Consumer usability of flexible vs. rigid lower anchor hardware in infant child restraint system (CRS) installations

Abstract

Objectives

Child restraint systems (CRS) can be installed using the Lower Anchors and Tethers for CHildren (LATCH) system. Two different types of lower anchors (LA) are available on CRS in the United States: (1) traditional flexible webbing LA, or (2) rigid LA, which is similar to the European ISOFIX system. The objective of this study is to evaluate consumer usability of flexible vs. rigid LA on infant CRS bases.

Methods

Thirty adult volunteers were recruited to install one infant CRS base with flexible LA and one infant CRS base with rigid LA into two different vehicles apiece, for a total of four installations per participant. One vehicle had easily accessible LA hardware and the other vehicle had less accessible LA hardware. A Child Passenger Safety Technician (CPST) inspected each installation and recorded any misuse or errors. Participants were asked to share their opinions and preferences on the two types of LA via written survey and verbal interview. Outcome variables include the number of errors committed, consumer preference, and ease of use (measured by time spent on each installation and participant feedback) of each LA system.

Results

Rigid LA installations resulted in higher rates of correct installation compared to flexible LA (72% correct vs. 47% correct, χ2 = 7.761, p = 0.0053). Participants reported better physical ease-of-use and higher confidence with rigid LA compared to flexible (Likert scale with Wilcoxon signed rank test, two-tailed, for paired data: p = 0.0326 and p = 0.0023, respectively). Rigid LA was faster to install during users’ second installation of each CRS compared to flexible (p = 0.0007) but was not faster during the first installation (p = 0.1165, matched-pair two-tailed t-tests). This indicates that each CRS required about the same amount of time for participants to familiarize themselves with the system, but once familiar, the rigid LA system was faster to install. Overall, 24/30 participants (80%) preferred rigid LA to flexible. Self-reported sociodemographic factors suggest that the participant sample was skewed toward highly-educated, white, native English speakers.

Conclusions

The rigid LA system was preferred by this group of consumers for installation of the base of infant CRS. The lower rate of installation errors suggests that rigid LA may help reduce the reported high rates of incorrect LA installation in vehicles in the US.

Keywords:

• Child restraint system
• CRS
• usability
• lower anchors
• LATCH

Introduction

Child restraint systems (CRS) can be installed using the Lower Anchors and Tethers for CHildren (LATCH) system. Two different types of lower anchors (LA) are available on CRS in the United States (US): (1) traditional flexible webbing LA, or (2) rigid LA, which is similar to the European ISOFIX system. Currently, most CRS on the US market are equipped with flexible webbing LA. Results of observational studies suggest that 40–56% of all types of LA installations are incorrect (Safe Kids USA 2011; Greenwell 2015). For flexible LA systems, misuse often includes loose installation, in which the flexible strap(s) are not pulled tightly enough to secure the CRS snugly to the vehicle seat. Most manufacturers instruct that CRS should be installed tightly enough to pass the “one inch test,” meaning that the CRS moves less than one inch when pushed laterally or forward. Results of observational studies also indicate that 32–40% of all LA installations had one inch of movement or more (Greenwell 2015). Other common misuses include incorrect orientation of LA connectors (19–21%), routing the LA strap through an incorrect belt path (38–33%), twisting the LA strap (16–30%), and attaching a LA connector to a non-anchor point in the vehicle (approximately 2%) (Greenwell 2015). Previous laboratory studies with adult volunteers revealed similarly high rates of misuse associated with flexible LA systems, including critical errors such as attaching the LA connector to a non-anchor point in the vehicle (Klinich et al. 2013; Mansfield et al. 2019). Sled testing has confirmed that incorrect belt path (for rear-facing (RF) CRS) and loose installation and top tether misuse (for forward-facing (FF) CRS) are the most consequential types of misuse in terms of safety (Manary et al. 2021). We hypothesize that rigid LA systems might reduce some of the most common errors. Rigid LA systems connect directly to the lower anchors without a strap, which means loose installations, mis-routed straps, twisted straps, and upside-down connector errors are unlikely or impossible. However, it is possible that rigid LA systems may result in types of misuse not identified with flexible strap LA systems.

