	Ref	Ref	Ref
Male	0.91 (0.59–1.12)	1.20 (0.96–1.49)	1.34 (0.85–2.15)	1.17 (0.70–1.96)
Caregiver Language ((as recorded in the EHR)
English-preferred	Ref	Ref	Ref	Ref
Spanish-preferred	1.32 (0.69–2.48)	1.13 (0.68–1.89)	0.40 (0.17–1.08)	0.61 (0.03–6.83)
Child Race/Ethnicity (as recorded in the EHR)
Asian, Non-Hispanic/Latine	1.76 (0.71–3.89)	1.06 (0.58–1.87)	2.28 (0.49–7.69)	0.42 (0.06–1.72)
Black, Non-Hispanic/Latine	2.94 (1.56–5.47)	1.80 (1.18–2.75)	4.80 (2.20–10.43)	1.80 (0.66–4.76)
Hispanic/Latine, any race	1.87 (1.11–3.15)	1.08 (0.77–1.51)	3.18 (1.60–6.34)	1.93 (0.90–4.18)
Multiple races, Non-Hispanic/Latine	0.61 (0.03–3.09)	0.60 (0.19–1.58)	NA	NA
Other or Unknown, Non-Hispanic/Latine	1.66 (1.00–2.74)	1.47 (1.02–2.11)	1.33 (0.43–3.40)	1.31 (0.54–3.03)
White, Non-Hispanic/Latine	Ref	Ref	Ref	Ref
Caregiver Age
18–24 yr	Ref	Ref	Ref	Ref
25–34 yr	0.80 (0.46–1.41)	0.97 (0.56–1.70)	Not estimable	NA
35–44 yr	0.67 (0.36–1.24)	0.72 (0.41–1.28)	Not estimable	1.12 (0.52–2.47)
45+ yr	0.86 (0.22–2.89)	0.69 (0.34–1.39)	Not estimable	1.30 (0.54–3.17)
Household size
2–4 people	Ref	Ref	Ref	Ref
5+ people	1.00 (0.67–1.47)	1.07 (0.84–1.38)	1.16 (0.71–1.87)	1.71 (1.01–2.91)
Household work status
No one looking for work	Ref	Ref	Ref	Ref
Looking for work	1.29 (0.87–1.89)	0.99 (0.73–1.33)	0.61 (0.33–1.08)	1.53 (0.78–2.95)
Annual household income
<$50k	Ref	Ref	Ref	Ref
$50k–99k	0.81 (0.47–1.39)	0.74 (0.50–1.10)	0.96 (0.48–1.91)	1.17 (0.45–2.99)
$100k+	0.37 (0.20–0.69)	1.00 (0.64–1.55)		2.45 (0.87–7.38)
Caregiver Education
High School Graduate or below	Ref	Ref	Ref	Ref
Associate’s degree/some college	0.82 (0.50–1.35)	0.85 (0.57–1.27)	1.61 (0.79–3.33)	1.54 (0.69–3.45)
Bachelor’s degree or above	0.99 (0.56–1.78)	0.66 (0.42–1.04)	1.46 (0.66–3.28)	0.49 (0.18–1.33)
Location
Urgent Care	Ref	Ref	Ref	Ref
Emergency Department	1.20 (0.76–1.97)	1.36 (1.04–1.80)	1.01 (0.57–1.87)	0.88 (0.45–1.76)

^*note: bold indicates statistical significance, p<0.05

Discussion

We found an overall CRS premature transition rate near 25%, with higher rates within certain age and demographic groups. Our approach is novel in using modern weight and height limits for typical CRSs on the US market and more nuanced in determining premature transitions by combining weight and height with age. With roughly one-third of 2- to 4-year-olds and 8- to 10-year-olds in our sample prematurely transitioned, our results identify new age groups of focus for interventions to promote size-appropriate CRS selection and use. By considering premature transitions to forward-facing CRS and including children from 6 months through 10 years, we expand upon prior premature transition studies that have largely focused on the premature use of booster seats and seat belts and children in the traditional booster seat age range of 4 to 8 years old.^{12,17,18,25–27} Consistent with other studies,^14,19 we found that the proportion of children prematurely transitioned varied across caregiver education and household income levels, caregiver employment status, household size, and child race and ethnicity.

In addition to variation in premature transitions between age and demographic groups, our findings shed new light on differences in premature transitions relative to care locations. The proportion of children prematurely transitioned was higher among the group who were seen at the main hospital ED location than the group seen in the affiliated urgent care centers. While the ED environment can be hectic, with clinicians, nurses, and social workers who are focused on the evaluation and treatment of children with acute needs, there are opportunities to innovate parallel processes that leverage digital technologies for child passenger safety screening, education, and referral to resources, during or after the encounter.²⁸ For children who rely on the ED for their healthcare needs, acute care encounters can fill critical voids in primary injury prevention. An acute care visit may also serve as a teachable moment that complements and reinforces injury prevention information shared in primary care settings. Local differences in proportions of premature transitions between ED and urgent care may reflect the location of the urgent care sites in more affluent urban and suburban areas. Other ED- and hospital-based studies have conducted educational interventions around CRS, finding that more than 50% of children were not following AAP recommendations and that caregiver knowledge and awareness increased after screening or intervention.^29–31

