Baby Carrying Method Impacts Caregiver Posture and Loading During Gait and Item Retrieval

Abstract

Background:

Human babies are carried by their caregivers during infancy, and the use of ergonomic aids to wear the baby on the body has recently grown in popularity. However, the effects of wearing or holding a baby in-arms on an individual’s mechanics during gait and a common object retrieval task are not fully understood.

Research question:

What are the differences in: 1) spatiotemporal, lower extremity kinematics, and ground reaction force variables during gait, and 2) technique, center of mass motion, and kinematics during an object retrieval task between holding and wearing an infant mannequin?

Methods:

In this prospective biomechanics study, 10 healthy females performed over-ground walking and an object retrieval task in three conditions, holding: (1) nothing (unloaded), (2) an infant mannequin in-arms, and (3) an infant mannequin in a baby carrier. Mechanics were compared using repeated measures ANOVA.

Results:

During gait, greater vertical ground reaction force and impulse and braking force was found during the in-arms and carrier conditions compared to unloaded. Significant but small (<5°) differences were found between conditions in lower extremity kinematics. Increased back extension was found during carrier and in-arms compared to unloaded. Step length was the only spatiotemporal parameter that differed between conditions. During object retrieval, most participants used a squatting technique to retrieve the object from the floor. They maintained a more upright posture, with less trunk flexion and anteroposterior movement of their center of mass, and also did not try to fold forward over their hips during the two loaded conditions. Lower extremity kinematics did not differ between unloaded and carrier, suggesting that babywearing may promote more similar lower extremity mechanics to not carrying anything.

Significance:

Holding or wearing an infant provides a mechanical constraint that impacts the forces and kinematics, which has implications for caregivers’ pain and dysfunction.

1. INTRODUCTION

At the beginning of life, human infants are helpless. Unlike many infant mammals, they are unable to transport themselves, creating a significant obligation for a caregiver. The use of ergonomic aids to attach an infant to the caregiver’s body for transportation (babywearing) is not a new practice but has recently grown in popularity. Caregivers benefit from holding the baby close while keeping their hands free to attend to other tasks. Other proposed benefits of infant wearing include increased mother-infant attachment [1], increased breastfeeding duration [2], and reduced agitation [3].

Researchers have begun to explore the effects of infant carrying on caregivers’ mechanics during walking. Juanqueira et al., [4] and Schmid et al. [5] demonstrated increased lumbar lordosis and thoracic kyphosis [4, 5], backward inclination of the trunk [4], and paraspinal muscle activity [5], indicating that individuals modulate their trunk mechanics while holding or front carrying infant mannequins compared to unloaded gait. Babywearing is ultimately a load carriage task, and differences in forces and loading have been recently reported. Brown et al. [6] showed an increase in normalized impact peak vertical ground reaction force (GRF) and anteroposterior GRF impulse while walking using baby carriers compared to unloaded. Lower extremity joint moments also differ, and a recent research concluded that knee and hip frontal plane loading moments were more similar between unloaded and babywearing conditions compared to in-arms carrying [7]. This line of research has begun to demonstrate that infant carrying affects caregivers’ biomechanics. However, no study has investigated lower extremity gait kinematics during these conditions.

Aside from gait, few other tasks have been studied with respect to baby carrying. Another common task for caregivers is picking up an object from the floor without putting the infant down. This task may be performed dozens of times a day, particularly when caring for other young children who drop objects and play on the floor. Object retrieval is an unconstrained task that can be achieved in many different ways [8], and requires individuals to both lower their center of mass and reach their arm out to lift the object. Both of these subtasks have been shown to require postural adjustments to prevent falling [9, 10]. Holding an infant may alter the technique and mechanics used, and a better understanding of this may reduce falls.