Rigid LA systems have shown potential safety benefits over flexible webbing LA in dynamic crash studies, even when both systems are correctly installed. Results of sled testing suggest that rigid LA systems can reduce head excursion during the crash event and rebound phases, reduce head injury criterion (HIC) values, and reduce neck loads compared to flexible LA systems (Lowne et al. 2002; Charlton et al. 2004; Hauschild et al. 2018). Results of finite element simulations showed similar benefits (Kapoor et al. 2011). Notably, when crash tests were conducted for FF CRS without the top tether attached, the rigid LA installation performed better than the flexible LA (Kapoor et al. 2011; Hauschild et al. 2018). These outcomes suggest that the performance of rigid LA systems in a crash might be less affected by top tether misuse than flexible LA systems. This is important because caregivers neglect to attach the top tether in approximately 44%–58% of FF CRS installations (Eichelberger 2014; Greenwell 2015).

Due to slow integration into the US market, most consumers in the US are currently unfamiliar with the rigid LA system. To the authors’ knowledge, no studies have compared the consumer usability of rigid LA compared to flexible LA. It is unclear whether consumers might prefer the rigid LA system, whether it would reduce some of the most common types of LA misuse associated with flexible LA straps, or whether it would result in different types of misuse not identified with flexible LA systems. The objective of this study is to evaluate consumer usability of rigid vs. flexible LA systems.

Methods

Recruitment

All research protocols were approved by the Ohio State University Institutional Review Board (protocol #2020H0257). Thirty adult volunteers were recruited via departmental emails and campus newsletter announcements. Inclusion criteria included: age 18 or older, able to lift and install a CRS into a vehicle, not currently pregnant, and English speaking. Previous experience with CRS installations was acceptable but not required. Participants were scheduled for a single 1-h appointment and given a $20 cash incentive payment. All participants provided their written informed consent to participate.

Equipment

Each participant was asked to install two different RF-only infant CRS bases. One CRS base had a standard single-loop flexible LA strap (the Combi Shuttle) and the other had rigid LA (the Clek Liing) (Figure 1). For the Combi Shuttle, the flexible LA strap was folded and stored inside of a compartment behind a plastic door (Figure A1, see online supplement). Each participant had to pull the LA strap out of the storage compartment, orient the connectors in the proper direction, and guide the strap through the appropriate belt path. The Clek Liing had rigid LA which could be adjusted to nine different length settings. The length of each LA connector could be set independently of the other. The nine length settings were marked so that researchers could easily record the participants’ selected setting(s). The Clek Liing includes a load leg. To keep the participants focused on the LA features, the load leg was folded into its stored position and taped over. The labels and instruction manual were modified to omit any references to or images of the load leg. The size and shape of the LA connectors were similar between the two CRS models (push-on mini-connectors on both).

Figure 1. CRS bases with flexible LA system (top) and rigid LA system (bottom).

Participants installed each CRS base into two different vehicles apiece, for a total of four installations per participant. The test vehicles were selected to represent different difficulty levels. One vehicle was a 2015 Chrysler Town & Country minivan with leather seat trim. The LA were visibly protruding from the seat bight. The Insurance Institute for Highway Safety (IIHS) rated this vehicle as “Acceptable” for LATCH Ease-of-Use (IIHS 2022). The other test vehicle was a 2015 Toyota Corolla sedan with cloth trim. The LA were positioned deeper in the seat bight and were not visible without reaching into the bight. The IIHS rates this vehicle as “Marginal” for LATCH Ease-of-Use (IIHS 2022).

Protocol

Participants completed the four installations in partially randomized order. Each was randomly assigned to their first CRS, which they installed in the first randomly selected vehicle followed by the same CRS in the second vehicle. Then, the second CRS was introduced and was installed into each vehicle in random order.

