The Pronation Control paradigm is one of the earliest and most well-known of all the footwear paradigms. Emerging in the late 1970s, footwear was seen as a possible avenue to respond to the increase in running-related injuries paralleling the rise in popularity of recreational running (Jacobs and Berson, 1986). The main premise of the Pronation Control paradigm was that excessive pronation at the subtalar joint during the stance phase of running increases stress on the anterior medial region of the knee, leading to pain and/or injury. Specifically, subtalar pronation promotes internal rotation of the tibia through joint coupling with the foot, which in turn was thought to place greater stress at the tibiofemoral and/or patellofemoral joints (Hintermann and Nigg, 1998). Since most of the injuries in recreational runners were—and still are—at the knee (van Gent et al., 2007; Tonoli et al., 2010), using footwear to prevent increased knee joint stress was a logical method to reduce knee pain and injury (Stefanyshyn et al., 2006; Liao et al., 2018).

Shoes aligning with this paradigm are constructed with “motion control” technology. Typical motion control technology involves a region of dense (harder) midsole material along the longitudinal mediolateral arch to reduce midfoot pronation during stance. Alternatively, or additionally, posting, or wedging is used at the rear portion of the shoe to reduce rearfoot motion and provide increased stability to the heel. Since pronation is a tri-planar motion of the foot and ankle complex (calcaneal/rearfoot eversion, forefoot abduction, and ankle dorsiflexion) and involves all three regions of the foot (fore-, mid-, and rear-foot), minimizing rearfoot eversion can influence pronation.

Runner assessment for recommending footwear in line with the Pronation Control paradigm typically begins with evaluating static foot posture but also may include assessing dynamic foot movement during activity. Static foot posture is typically evaluated in double-leg standing and includes medial longitudinal arch height, navicular height, and amount of rearfoot eversion from subtalar neutral. Additionally, the Foot Posture Index (FPI-6) can be used clinically to manually quantify or classify foot posture (Redmond et al., 2006). The assessor may ask the runner to perform a series of single leg squats to observe the amount of change in arch/navicular height in combination with knee frontal plane position (Magrum and Wilder, 2010). If room or equipment is available, runners may be asked to run while the practitioner examines rearfoot eversion and pronation position at midstance. Typically, health care professionals use qualitative analysis (Nicola and Jewison, 2012), but recent advancements and access to video annotation software allow for quantification of rearfoot eversion angle using 2D video (Souza, 2016).

From this assessment, a health care professional would determine whether foot posture and/or motion during running was considered a risk factor for injury and needed to be controlled, in part, using shoe technology. If so, a stability or motion control shoe would be recommended based on the severity of foot posture/motion. Specifically, “low” arch height and/or “very excessive” pronation/rearfoot motion would be matched to a motion control shoe, which has significant motion control technology. A “medium” arch height and/or “excessive” pronation/rearfoot motion would be matched to a stability shoe, which has some motion control technology but less than a motion control shoe. Importantly, there is no standard accepted threshold for excessive pronation or medial (frontal plane) knee motion, although normative running biomechanics suggest that the rearfoot is typically in 6°-8° of calcaneal inversion at initial contact and shifts to ~6°-8° of eversion by midstance (Dicharry, 2010).

It is important to highlight that while we have laid out the original paradigm rationale and accompanying approach and recommendation, most of the current evidence does not support the underlying rationale of injury paradigm or that the key footwear feature reduces risk. Specifically, excessive pronation was not found to be a risk factor for non-specific injuries in a large prospective study of novice runners (Nielsen et al., 2014) and was not associated with specific types of running-related injuries (Chuter and Janse de Jonge, 2012). Moreover, limited evidence exists to indicate that structural alignment is a primary risk factor for injury (Wen et al., 1997, 1998; Saragiotto et al., 2014a) or that static foot posture accurately reflects dynamic foot motion during running (Behling and Nigg, 2020).