The purpose of this study was to characterize the biomechanical differences between holding and carrying an infant during gait and a retrieval task. We hypothesize that (1) spatiotemporal, GRF and sagittal plane lower extremity and trunk gait mechanics will differ between carrying an infant in arms, holding an infant in a carrier, and an unloaded control condition, and (2) that individuals will choose different techniques and exhibit differences in timing, center of mass displacement and sagittal plane kinematics when retrieving an item from the ground between the three conditions. Understanding the biomechanical impact of different carrying methods may ultimately inform caregivers on best practice to avoid injury or pain.

2. METHODS

2.1. Participants

The Institutional Review Board at the University of Denver approved this study. Ten nulliparous healthy females (27.4±4.1 years; 62.6±12.2 kg; 1.7±0.1 m) with no previous babywearing experience provided informed consent. No participants had any musculoskeletal or neurological disorders.

2.2. Instrumentation

Marker-based motion capture (100 Hz; Vicon Motion Systems, Oxford, UK) monitored movement (detailed in Figure 1). Two force platforms embedded into the floor recorded GRF (1000 Hz; Bertec Corporation, Columbus, OH).

Figure 1.

Experimental setup of one participant in each testing condition: (A) unloaded, (B) in-arms, and (C) carrier. 9 mm reflective markers were placed on anatomical landmarks: C7, T10, manubrium, xiphoid process, and bilaterally on the anterior superior iliac spine, posterior superior iliac spine, greater trochanter, lateral and medial femoral epicondyles, tibial tuberosity, medial and lateral malleolus, calcaneus, 1^st and 5^th metatarsal heads and the base of the 2^nd toe. Marker clusters were also placed on the lateral surface of the thighs and shanks half-way between the proximal and distal joint centers.

2.3. Procedures

Testing was conducted at the Human Dynamics Laboratory in the Center for Orthopaedic Biomechanics at the University of Denver. Age, height, and weight were recorded for each individual.

Participants performed gait and an item retrieval task for each of three conditions: (1) holding nothing (unloaded), (2), holding an infant mannequin (5000 g, Dietz, Freiburg, Germany) in her arms (in-arms), and (3) holding an infant mannequin in a soft-structured baby carrier (carrier) (Ergobaby, Inc., Los Angeles, CA; Figure 1). For in-arms, individuals were instructed to carry the mannequin close to their own bodies as if it was a living infant with good head control, and participants self-selected their carrying strategy with the infant in contact with their own bodies. For carrier, individuals watched an instructional video from a certified babywearing consultant explaining the correct way to fit the All-Position 360™ baby carrier before self-fitting the carrier and the mannequin inward-facing on their own bodies [11].

For gait, individuals were instructed to walk overground at a self-selected, comfortable pace (20 meters, with force platforms near 10 meters). For the retrieval task, a hand towel was placed beyond the force platform’s edge. Individuals were instructed to walk to the towel, pick it up, and return to standing. Participants completed several trials of each task until three trials with clean force plate strikes were completed for each condition (in-arms, carrier, unloaded), in a randomized order.

2.4. Data Analysis

Marker and GRF data were low pass filtered using 6 Hz fourth-order Butterworth filter, in accordance with previous studies [12–14]. Reconstructed marker coordinate data were used to compute segmental kinematics (Visual3D v.6.01.35, C-Motion, Germantown, MD). Joint kinematics were calculated using the joint coordinate approach following Cardan sequence of rotations [15].

2.4.1. Gait

Spatiotemporal gait parameters were calculated and expressed as dimensionless values by using leg length as the anthropometric factor [5, 16]. Variables of interest included step length, step time, stance time, gait speed, and step width.

GRF impulse was calculated for posterior (braking), anterior (propulsive), mediolateral, and vertical direction during stance [17]. Time of posterior and anterior GRF (braking and propulsive time) were calculated. Peak GRFs were also identified. Vertical peak GRF was identified during midstance (∼25% stance) [18]. Force data were normalized to body weight of participant [6].

Sagittal plane kinematics of the ankle, knee, hip, pelvis, and trunk were identified at key gait events [18]: peak ankle plantarflexion during loading response, peak ankle dorsiflexion at terminal stance (∼75% stance); peak knee flexion during midstance (∼25% stance) and pre-swing (toe-off); peak hip and trunk flexion and extension; peak maximum and minimum pelvic posterior tilt.