Premature transitions were more common among older children and larger family sizes. This may reflect the prioritization of larger, more expensive rear-facing and forward-facing CRSs as hand-me-downs for younger siblings. In addition, caregivers perceive their child as being a “big kid” or “too grown up” for a CRS, or they and their older children may prefer the convenience of using a booster seat or seat belt alone.³²

Our study is unique in that we inquired about all the ways a child is restrained when traveling, not just the restraint that was in use at the time of data collection. Premature transitions were slightly more prevalent when considering the least protective restraint rather than the child’s usual CRS alone. This finding emphasizes the importance of inquiring about all the ways a child may be restrained when providing child passenger safety counseling. More work is needed to explore the factors underlying the use of multiple CRS modes. The use of a less protective CRS when the family has multiple CRS modes available may relate to traveling with multiple caregivers or drivers (e.g., grandparents, babysitters) or challenges using CRS across different circumstances (e.g., seeing friends who have transitioned to a seat belt may reduce booster seat use when carpooling).^33,34 Other literature has found that caregivers allow children to sit in the front seat as a “treat”.^19,35

Taken together, these results reveal child characteristics available in the EHR that could be incorporated into the development of precision prevention strategies that incorporate screening for child passenger safety behaviors in acute care settings. One example of such an approach would be to send an automatic EHR-based message that contains a size-specific recommendation for CRS type. A second example would be to use two-way text messaging to ask if a family feels confident in installing their child’s CRS, and if they do not, automatically respond with a referral to a local child passenger safety technician. Such efforts could be integrated with screening for other social needs that have been promoted in ED settings in recent years.^36–38

There are several limitations of our study. First, as with many surveys, our findings are subject to participation bias. Only 13% of those invited to screen did so. This response rate is similar to other ED-based injury prevention studies.³⁹ COVID-19 pandemic restrictions on in-person activities prevented our plan to complete baseline data collection during the ED visit. Our budget did not include incentives for screening. We relied primarily on after-visit communication, and our response rate denominator was inflated by an unspecified number of families who did not receive or open the invitations sent via email or text message. We know 1,428 messages were flagged as having an email that “bounced”, an invalid phone number, or the family selected “opt out”, but we do not have a complete picture of emails that went to spam filters and the number of families who considered text messages with hyperlinks from unknown numbers to be spam or a security risk. Respondents disproportionately had English listed as their preferred language in the child’s EHR, and race or ethnicity listed as white or other, as compared with the overall population invited to complete a screening survey. Responses were also higher among caregivers whose children were younger than 5 years, an age group for whom child passenger safety may feel more relevant, and children who had height data in the EHR, which may be an indicator for primary care in the Lurie Children’s system and greater familiarity with Lurie Children’s. Second, there is potential for CRS misclassification bias, which may contribute to over- or under-estimation of premature transitions. For example, words and pictures in the survey may have been interpreted differently by different caregivers. Next, because we did not collect specific information on the make and model of each CRS, we used typical weight and height limits for CRSs on the US market and may have overestimated premature transitions. Some children will have appropriately stopped using a CRS with a lower weight limit, and our approach categorized them as having been prematurely transitioned. Additionally, a child’s height, as reported by a caregiver and, less likely, weight and height as recorded in the EHR, may be subject to error. We cannot estimate the magnitude or direction of this measurement bias, but we did attempt to mitigate size measurement errors by excluding 37 cases with extreme height or weight values per aforementioned CDC growth curves. We acknowledge that some of these measurements, particularly for weight, could have been accurate. Children evaluated by age or weight alone may have been incorrectly classified as appropriate when they were too tall for the CRS they were using. Additionally, a child’s length or standing height might not reflect their fit in a CRS; children with longer torsos are prone to outgrow a CRS sooner than those who are long in the legs, as fit is determined by where the top of the child’s head/ears is relative to the top of the CRS shell/headrest. Third, social desirability bias may have affected a caregiver’s willingness to accurately report on the CRS they are using, and social norms may vary by age group or cultural background. Fourth, data were collected during the COVID-19 pandemic, which may have acutely impacted household income and employment status. Finally, findings from this study population, screened for child passenger safety behaviors after an acute care visit occurring in metropolitan Chicago, may not be generalizable to other dissimilar settings.

Conclusion

Premature transitions to less protective CRS are common among children receiving acute care in metropolitan Chicago, differing across child age groups and most frequent among Hispanic/Latine children and Black children, children in larger families, and children seen in the emergency department. The ED, with a higher proportion of premature transition than observed for children seen in two urgent care settings, may provide an important location for child passenger safety behavior screening, reinforcement of safety information traditionally delivered in primary care settings, initiation of interventions that utilize community-based resources, and referrals to child passenger safety technicians.

Supplementary Material

1

Appendix Figure 1. Visuals used in CRS questions

What’s New:

We developed criteria to categorize caregiver-reported child restraint system (CRS) use as appropriate or prematurely transitioned based on child age, weight, height, and typical CRS size limits. Nearly a quarter of children in our acute care sample were prematurely transitioned.

Abbreviations:

AAP

American Academy of Pediatrics

CRS

child restraint systems

yr

year

k

thousand

EHR

Electronic Health Record

HB

high back [booster]

NB

no back/backless [booster]