The Mini-BESTest is a clinical balance assessment tool that is a shortened version of the Balance Evaluation Systems Test (Franchignoni et al., 2010). It targets four different balance control systems that contribute to dynamic balance, which are considered functional and influenced by foot posture and proprioception (Menz et al., 2006). The four balance control systems are 1) anticipatory postural adjustments, 2) reactive postural control, 3) sensory orientation, and 4) dynamic gait; all of which may be affected by exercises targeting the IFM. There are 14 items scored on a 3-level ordinal scale. The maximum score is 28, indicating little to no fall risk, and the minimum score is 0. Based on a Rasch validation study by Franchignoni et al. (2015), we created three fall risk categories for participants’ total score (none to mild risk 28–24, moderate risk 23–18, high risk ≤17). The equipment required for the test includes a 4-inch foam pad (medium density T41 firmness rating), an incline board (10°–15°), a 9-inch-tall box, a chair without arm rests or wheels, a stopwatch, and a 3-m distance marked on the floor. It has been tested on several populations with balance disorders and was found to have high reliability and validity for measuring balance function and its change over time (Godi et al., 2013; O’Hoski et al., 2014). The Mini-BESTest has excellent concurrent validity with the Berg Balance Scale (r = 0.85, CI 0.78–0.90) (Franchignoni et al., 2010). It also has high content validity, as the test consists of other well-known balance tests including the Berg Balance Scale, the Timed Up and Go, the Dynamic Gait Index, and the Clinical Test of Sensory Integration in Balance (Franchignoni et al., 2010). The Mini-BESTest is cited as a good tool for predicting fall risk in community-dwelling older adults ages 60–102 (Magnani et al., 2020), and there are published normative values for healthy adults ages 50–89 (O’Hoski et al., 2014). When used to measure community-dwelling older adults with balance disorders, it had a low ceiling effect, an established minimal detectable change (MDC) of 3.5 points and cut-off scores for predicting falls based on age (Godi et al., 2013; O’Hoski et al., 2014; Marques et al., 2016). A minimal clinically important difference (MCID) of two points was established as significant progress in balance in adults with recent total knee arthroplasty, and authors recommended it as a balance assessment tool (Chan et al., 2020). The Mini-BESTest will be conducted by supervised DPT students and scored by a physical therapist researcher with 17 years of clinical experience. Further, each test will be video recorded for review and assistance with scoring.

DPT students will be trained in performing the Mini-BESTest by two physical therapy faculty members, one with over 30 years of clinical practice and teaching experience and the other with 17 years of clinical practice and 6 years of teaching experience. The Mini-BESTest is part of these students’ regular curriculum. They individually read and learn the components of the test along with the standardized script. They next practice the test on other students, and then practice the test on faculty members unaffiliated with the current research study. They also observe videos of older adults with different balance abilities partake in the Mini-BESTest.

Proprioception is a sense or perception of position in an environment and employs a complex interaction of the visual, vestibular, and somatosensory systems (Rosker and Sarabon, 2010). It is a difficult element to measure given the number of structures that contribute to this sense, but from a clinically practical standpoint, the focus in this research will be on kinesthesia, or consciously perceived joint position sense. Literature regarding the reliability and validity of clinical ankle/foot proprioceptive measures is scarce. This body of literature tends to use highly technical measuring devices, oftentimes custom-made by researchers. A more simple and clinical approach to proprioception measurement can be done with active or passive positioning of a joint (Rosker and Sarabon, 2010). The LEPT is a passive positioning test that has recently been developed as a clinical tool without the use of technical equipment, and has been tested for reliability in healthy older adults (fair to good test-retest reliability, ICC = 0.60–0.68) and in populations post-stroke (good test-retest reliability ICC = 0.79–0.85) (Ofek, 2019; Ofek et al., 2019). This tool assesses a combined movement of the knee and ankle to two distances (12 and 22 cm) while the participant is in a seated position, eyes closed, wearing a thin sock, with their foot resting on a non-friction surface. The examiner slides the foot to a measured point on the floor (12 cm), returns it to the starting position, and then repeats the passive movement until the participant says “stop” at the point they think replicates the first position. The test is repeated using a 22 cm distance. The difference between the designated point and where the participant is asked to stop is calculated for each distance. Standard Error of Measurement (SEM) calculations from a reliability study of the LEPT by Ofek, (2019), found the 12 cm distance SEM to be 1.14 cm and the 22 cm distance SEM to be 1.28 cm. For the present study, measures within the SEM for each distance will be considered “accurate”. If a participant is outside the SEM, it will be considered “inaccurate”. A change from inaccurate to accurate performance will be considered an improvement in proprioception. Because there is evidence that IFM strengthening interventions can affect loading and alignment of the knee (Trombini-Souza et al., 2015), and because previous research participants’ subjective reports suggested enhanced overall lower extremity proprioception (Franklin et al., 2017b), it seems practical to include a proprioceptive measure that may capture effects beyond the foot and ankle. A physical therapist researcher with 17 years of clinical experience will conduct this measure.