2.4.2. Retrieval

Techniques used during the retrieval task were qualitatively assessed and grouped into three categories: squat, lateral bend, and stoop [8]. Squat was defined as a deep knee bend with little lateral bend or forward flexion. Lateral bend was defined as an asymmetric bend at the waist. Stoop was defined as an anterior bend at the waist with little knee flexion. The lowest point of the squat (squat depth) was identified as the minimum vertical position of the center of mass (COM). Times from standing upright to squat depth (downward phase) and from squat depth back to standing (upward phase) were also calculated. The COM displacement during the downward phase was calculated in the anteroposterior, mediolateral, and vertical directions and normalized to subjects’ body height. Bilateral kinematics were identified at squat depth, including sagittal plane ankle and knee, and tri-planar angles of the hip and trunk.

2.5. Statistics

Shapiro-Wilk tests of normality indicated that parametric tests were appropriate. To determine differences in mechanics between conditions, repeated measures ANOVA was used (α≤0.05). Effect sizes were determined using partial eta-square and interpreted as small (0.01), medium (0.09) and large (0.25) [6]. Paired t-tests were used for pairwise comparisons, when appropriate. Bonferroni adjustments were used for multiple comparisons (α=0.017). Statistical analyses were performed using SPSS (Version 18, Chicago, IL).

3. RESULTS

3.1. Gait

All gait values are presented in Table 1. Few differences were found between conditions in spatiotemporal gait parameters. Only step length was significantly different between conditions, with 2.2% greater step length for unloaded compared to in-arms.

Table 1:

Gait spatiotemoral, ground reaction force, and kinematic data by condition (unloaded, arms, and carrier) with corresponding statistical results.

Variables	Unloaded	Arms	Carrier	F-statistics	p	η2
Spatiotemporal parameters
Step length*	0.90 ± 0.07	0.88 ± 0.06 a	0.87 ± 0.08	(2, 0.004) = 3.877	0.040	0.301
Step time*	2.03 ± 0.16	2.04 ± 0.21	2.00 ± 0.14		0.586
Stance time*	2.26 ± 0.18	2.33 ± 0.25	2.28 ± 0.20		0.173
Speed*	0.45 ± 0.06	0.44 ± 0.05	0.45 ± 0.05		0.327
Step width*	0.16 ± 0.05	0.16 ± 0.04	0.16 ± 0.04		0.939
Impulse
Braking (BWs)	0.033 ± 0.006	0.038 ± 0.006 a	0.038 ± 0.006 a	(2, 0.0001) = 11.822	0.001	0.568
Propulsive (BWs)	0.030 ± 0.006	0.032 ± 0.004	0.032 ± 0.005		0.227
Braking time (s)	0.387 ± 0.024	0.393 ± 0.033	0.392 ± 0.031		0.678
Propulsive time (s)	0.314 ± 0.026	0.327 ± 0.033	0.312 ± 0.026		0.158
Mediolateral (BWs)	0.020 ± 0.011	0.021 ± 0.010	0.022 ± 0.010		0.694
Vertical (BWs)	0.553 ± 0.033	0.615 ± 0.050 a	0.608 ± 0.042 a	(2, 0.012) = 23.487	<0.001	0.723
Ground reaction force
Braking (BW)	−0.19 ± 0.04	−0.22 ± 0.04 a	−0.22 ± 0.04 a	(2, 0.002) = 8.13	0.003	0.475
Propulsive (BW)	0.20 ± 0.04	0.21 ± 0.04	0.21 ± 0.04		0.192
Medial (BW)	0.07 ± 0.02	0.07 ± 0.02	0.07 ± 0.02		0.578
Vertical (bW)	1.14 ± 0.10	1.24 ± 0.11 a	1.26 ± 0.12 a	(2, 0.065) = 47.071	<0.001	0.839
Peak Kinematics
Ankle plantarflexion (°)	4.6 ± 2.2	4.0 ± 2.9	4.3 ± 2.8		0.728
Ankle dorsiflexion (°)	13.8 ± 2.8	15.2 ± 2.2 a	15.6 ± 2.1 a	(2, 8.748) = 8.221	0.003	0.477
Knee flexion midstance (°)	12.0 ± 7.6	12.8 ± 8.0	14.4 ± 7.2	(2, 15.339) = 5.374	0.015	0.374
Knee flexion toe off (°)	46.4 ± 3.4	49.5 ± 3.8 a	49.5 ± 5.0 a	(2, 31.997) = 7.034	0.006	0.439
Hip flexion (°)	29.2 ± 6.6	28.7 ± 6.4	32.3 ± 6.3	(2, 37.222) = 3.968	0.037	0.306
Hip extension (°)	8.6 ± 6.1	9.3 ± 7.6	6.6 ± 6.6		0.143
Pelvis max posterior tilt (°)	11.6 ± 4.6	10.3 ± 5.1	12.9 ± 4.6		0.077
Pelvis min posterior tilt (°)	8.3 ± 4.5	6.1 ± 4.9 a	9.4 ± 4.1 b	(2, 27.637) = 6.61	0.007	0.423
Trunk flexion (°)	4.27 ± 4.3	0.30 ± 5.1 a	−2.22 ± 3.7 a	(2, 107.059) = 15.656	<0.001	0.635
Trunk extension (°)	0.7 ± 4.7	4.1 ± 5.9 a	8.0 ± 4.4 a	(2, 132.126) = 14.766	<0.001	0.621