The researcher explained that the participant should install each CRS to the best of his/her ability. No guidance or feedback was given to the participant between tests. The researcher timed each installation but participants were encouraged to take as much time as they needed to ensure proper completion of the tasks. The CRS and vehicle manufacturers’ instruction manuals were provided. Upon completion of each installation, the participant was led away from the vehicle while a certified Child Passenger Safety Technician (CPST) evaluated their work. Visually identifiable errors were noted. The tension in the flexible LA strap was measured using a clip-on tension gauge. Tensions were considered correct at levels above 5.5 lbs in the sedan and above 7.5 lbs in the van. These thresholds were determined by the CPST installing the base several times in each vehicle to identify the approximate tension at which each passed the one inch test. More tension was required in the van due to the higher stiffness of the seat cushion and the low friction leather trim.

When all four installations were complete, participants completed a written survey. The survey contained open-ended queries to describe what they liked or did not like about each LA type. They used a seven point Likert scale to rate the intuitiveness, physical ease-of-use, and their self-confidence associated with each LA type. Participants were also asked which LA system they preferred overall. Sociodemographic data were collected (age, sex, education level, household income, race, ethnicity, and native language). Previous CRS experience was captured by asking participants to estimate how many CRS installations they had ever done in their lifetime (none, 1-5, 6-20, or more than 20), whether they had ever installed a CRS with rigid LA before, and whether they had ever attended a car seat checkup event. After study procedures were completed, the researcher demonstrated how to correctly install each CRS and provided the participant with educational materials, if desired.

Data analysis

The frequencies of individual installation errors were calculated. Installations were dichotomized into “correct” (zero errors) or “incorrect” (one or more errors). Pearson’s chi-square test was used to compare the rate of correct vs. incorrect installations between LA types. The amount of time taken for each installation was compared between LA types using two-tailed t-tests. Likert scale responses regarding intuitiveness, physical ease-of-use, and confidence were compared between LA types using Wilcoxon signed rank tests for paired data (two-tailed). Pearson’s chi-squared tests were conducted to determine which sociodemographic factors or previous experience factors influenced installation errors and overall LA preference.

Results

Participant characteristics

The thirty adult volunteers had an average age of 35.0 ± 10.9 years, with a range of 23–60 years. The cohort was comprised of 24 females and 6 males. All participants reported having a college degree or higher. The cohort was predominantly non-Hispanic/non-Latino (93%), white (90%), native English speakers (97%). Six participants had never installed a CRS, ten participants reported extensive experience of twenty installations or more, and the rest of the sample fell in between. One participant reported having installed a rigid LA CRS before, one was unsure, and twenty-eight had never installed with rigid LA before. Nine participants reported prior attendance at a child seat checkup event to have their child’s CRS checked by a professional. Full participant information is provided in Tables A1 and A2 (see online supplement).

Installation errors

The rigid LA base was installed a total of 60 times (once in each vehicle by each participant). Twenty-eight participants (93.3%) referred to the CRS instruction manual at least once for their rigid LA installations. The following errors were observed:

• 16/60 (26.7%) did not snug the base against the vehicle seat back (i.e., the LA connectors were set too long).

• 2/60 (3.3%) used the seat belt along with the rigid LA.

Some installations contained more than one error. In total, 43/60 installations (71.7%) were completely correct, while the remaining 17/60 installations (28.3%) contained at least one error.

The flexible LA base was also installed a total of 60 times (once in each vehicle by each participant). Twenty-eight participants (93.3%) referred to the CRS instruction manual at least once for their flexible LA installations. The following errors were observed:

• 22/60 (36.6%) installations were too loose.

• 12/60 (20.0%) had upside-down connector(s). Of these:

  • Five participants had both connectors upside-down for both installations (10/60 installations total).

  • One participant attached the LA connector upside-down in the sedan on purpose because the connector seemed to fit better in the bight in that orientation.