Studies support that midsole technology can effectively alter rearfoot motion (Cheung et al., 2011). VanWoensel and Cavanagh (van Woensel and Cavanagh, 1992) tested customized shoes with the same midsole hardness and varied the degrees of rearfoot posting (10 degrees of valgus, 10 degrees of varus, and neutral shoe). Shoes with varus or valgus posting produced forced supination or pronation, respectively, when compared to running in a neutral midsole. Lilley and Dixon (2013) examined foot kinematics and knee forces in 30 females with a minimum of 3 years of running experience while running in a motion control (Adidas Supernova Sequence) and a neutral shoe (Adidas Supernova Glide). Peak rearfoot eversion and knee internal rotation moment were reduced in the motion control shoe compared to the neutral shoe. Butler et al. (2006) studied several measures of rearfoot motion, tibial motion, and loading in 20 low-arched and 20 high-arched recreational runners. Arch height was classified with the Arch Height Index Measurement System. All participants ran in a motion control shoe (New Balance 1122) and a neutral shoe (New Balance 1022). While the motion control and neutral shoes elicited significant differences in most of the biomechanical parameters, only instantaneous vertical loading rate was significantly different between low- and high-arch runners. Thus, study authors questioned the validity of matching an arch type to a shoe type to reduce injury risk.

Several studies have examined the incidence of injury following prescription of running shoes based on foot posture assessment. Two large prospective studies (Knapik et al., 2010, 2014) were performed on military cadets who were assigned one of three running shoes (motion control, stability, neutral) based on their arch height (low, medium, high). Arch height was measured by two independent classifiers who evaluated an image created from standing on a reflective mirror and were classified into three categories (high, medium, or low arches). Assigning shoes based on arch type did not significantly reduce the rate of injuries during 12 weeks of basic training for male or female military recruits (Knapik et al., 2010, 2014). Ryan et al. (2011) found that female recreational runners who were randomly assigned motion control shoes incurred more missed days (79 days) due to pain during a 13-week half marathon training program than those assigned a neutral (64 days) or stability (51 days) shoe. Likewise, Nielsen et al. (2014) did not find significant differences in injury risk among 927 novice runners with varying arch types who trained in neutral shoes. Malisoux et al. (2016) found contrasting results when randomly assigning either a neutral or motion control shoe to a group of 372 recreational runners. During 6-months of observation, sixty (of 185) runners in the neutral shoe group reported an injury compared to 33 (of 187) in the motion control group. However, further research is needed to substantiate these findings as twice as many runners in the motion control group were censored in analysis with both groups having relatively equal numbers of injury free runners (91 in the neutral shoe group compared to 96 in the motion control shoe group). Together these findings suggest that recommending footwear from static foot posture assessment or degree of rearfoot eversion during stance phase of running is not currently supported in most cases and that reducing foot motion through motion control shoes may even be injurious.

The Impact Force Modification Paradigm theorizes that excessive forces imparted on the body during the impact phase of running (i.e., landing and loading response phases) contribute to injury. Primarily two biomechanical features, the impact peak and the loading rate (slope of the rising portion) of the vertical ground reaction force (Figure 3), have been targeted under this paradigm. These features were selected because they occur during the loading phase of stance and were thought to correlate with larger (injurious) body loads. While the impact peak itself has not been well-correlated to injury, a larger (steeper) vertical ground reaction force loading rate has been suggested as a risk factor for injury in distance runners (Hreljac, 2004; Zadpoor and Nikooyan, 2011; van der Worp et al., 2016). However, recent prospective studies call into question the evidence supporting this association, particularly as a global risk factor for running-related injuries (Ceyssens et al., 2019). The presence of an impact peak also can provide information about the foot strike pattern. A visible impact peak is typically associated with rearfoot striking (Lieberman et al., 2010; Daoud et al., 2012), whereas impact peak is typically less pronounced or absent in non-rearfoot striking. This phenomenon becomes important to understanding the development of minimalist footwear, which aims to reduce impact by promoting a non-rearfoot strike pattern, and maximalist footwear that aims to reduce impact forces with increased midsole thickness.