^*Expressed as dimensionless number according to Hof [16];

^asignificantly different to the unloaded condition (p<0.017),

^bsignificantly different to the in-arms condition (p<0.017)

Compared to unloaded, total vertical impulse was 10.1% greater during carrier and 11.3% greater in-arms, and vertical GRF during midstance was 10.0% greater during carrier and 8.6% greater in-arms. Braking impulse was 16.0% greater during carrier and 14.8% greater in-arms, and braking GRF was 11.9% greater during carrier and 11.3% greater in-arms compared to unloaded.

At the ankle, peak ankle dorsiflexion was significantly different between conditions, with about 2° more dorsiflexion in carrier and arms compared to unloaded. Knee flexion angle at toe-off was significantly different between conditions, with 3° greater flexion during both carrier and in-arms compared to unloaded. The ANOVA revealed a significant difference in knee flexion angle during mid-stance and in peak hip flexion during loading response, but pairwise comparisons did not reach significance. Both peak trunk flexion and extension differed between conditions. Individuals exhibited more trunk extension during the in-arms (4°) and carrier (7°) conditions compared to unloaded. Compared to in-arms, participants exhibited 2–3° more minimum pelvis posterior tilt in the unloaded and carrier conditions.

3.2. Retrieval

Participants used different techniques to retrieve the object from the floor (Figure 2). Most individuals performed the squat technique to retrieve the object: 70% (21/30) of trials for unloaded and carrier and 77% (23/30) for in-arms. Some individuals used the lateral bending technique: 10% (3/30) for unloaded, 30% (9/30) for carrier, and 23% (7/30) for in-arms. The stoop technique was only used during unloaded in 20% (6/30) of the trials. Most individuals utilized the same technique for all trials of a particular condition. Because the majority of trials from all participants were performed using the squat technique, only these trials were used for the subsequent analyses.

Figure 2.

Examples of one participant performing the item retrieval task using the (A) stoop and (B) squat technique.

There were no timing differences between the downward and upward phases of the squat between the conditions (Table 2). However, the anterior-posterior translation of the COM was different between the conditions during the downward phase, with 19% more motion during unloaded compared to carrier.

Table 2.

Retrieval timing, center-of-mass (COM), and kinematic data; COM variables all reported as dimensionless values (normalized to body height).