  • One installation had one connector upside down with a 180 degree twist in the webbing.

• 2/60 (3.3%) installations had the entire strap facing backwards, such that the adjustor pull strap was stuck beneath the webbing itself.

• 2/60 (3.3%) installations, both completed by the same participant, incorrectly routed the LA strap through the seat belt lock-offs.

• 3/60 (5.0%) installations used the seat belt along with the flexible LA.

Some installations contained more than one error. In total, 28/60 installations (46.7%) were completely correct, while the remaining 32/60 installations (53.3%) contained at least one error. This overall LA misuse rate aligns with rates reported in previous observational studies (Safe Kids USA 2011; Greenwell 2015).

Overall, the rigid LA CRS was installed correctly more often than the flexible LA CRS (71.7% correct vs. 46.7% correct, respectively, χ2 = 7.761, p = 0.0053).

Since loose installation was the most common error, this error was examined with respect to order of installation (i.e., each participant’s first or second installation of the flexible LA CRS) and also with respect to test vehicle (van or sedan). There was no significant difference with respect to each participant’s first or second installation (tight installation 56.7% and 69.0% among first and second installations, respectively, χ2 = 0.672, p = 0.412). Therefore, repetition of the task did not improve outcomes. This result is consistent with other studies where constructive feedback was not offered between installations and thus no improvement occurred (Klinich et al. 2014a; Mansfield et al. 2018). There was a significant difference with respect to vehicle, with higher rates of tight installations in the sedan compared to the van (75.8% vs. 50.0%, respectively, χ2 = 4.23, p = 0.040). This is likely because of the leather trim and stiffer cushion on the van, which required more tension on the strap to pass the one inch test.

Pearson’s chi-squared tests were conducted to investigate relationships between installation correctness and participant factors (Table A3, see online supplement). For the flexible LA installations, none of these factors were significant: age, sex, previous CRS experience level, or previous attendance at a car seat checkup event. For the rigid LA installations, a significant relationship was found for sex with males having higher rates of correct rigid LA installations compared to females (100% vs. 64.6%, respectively, χ2 = 5.93, p = 0.015). The remaining factors were not evaluated due to skewed sampling rates (education level, income level, race, ethnicity, native language).

Ease of use

The amount of time taken for each installation was investigated as a proxy for ease-of-use. Participants’ first installation of each CRS took similar amounts of time: means of 4.9 min for rigid LA and 5.6 min for flexible LA (p = 0.1165, two tailed t-test). On the second installation of each CRS, participants were able to install the rigid LA significantly faster than the flexible LA (means of 1.6 min vs. 2.9 min, respectively, p = 0.0007).

User preferences

Likert scale responses ranged from −3 (strongly disagree) to +3 (strongly agree), with zero being neutral. Responses were compared between rigid and flexible LA installations using the Wilcoxon signed rank test (two-tailed) for paired data (Table 1).

Table 1. User preferences.

Statement	Rigid LA, mean response	Flexible LA, mean response	p-value (Wilcoxon signed rank test)
The seat was physically easy to install (It did not take a lot of strength or physical effort)	1.87	0.93	0.033
The seat was intuitive to figure out (I could just look at it and figure out what to do).	0.77	0.13	0.067
I am confident that I installed the seat correctly	2.17	1.07	0.002

Therefore, the rigid LA system was rated as significantly easier to install. Participants reported no significant difference in the intuitiveness of the two LA types. Participants had significantly more confidence in their ability to install the rigid LA compared to flexible LA. For flexible LA, higher confidence was associated with correct overall installation (χ² = 4.401, p = 0.036). For rigid LA, higher confidence was not associated with correct overall installation (χ² = 1.788, p = 0.181). However, all participants expressed a positive degree of confidence with rigid LA so the range of responses was less varied.

When asked which LA system they preferred overall, 24/30 participants (80%) said they preferred the rigid LA system, 5/30 said they preferred the flexible LA, and one participant (3.3%) provided an ambiguous response stating that they preferred using the flexible LA during the installation but preferred the post-installation results of the rigid LA.