Historically, midsole cushioning has been the primary footwear feature targeted by this paradigm. Initial studies found that vertical ground reaction force loading rates were less steep when runners wore a more cushioned shoe (Clarke et al., 1983). Thus, since the majority of recreational runners are rearfoot strikers (de Almeida et al., 2015), meaning they contact the ground first with their heel, a cushioned heel became the target feature to mitigate injury (Hintermann and Nigg, 1998). Conventional running shoes typically did not vary in the amount of thickness until mid 2000s. Prior to this time most changes in midsole cushioning were driven by selecting materials that absorbed more energy at impact. Beginning in the mid 2000s alterations to the midsole thickness also have been employed to alter cushioning properties. Increased thickness also allows for variations in the ratio of rearfoot to forefoot thickness (heel-to-toe drop) and the ability to add technology to the midsole (e.g., carbon fiber plates). These allowances have been targeted to alter foot strike pattern and promote a shift of the ground reaction force anteriorly during stance (Nigg et al., 2021), respectively.

The most common assessment method to estimate injury risk is still observation of running form. Traditionally, measuring ground reaction forces to calculate loading rates and estimate joint resultant forces and moments required expensive, laboratory-based force-, or pressure-sensing devices like platforms and instrumented treadmills. More recently, wearable accelerometers have been used to measure peak positive tibial acceleration as a surrogate for loading rate (Henning and LaFortune, 1991; Van den Berghe et al., 2019) while force- or pressure-sensing insoles have been used to directly measure vertical ground reaction force during running. However, these wearable devices do not provide sufficient data to calculate joint kinetics. Foot strike pattern, foot inclination angle (the angle between the foot and the horizontal in the sagittal plane), step rate, and stride length are considered kinematic correlates to ground reaction or joint forces (Heiderscheit et al., 2011; Lenhart et al., 2014a,b; Wille et al., 2014; Souza, 2016). Runners who rearfoot strike tend to have low step rates, use longer steps, and run with higher magnitudes of peak vertical ground reaction force, loading rates, peak braking forces, and potentially larger lower extremity joint forces than non-rearfoot strikers. Following the paradigm, clinicians would recommended a more cushioned running shoe to these runners to mitigate the higher repetitive forces. Importantly, there are no quantitative cut-points or thresholds used to determine a “risky” kinematic measurement. Moreover, as stated above, strong evidence to support the underlying kinetic measure to which these kinematic features are associated is lacking (Ceyssens et al., 2019).

Currently, there is not sufficient evidence to support that increasing cushioning via changes in midsole material reduces external vertical ground reaction forces (Baltich et al., 2013, 2015; Addison and Lieberman, 2015; Malisoux et al., 2021), which is the rationale behind the paradigm. It is important to note that the validity of using the impact peak of the vertical ground reaction force to evaluate the effect of shoe cushioning has been questioned Shorten and Mienjtjes (Shorten and Mientjes, 2011). Shorten and Mienjtjes demonstrated that the effect of cushioning on reducing impact forces generated under the heel was pronounced when specific analyses (frequency-domain rather than traditional time-domain) were applied. However, studies using in-sole pressure sensors have found differences in impact across cushioning levels (Kersting and Bruggemann, 2006; Dixon, 2008; Shorten and Mientjes, 2011). With respect to injury, more research is needed to determine the influence of cushioning on injury risk as only two studies have examined this relationship and results are mixed (Theisen et al., 2014; Malisoux et al., 2020), with Malisoux et al. (2020) finding that only lighter runners experience the protective effect of cushioning.

The minimalist shoe movement emerged as a variation of the impact modification paradigm, beginning in the mid 2000s. It was a response to the insufficiency of cushioned running shoes to reduce impact forces or injuries. This movement was fuelled by the rapidly rising interest in barefoot running around the same time (Rothschild, 2012; Davis, 2014). Minimalist shoes are designed to be thin, have a flexible midsole, and offer minimal to no arch stabilizing technology. Minimalist shoes are intended to permit free motion of the foot to achieve as close to a barefoot style of running as possible, while still providing protection from glass, debris, and rough road surfaces. The rationale for using minimalist shoes for impact force modification is that a barefoot running style is associated with a forefoot strike pattern, which eliminates the appearance of an impact peak on the vertical ground reaction force curve (Lieberman et al., 2010). A forefoot strike pattern has also been associated with lower magnitude and loading rate of impact forces in forefoot compared to rearfoot strikers (Almeida et al., 2015; Xu et al., 2021). The magnitude of midsole thickness is the main feature distinguishing minimalist shoes from each other and traditional footwear. Esculier et al. (2015) developed a standardized method for classifying minimalist shoes called the Minimalist Index Rating Scale. It is important to note that minimalist shoes do not always lead to a significant shift in footstrike pattern. Not all rearfoot strikers transition to a forefoot strike pattern when footwear is reduced (Willson et al., 2014; Tam et al., 2016). Moreover, runners who rearfoot strike in minimalist shoes tend to demonstrate higher peak ground reaction forces and vertical loading rates compared to conventional shoes (Paquette et al., 2013; Willson et al., 2014; Willy and Davis, 2014; Rice et al., 2016). Thus, debate continues about whether foot strike pattern (Lieberman et al., 2010; Lieberman, 2012; Shih et al., 2013), type of running shoe (Giandolini et al., 2013; Rice et al., 2016), or a combination of the two is best to reduce impact forces.