Variables	Unloaded	Arms	Carrier	F-statistics	p	η2
Timina
Time down (s)	0.70 ± 0.08	0.87 ± 0.21	0.87 ± 0.13		0.152
Time up (s)	0.97 ± 0.11	1.02 ± 0.11	1.07 ± 0.08		0.234
COM translation
ML	0.032 ± 0.012	0.039 ± 0.022	0.031 ± 1.01 ,		0.326
AP	0.241 ± 0.043	0.166 ± 0.068	0.202 ± 0.090 a	(2, 0.011) = 8.237	0.011	0.673
Vertical	0.206 ± 0.047	0.243 ± 0.046	0.229 ± 0.062		0.174
Kinematics
Left ankle dorsiflexion (°)	−4.5 ± 6.8	4.1 ± 6.9	2.8 ± 4.0		0.073
Right ankle dorsiflexion (°)	3.2 ± 7.2	9.3 ± 5.2	8.5 ± 5.1		0.080
Left knee flexion (°)	95.2 ± 28.4	125.7 ± 24.5 a	116.6 ± 27.8	(2, 544.373) = 10.236	0.006	0.719
Right knee flexion (°)	95.8 ± 30.4	118.5 ± 23.1 a	111.9 ± 24.1 b	(2, 369.444) = 10.481	0.006	0.724
Left hip flexion (°)	83.3 ± 10.7	90.1 ± 12.8	93.6 ± 9.6		0.102
Right hip flexion (°)	74.8 ± 9.8	79.1 ± 16.0	86.7 ± 12.5		0.123
Left hip abduction (°)	1.9 ± 5.7	7.7 ± 12.9	19.5 ± 7.3		0.086
Right hip abduction (°)	10.3 ± 8.1	19.4 ± 13.2	24.2 ± 10.7	(2, 435.813) = 7.813	0.013	0.661
Left hip internal rotation (°)	−4.7 ± 6.6	4.3 ± 14.6	−3.3 ± 14.1		0.898
Right hip internal rotation (°)	−18.8 ± 14.0	−16.3 ± 9.7	−16.6 ± 10.6		0.580
Trunk flexion (°)	70.9 ± 13.8	49.0 ± 10.5 a	37.8 ± 16.5 a	(2, 1037.833) = 19.109	<0.001	0.827
Trunk frontal (°)	−3.0 ± 3.2	−0.9 ± 4.2	3.5 ± 8.4	(2,128.743) = 9.053	0.009	0.694
T runk rotation (°)	19.2 ± 9.2	15.3 ± 17.5	16.5 ± 14.0		0.539

^asignificantly different to the unloaded condition (p<0.017),

^bsignificantly different to the in-arms condition (p<0.017)

Individuals performed the squat asymmetrically to retrieve the object. There were significant differences in lower limb kinematics at the squat depth between right and left legs when all conditions were considered; therefore, kinematics of right and left side were analyzed separately.

Knee flexion angles were significantly different between conditions for both right and left limbs, where 23–30° greater flexion was found in-arms compared to unloaded for both legs, and 7° greater in-arms compared to carrier for the right leg. Although the right hip frontal plane angle was significantly different between conditions, pairwise comparisons did not reach significance and the left frontal plane hip angle was not different between conditions. Trunk flexion angle was significantly different between conditions, with 22° less flexion for in-arms, and 33° less flexion for carrier compared to unloaded. Trunk frontal plane angle was also significantly different between conditions, but the pairwise comparisons did not reach significance. Figure 3 highlights the greater flexion of the knee and less flexion at the trunk for in-arms and carrier compared to unloaded.

Figure 3:

The relative sagittal plane joint angle contribution to the vertical movement of the body when subjects squatted to retrieve the object from the floor

4. DISCUSSION

The purpose of this study was to identify differences in gait and retrieval mechanics between carrying an infant mannequin in-arms, in a structured carrier, and an unloaded control condition. The data supported our hypotheses as (1) biomechanical differences existed between the conditions for both the gait and retrieval tasks, and (2) participants demonstrated different techniques when retrieving an object from the ground.