Factors such as age, sex, education level, and previous seat check attendance were not significantly associated with participants’ stated preference of LA type (Pearson’s chi-squared tests with the ambiguous response omitted; see Table A4, online supplement). Participants’ levels of previous CRS installation experience were also not significantly associated with LA preference (Pearson’s chi-squared, χ2 = 7.59, p = 0.055 with the ambiguous response omitted). However, most of the participants who preferred flexible LA or responded ambiguously were among the most experienced CRS installers (Figure 2).

Figure 2. Distribution of participants’ LA type preference with respect to level of previous CRS experience.

Participants who preferred the rigid LA gave the following reasons for their preference: “Overall easier to install,”, “less effort to ensure a good fit,” “felt very solid,” “seemed more secure,” “no need to adjust tension and no guess work to hope you have the belt tight enough,” and “easy to install once I figured out how.” Participants who preferred the flexible LA gave the following reasons for their preference: “Easier to latch before tightening (because latch points were harder to line up for the rigid LA),” “more intuitive and easier to install since the strap was flexible,” “easier to figure out without looking at directions,” and “overall physically easier.”

Discussion

Correct installation rates were higher for rigid LA compared to flexible LA (71.7% correct vs. 46.7% correct, χ2 = 7.761, p = 0.0053). The most common installation error for the rigid LA system was using a LA length setting that was too long (i.e., neglecting to snug the base against the seat back after attaching the LA). At present, there are no published data to determine possible safety consequences for this type of misuse. The most common installation error for the flexible LA system in the current study was loose tension on the LA strap. Loose installations rates in the current study match observational rates reported in the literature (Greenwell 2015). Crash testing literature suggests that RF CRS with minor/moderately loose installations, such as those documented in this study, may not have a significant effect on crash safety outcomes (Menon and Ghati 2007; Rudin-Brown et al. 2017; Manary et al. 2021). Loose installations appear to be more detrimental for FF CRS, especially when the top tether is also misused (Manary et al. 2021). Currently there are very few harnessed FF CRS with rigid LA on the US market. Future study may be warranted to determine if rigid LA can reduce loose installation rates for this type of CRS. In addition to sled testing data, real-world studies of CRS effectiveness with respect to misuse should also be considered. Elliott et al. (2006) report that CRS were associated with a 21% reduction in risk for death compared to seat belts. When cases of serious CRS misuse were excluded, the effectiveness was slightly higher at 28%. It is unclear whether effectiveness might increase further if mild/moderate instances of misuse (such as the loose installations documented in this study) were also considered. Given the relatively small effect of excluding serious misuse in the Elliott et al. study, it might be unlikely that minor misuse has a significant effect. However, more recent and more detailed crash data for modern CRS occupants is needed to answer these questions.

Previous studies also document other critical errors associated with flexible LA systems, such as attaching the LA connector to a non-anchor location in the vehicle or attaching only one LA instead of both (Greenwell 2015; Mansfield et al. 2019). In the current study, the LA of both the rigid and flexible systems were always attached to the proper LA hardware in the vehicle. Other types of misuse observed in this study have less clear consequences in terms of safety. Upside down connectors and using LA and seat belt together are generally considered non-severe errors and have been categorized as such in a survey of experts (Rudin-Brown et al. 2017). Regardless, the CRS manufacturer’s instructions directly advise against these practices. The manufacturers test and certify the CRS under proper use conditions, thus the performance of the CRS cannot be guaranteed in any other untested misuse conditions.