However, studies examining the influence of minimalist shoes on injury or running biomechanics related to injury do not support initial claims. Specifically, higher vertical ground reaction force loading rates and impact peaks were found when running in minimalist compared to conventional shoes (Sinclair et al., 2016a; Agresta et al., 2018; Jandová et al., 2019). Tibial acceleration, which has been suggested as an in-field measure for vertical loading rate (Van den Berghe et al., 2019), also was higher when running in minimalist compared to conventional shoes (TenBroek et al., 2014; Agresta et al., 2018). Moreover, there is evidence to suggest that switching from traditional to minimalist shoes may contribute to short-term (up to 12 weeks) foot and lower leg pain or injury (Ryan et al., 2014; Agresta et al., 2018) but may result in plantar-flexor strength gains in the longer-term (20 weeks) (Fuller et al., 2019). Body mass also appears to modulate the relation between cushioning and injury risk as heavier runners seems to be at greater risk of injury in minimalist shoes (Fuller et al., 2017). Fuller et al. (2017) found an influence of weekly training distance (>35 km/week) on injury risk in minimal shoes. However, this finding may relate more to a global running consideration rather than a runner-shoe interaction. Previous research (Rasmussen et al., 2013) has found injury risk increases at a similar weekly volume (30 km/week) when footwear wasn't controlled, and material testing has demonstrated deterioration of the functional properties of shoes after only 1,000 km of use (Hennig, 2011).

Maximalist shoes are another recent evolution of the Impact Force Modification Paradigm and a counterpart to minimalist shoes. Like minimalist shoes, midsole thickness is the primary distinguishing feature from traditional shoes. However, unlike minimalist shoes, no standardization currently exists for classifying maximalist shoes. Generally, shoes having more than 20 mm midsole thickness and minimal support technology are considered maximalist shoes. Less research exists on maximalist shoes compared to minimalist shoes. Maximalist shoes were first adopted by those who felt the extra cushioning protected against joint stress but otherwise wanted minimal support structure to promote a “natural” gait pattern. While the increase in midsole thickness of a maximalist shoe does not seem to reduce vertical ground reaction force (Agresta et al., 2018; Kulmala et al., 2018) or joint resultant forces (Sinclair et al., 2016b; Chan et al., 2018), it does appear to be less likely to cause running-related pain or time-loss from running compared to minimalist shoes (Agresta et al., 2018), suggesting there may be some protective value in shoes having large midsole thickness.

The main premise of the Habitual Joint (Motion) Path Paradigm is that each runner has a unique trajectory for the motion of their joints (Nigg et al., 2015). The magnitude of the joint motion may fluctuate, but the trajectory, or path, is stable. Alterations in footwear that resist or do not allow for skeletal movement along this path may increase tissue stress—either from a deviation from their pathway or from increased muscle activity to keep their individual habitual joint path—which in turn increases the risk of injury (Nigg, 2001; Enders et al., 2013; Willwacher et al., 2020). Originally called the Preferred Movement Path, this paradigm has since been updated and renamed the “habitual joint (motion) path,” adding that joint motion takes the path of least resistance due to an individual's anatomy and passive tissue properties (Trudeau et al., 2019).