4.1. Gait

Our results are consistent with previous studies showing a shorter step during infant carrying compared to a control condition [5, 19]. However, step length was the only spatiotemporal gait parameter that differed between conditions in our study. While Juanqueira et al [4] demonstrated that individuals carrying an infant mannequin in arms walked slower than when not carrying anything, other researchers have not shown a difference when an infant mannequin was carried in a structured carrier [5, 6]. Our data are consistent with these studies, as we found that individuals did not walk with different speeds, or step or stance times between conditions. Similar to Schmid et al. [5], we expressed spatiotemporal gait parameters as dimensionless numbers by scaling to body size, which may account for the difference between our findings and Juanqueira et al. [4].

Vertical GRF and impulse were both greater during the carrying conditions compared to unloaded. The mannequin added a 5–10% body weight load. The increased vertical GRF was approximately proportionate to the increase in load, which is consistent with other research on infant carrying [6] and backpack loading [20, 21]. As stance time was not affected by the loading conditions, the increase in vertical impulse was likely driven by the increased force rather than time of force application. No difference was found between the carrier and in-arms conditions, suggesting that the added load of the carrier itself did not significantly contribute to the interaction force at the ground. Braking force and impulse were also larger during the two loaded conditions. Time spent braking was not different between conditions. Similar to vertical, this finding indicates that braking impulse differences were also driven by the magnitude of force. Differences were also not found in gait speed, suggesting that the increased braking force did not slow the individuals during gait.

Sagittal plane kinematics differed between the carrying conditions, particularly for trunk. When compared to the unloaded condition, individuals exhibited more trunk extension during both loaded carrying conditions. In fact, during the carrier condition, individuals never flexed their trunks during stance but remained extended throughout. This is similar to previous research demonstrating more backward trunk inclination during front pack carrying [22] and infant carrying [4], which may be a compensation to keep the body-plus-infant COM over the base of support. Slightly more peak ankle dorsiflexion was found during arms and carrier compared to unloaded, as seen in front pack carrying [22]. This ankle position at terminal stance may facilitate forward progression of the anteriorly shifted COM beyond the base of support in the carrying conditions. At the hip, individuals exhibit peak hip flexion at or near initial contact. This facilitates weight bearing stability while the knee and ankle go through an initial arc of motion [18]. We found a trend towards increased hip flexion in the carrier condition, which is similar to previous research on military backpack carrying. Majumdar et al [23] demonstrated a trend towards increased hip flexion with increased load magnitude.

Previous research has suggested that knee flexion during early stance is important to attenuate forces during increased loading [20, 21, 24]. This makes sense; peak flexion occurs during the transition between loading response and midstance and contributes to controlled shock absorption [18]. Williams et al. found increased knee extension moment during infant carrying and wearing [7], similar to the results found during front-pack carrying [22]. Similarly, our study identified differences in knee flexion at ∼25% of stance and toe-off, which may indicate a pattern of more flexed lower extremity posture to attenuate the larger vertical forces and accommodate the added load.

Together our findings suggest that individuals alter their mechanics in order to accommodate an infant mannequin. They do this similarly whether holding in arms or in a baby carrier during gait, but alter their kinematics compared to holding nothing. However, our results and those of others studying load carriage [22, 23] demonstrate small differences in joint angle (<5°), making clinical relevance unclear [7, 25], as these kinematic differences may not result in significant changes to joint moments. We did not analyze joint moments during this task; however, Williams et al [7] reported no differences in hip or ankle sagittal plane joint moments during baby wearing or holding. However, it is important to note that even small changes in joint moments over many cycles may cause pain or discomfort in some individuals, so further study is required to truly understand the impact of small changes in kinematics.