The results of this study suggest that rigid LA systems might be effective in reducing CRS misuse rates compared to flexible LA. Several European observational studies have compared the misuse rates of ISOFIX (similar to the rigid LA system studied here) compared to seat belt installations. Some studies found significantly fewer installation errors for CRS installed with ISOFIX compared to seat belts (Roynard et al. 2014; Kuhn et al. 2020). However, at least one study found no significant difference between ISOFIX errors and seat belt errors (Cornelissen et al. 2021). The most common errors associated with ISOFIX in these studies were failure to attach the LA on both sides and/or failure to use an accompanying anti-rotation feature such as the top tether or load leg. Therefore, while the rigid LA system cannot be expected to eliminate installation errors entirely, there is some evidence that it can reduce errors compared to seat belt installations (Roynard et al. 2014; Kuhn et al. 2020) and compared to flexible LA installations (current study).

Overall, 24/30 participants (80%) preferred rigid LA to flexible LA. Participants rated rigid LA as physically easier to install and reported higher confidence in installing rigid LA correctly compared to flexible LA. In general for flexible LA, higher confidence was associated with correct overall installation (χ2 = 4.401, p = 0.036). However, it should be noted that eight participants expressed the highest level of confidence that their flexible LA installation was correct when it was not actually correct. Similarly, eight participants also expressed high confidence in correct rigid LA installation when it was not actually correct. Of the participants with extensive CRS experience, approximately half preferred the traditional flexible LA compared to rigid LA. This suggests that consumers are sometimes more comfortable with features that are familiar to them. However, inexperienced CRS users generally preferred the rigid LA system, along with about half of the highly experienced users (Figure 2). These results suggest that, overall, rigid LA could be accepted by a wide range of CRS users.

Despite these potential benefits, the prevalence of rigid LA in the US market is low. Concerns about compatibility between the rigid LA system and vehicles has likely played a role. Rear row center seating positions often do not have dedicated LA hardware. Manufacturers sometimes allow a user to “borrow” the innermost LA from each adjacent outboard position to install a CRS in the center seat. These LA points are typically further apart than the standard 280 mm (Aram and Rockwell 2012; Klinich et al 2014b). Flexible LA systems can generally reach these anchor points but rigid LA systems cannot be adjusted to a wider lateral span. Additionally, anchors positioned several centimeters above the surface of the seat cushion might create an undesired pitch angle or gap underneath the CRS base when rigid LA are attached. Rigid LA systems are also more expensive to produce than flexible LA systems and these cost increases are passed down to the consumer. Design changes that add cost should have clear safety benefits because cost increases can lead to reduced use, particularly by marginalized low-income communities.

This study includes several limitations. Only one type of CRS (RF-only infant bases) were presented to participants. Convertible or other types of CRS might have additional considerations associated with rigid vs. flexible LA. Participants were not requested to perform complete CRS installations, so the results only apply to the LA portion of the bases. The two CRS in the study were produced by different manufacturers, so the labels and instruction manuals were formatted differently from each other. There are many different types of rigid LA. This study included only one type (adjustable length, non-rotating). Only two test vehicles were used. There are likely other relevant vehicle compatibility factors which were not included in the study. The education level of the participants was skewed, with all of the participants holding at least a college degree or higher. This is not representative of the average population, and was a result of recruiting on a college campus. The race, ethnicity, and native language of the participants was also skewed toward white, non-Hispanic, native English speakers. Previous studies have found significant differences in CRS use and misuse among communities of different races, ethnicities, and education/income levels, so the skewed cohort is an important limitation of this study. Participants completed the tasks in a controlled setting with no distractions or urgent time pressure, which is not always reflective of real-world situations. Ultimately, participants knew that no children would be injured as a result of their mistakes in this study, which might have affected the amount of effort they exerted for the tasks.

Acknowledgments

The authors would like to acknowledge the National Science Foundation (NSF) Center for Child Injury Prevention Studies (CChIPS) at the Children’s Hospital of Philadelphia (CHOP) and the Ohio State University (OSU) for sponsoring this study and its Industry Advisory Board (IAB) members for their support, valuable input and advice. The views presented are those of the authors and not necessarily the views of CHOP, OSU, the NSF, or the IAB members.

Additional information

Funding

The funding source is: Center for Child Injury Prevention Studies (CChIPS).