Highly controlled studies examining segment-to-shoe motion support the concept that runners have an individual joint path of motion. Nigg (2010) used bone pins in the femur, tibia, and calcaneus to measure skeletal motion during running and found minimal variation in skeletal movement despite changes in footwear. Additional studies using skin or shoe mounted markers to measure limb motion while manipulating shoe inserts have illustrated similar results (Eng and Pierrynowski, 1994; Nawoczenski et al., 1995). Importantly, both the Pronation Control and Impact Force Modification Paradigms focus on a specific biomechanical parameter with generalized norms for how footwear should influence it. The Habitual Joint (Motion) Path Paradigm moves away from generalizations and toward a more runner-specific recommendation method.

Since, by definition, habitual joint motion patterns are subject-specific, there is no specific key feature associated with this paradigm. However, an example may be minimal arch support that allows for the preferred movement of the foot into pronation rather than a medial post that may promote a more supinated position of the foot (opposite of the preferred path).

The Habitual Joint (Motion) Path Paradigm attempts to match runners to footwear that minimizes variability, or deviation, away from their habitual joint motion path. Currently, there are no standardized or common clinical tests to assess variability of or deviation from the habitual path. Trudeau et al. (2019) developed a field method to assess deviation in order to recommend footwear. The field method evaluates a runner's baseline habitual motion path by measuring lower-limb kinematics during double-legged half squats. The half squat position was chosen because it induces lower-limb flexion/extension movement similar to running but with less than half the force on the limb. Subsequently, participants completed a short run in a sock shoe, which has no footwear technology but a minimum level of cushioning to promote a “normal” gait pattern. Knee and ankle kinematics were compared between the squat and the run to determine whether the runner is a “high deviator” or “low deviator.” High deviators displayed large differences between the half-squat and run, and vice versa for the low deviators. Finally, they were tested with selected running shoes. Running shoes were deemed as appropriate if they decreased the difference between running and half-squat kinematics and inappropriate if they increased this deviation.

A few studies (Fuller et al., 2016; Schrödter et al., 2016; Weir et al., 2020a) have investigated the Habitual Joint (Motion) Path paradigm by examining the influence of running shoes on different measures of movement variability. In contrast to Pronation Control and Impact Force Modification, which relied on joint motions or forces measured at specific time points (e.g., initial contact and midstance) to test the effect of footwear, the Habitual Joint (Motion) Path requires a more wholistic calculation of how either whole body or select body segments are adapting across the entire gait cycle or multiple gait cycles. Testing the consistency of movement patterns—or its opposite, variability—is one such wholistic approach to studying movement.

Schrödter et al. (2016) defined “footwear-related variability” as the influence of shoes to force an individual outside their habitual motion path. They calculated the similarity of a runner's kinematical and kinetical patterns across a variety of footwear conditions. Runners with high footwear-related variability had patterns that changed more across footwear conditions. Based on their findings, the authors considered these runners to be more sensitive to running footwear, less able to maintain their habitual motion path, and at higher risk of injury compared to runners with low footwear-related variability. Runners having higher footwear-related variability also tended, albeit with weak association, to have lower peak knee adduction angle and moment, initial contact and peak hip external rotation angle, and peak ankle external rotation moment, suggesting a difference in running style between those runners who are more and less sensitive to footwear. Additionally, footwear-related variability was significantly correlated with gender, weight, and some measures of sagittal plane strength, which the authors posited could suggest female, lighter, and weaker runners are less able to maintain their habitual motion path.

Coordination variability is another biomechanical measure that has been used to explore this paradigm. In this case, coordination refers to the segment and/or joint pattern that is used to accomplish a given movement task, while variability refers to breadth of coordination patterns used by the individual to achieve the task. Coordination variability is considered indicative of the amount of flexibility of the motor control system and may have an optimal range, above or below which could lead to injury (Hamill et al., 2012). Weir et al. (2020b) examined coordination variability of the lower limb during a prolonged fatiguing running in a group of male recreational runners. Runners exhibited higher coordination variability for three of the seven measures in neutral shoes compared to stability shoes. However, it is not clear whether increased coordination variability in the neutral shoe represents a negative deviation from the habitual motion path in response to fatigue or a positive adaptation to fatigue that is not necessary or not possible in the stability shoe.