4.2. Retrieval

During the retrieval task, individuals approached an object on the floor and picked it up using any self-selected technique. The techniques varied between conditions, supporting our hypothesis. For most trials, participants performed a squat-like technique by flexing their knees while maintaining an upright torso. Only during the unloaded condition did participants use the stoop technique. This makes sense, given the added load was meant to simulate carrying a living baby. While an individual could bend at the waist while holding or wearing baby, they likely prefer not to invert the baby’s body in the process. The stoop lifting strategy is performed by using a large amount of trunk flexion, which results in a large external moment arm for the low back [26]. This strategy may increase spinal loads compared to other lifting techniques that rely on increased knee and hip flexion and decreased trunk flexion [27, 28].

Foot placement was not constrained during the retrieval task and individuals achieved the task goal through a variety of joint positions and strategies, which resulted in asymmetrical lower extremity positions. This differs from other retrieval research that required participants to pick up an object in a prescribed way [8]. Because literature suggests that a free style lifting strategy allows for more self-perceived natural and safer functional strategies [29], we purposefully chose to leave the task unconstrained to allow for self-selected strategies. Generally, participants used a large amount of knee flexion in conjunction with hip flexion, abduction and some external rotation. This appears consistent with the pattern used during late pregnancy for a stand to sit task [30]. For both pregnancy and baby carrying, individuals must manage a large and delicate mass in front of them while maintaining their balance as they go from standing to a seated or squatting posture.

Participants reached the same vertical depth when using the squat technique during all conditions. This indicates that all participants lowered their COM to the same degree, but they used a variety of joint kinematics to accomplish this. Participants relied on trunk flexion to accomplish the retrieval task while unloaded, and they flexed at their knee during in-arms. In the carrier condition, individuals used a combination of sagittal and frontal plane movements to reach the retrieval goal. Interestingly, no differences existed in the lower body kinematics between the carrier and unloaded conditions, which suggests that the use of a baby carrier may promote lower extremity mechanics more similar to unloaded retrieval compared to carrying an infant in arms. This is likely beneficial when considering joint moment differences during lifting. Similar joint positions between unloaded and carrier suggest that lower extremity external moment arms were not increased. Joint torque would thus be affected only by the relatively small increased load of the infant mannequin but not increased moment arms.

4.3. Limitations and Future Work

While other studies have examined gait during infant holding and carrying [4–7, 31], this study also uniquely examined lower extremity mechanics during these conditions. Limitations of this study include the tasks chosen, and subject population and size. The results of this study did not identify large effects of infant transportation method during short duration overground gait. Other studies have used longer duration walking (10–20 minutes on treadmill) [7, 31, 32], which may impact the mechanics used. This is the first study to our knowledge to analyze a retrieval task during baby carrying. The unconstrained retrieval task allowed participants to use any single method, which reduced the number of trials available for comparative analysis. Also, we used nulliparous females and infant mannequins, and future studies could consider analysis of postpartum mothers or other caregivers with their own infants.

4.4. Conclusions

This study characterized the biomechanical differences between holding and carrying an infant mannequin during gait and a retrieval task. While GRF variables differed during gait, spatiotemporal and lower extremity kinematics showed small differences, and clinical significance is unclear. During an item retrieval task, participants selected more upright postures and strategies while holding or carrying the infant mannequin. Lower extremity joint positions were similar during the unloaded and carrier conditions. Taken together, the results of this study suggest that infant transportation affects caregiver mechanics, particularly during object retrieval. As few differences were found between lower extremity kinematics during both tasks between the baby carrier and unloaded conditions, this research agrees with Williams et al. [7], suggesting that the use of a baby carrier may promote mechanics that more closely resemble unloaded mechanics.

HIGHLIGHTS.

• Mechanics of babywearing, in-arms holding, and unloaded conditions were investigated

• Greater forces and back extension found in-arms and baby carrier vs unloaded in gait

• Small (<5°) differences found in lower extremity kinematics during gait

• During retrieval, a more upright posture was used in-arms and carrier vs unloaded

• Babywearing may promote more similar lower extremity postures to unloaded conditions