Fuller et al. (2016) examined the similarity of stride time from one stride to the next in traditional and minimalist footwear and found that there were generally no differences between shoes. The only exception was a subset of runners who spontaneously switched from their normal rearfoot pattern to a midfoot pattern when introduced to the minimalist shoe. This group reduced consistency, as measured by the long-range correlation of stride time, at the fastest speed tested. However, difference in variability between footfall groups did not reach significance. Long-range correlations indicate how similar a pattern is over time, with an alpha of 0.50 indicating that strides are uncorrelated. Debate exists as to whether a reduction in long-range correlations indicates a positive or negative response (Dingwell and Cusumano, 2010; Van Orden et al., 2011; Cusumano and Dingwell, 2013). In this study, the decrease in alpha combined the increase in coefficient of variation, as was the case with the runners who switched to a midfoot strike, suggest that these runners needed to make more corrections in stride and, thus, are potentially less able to adapt their gait pattern to novel running footwear and potentially more susceptible to injury.

To our knowledge, no studies have directly tested the effect of matching footwear to minimize biomechanical variability and/or deviation from a specific motion path on running related injury. However, one study (Willwacher et al., 2020) suggested that increased time outside one's habitual motion path was associated with tissue-related changes in the knee joint. In this study of 12 healthy recreational runners, medial femur, medial tibia, and patella cartilage volume reductions were larger after 75 min of running in a shoe that increased a runner's deviation from their habitual joint path compared to one that reduced the deviation.

The Comfort Filter Paradigm, proposed by Nigg et al. (2015), posits that a runner intuitively selects a shoe that is biomechanically optimal based on comfort. This emerging viewpoint suggests that subjective comfort is the most important factor for selecting running footwear to reduce injury risk (Nigg et al., 2015). One hypothesis about why comfortable shoes are less injurious is that increased impact forces, which are expected to increase injury risk, lead to more soft tissue vibrations. These vibrations must be counteracted by muscle activation, also referred to as muscle tuning, which feels uncomfortable and requires higher energy expenditure (Nigg, 2001). The hypothesis that comfort is linked to reduced injury risk was partially developed from research on the effect of insoles on movement patterns and injury in soldiers. Mundermann et al. (2001) provided six different insoles to 206 military training personnel without lower extremity injuries and asked them to assess comfort of the insoles. The test group received the insole that they rated most comfortable and used it for the next 4 months. They were compared with a control group of soldiers with no insoles who were exposed to the same military training. The test group had 53% fewer lower-extremity injuries than the control group. Interestingly, five out of six insoles were selected as most comfortable at a similar rate, which suggests that comfort was linked to individual-specific rather than insole-specific factors.

Since this paradigm is runner-specific and relies on subjective perception, there is not a key feature of shoe design associated with it. However, one could argue that the runner's perception of comfort is heavily weighted by the properties of midsole thickness, cushioning, and resiliency as well as shoe fit.

Methods to accurately assess shoe comfort are available (Mundermann et al., 2002; Lindorfer et al., 2019). Practically, there is no standardized method to assess this subjective measure other than asking the runner to rank and/or select based on individual preference.

Two studies have explored the connection between comfort and biomechanical variability. The possible association between footwear comfort and biomechanical variability was postulated after findings in ergonomics literature. Sondergaard et al. (2010) found a positive correlation between measure of self-reported discomfort and degree of variability (measured by standard deviation) in center of pressure (COP) displacement during sitting tasks. The degree of biomechanical variability is also thought to contribute to running injury risk (Hamill et al., 2012) as some studies have found a connection between differences in coordination variability and injury populations (Heiderscheit et al., 2002; Miller et al., 2008; Seay et al., 2011). Meyer et al. (2017) calculated kinematic relative variability of foot motion (root mean square of the standard deviation/root mean square of the mean) from inertial sensor data. Thirty-six recreational runners performed trials in five different running shoes and ranked subjective comfort. The most and least comfortable shoes were compared. Only transverse plane angular velocity and frontal plane acceleration of the foot during the end of swing were significantly different between most and least comfortable shoe. Since these measurements represent transverse and frontal plane motions that are influenced by more proximal segments, the relationship to footwear is inconclusive. Lindorfer et al. (2020) examined coordination variability of lower extremity joint pairings in relation to most and least comfortable running shoes and did not find evidence to support this connection. Other than the seminal study by Mundermann et al. (2001) linking insole usage to injury reduction, no other studies have linked comfort and injury conclusively.