Intrinsic foot muscle cross-sectional area (CSA, cm2) of the right abductor hallucis, flexor hallucis brevis, flexor digitorum brevis, quadratus plantae, and abductor digiti minimi muscles will be measured using ultrasound imaging (Figure 4). Although there are additional plantar intrinsic foot muscles, these five muscles were selected based on existing research regarding ultrasound imaging measurement reliability, and based on links to foot function. The decision to measure only one foot (right) is due to the number of participants (n = 90), time points (n = 4), and muscles (n = 5) totaling 1800 data points related to muscle morphology alone. Ultrasound imaging is a reliable method to measure morphology of IFM in young and middle-aged adults as shown by previous researchers (McKeon et al., 2015; Fraser and Hertel, 2019a; Mickle et al., 2013; Crofts et al., 2014; Latey et al., 2018; Fraser et al., 2018; Wong, 2007). In these previous studies, researchers measured IFM muscle thickness or cross-sectional area and found good to excellent between-session and inter and intra-operator reliability. In brief, for this study, images of each muscle will be captured at a frequency of 10 MHz with a GE NextGen LOGIQ e R7 ultrasound unit (GE Healthcare, United States) with a wide-band linear array (12 L-RS) probe using techniques described by Latey et al. (2018) and Mickle et al. (2013) With the participant positioned in right side-lying on a height-adjustable treatment table, each muscle will be located using adjacent bony landmarks, then viewed in the long axis. Passive and active motion of the corresponding toe(s) will be performed to confirm the muscle. The thickest portion of each muscle will be visually identified, centered on the monitor, and then the probe will be rotated 90° to obtain a short-axis view. Passive and active movements again can be used to confirm the muscle as needed, and an image will be captured. Two images of each muscle will be captured using this same technique. A single physical therapist researcher will perform all ultrasound imaging. Intra-reliability analyses using intraclass correlation coefficients (ICC) and Fleiss’ kappa will be conducted to confirm accuracy of this measure. Values will be interpreted as 0–0.5 poor, 0.51–0.74 fair, 0.75–0.9 good, and >0.9 excellent reliability, with ≥0.75 being acceptable agreement. A separate researcher blinded to intervention group will view each captured image and determine the CSA using ImageJ software (National Institutes of Health). For each muscle, the two CSA measures will be averaged and analyzed for changes in size. An increase in CSA of a muscle indicates hypertrophy, and thus, increased strength (Mayer et al., 2011). Calculations of muscle size change will be presented as percent changes.

Navicular drop is a simple and clinically applicable measure of foot pronation requiring only a ruler. It involves measuring the height of the navicular bone from the floor when a participant is seated with the hip and knee at 90° and the foot and ankle positioned in subtalar joint neutral. The participant then moves to bilateral full weight bearing relaxed stance, and the height of the navicular bone is re-measured. The difference is taken between the seated and relaxed stance measures. A normal navicular drop is ≤10 mm while excessive and possibly pathological navicular drop is ≥15 mm (Brody, 1982). We will use these values to indicate improvement in arch height following interventions. Navicular drop has been found a reliable measure for foot pronation by several authors when performed by an experienced clinician (Menz, 1998). Excessive foot pronation is dysfunctional, can lead to several foot disorders and can contribute to falls in older adults (Menz et al., 2006; Hagedorn et al., 2013).

Hallux valgus angle is a simple clinical measure performed with a handheld goniometer with the participant seated with the hip, knee, and ankle joints at 90° and the foot resting on the floor. Hallux valgus is defined as ≥15° of abduction of the hallux from the first metatarsal, and has been further classified into stages of severity (mild 15°–25°, moderate 25°–35°, severe >35°) (Hagedorn et al., 2013). Greater hallux valgus angle has been linked to fall risk by several authors (Menz et al., 2006; Mickle et al., 2009; Koski et al., 1996). Measurement of hallux valgus is typically included in a comprehensive foot exam and was found to have good inter examiner reliability in a population of older adults (kappa values > 0.85, p < 0.01). (Hagedorn et al., 2013). When compared to the gold standard of radiographic measures of hallux valgus, clinical goniometry performed by an experienced clinician was within one°–5.5° (Janssen et al., 2014). Tang et al. conducted a 3 month intervention to reduce hallux valgus angle and found a reduction of 6.5° ± 3.8 resulted in significantly reduced pain and improved walking ability (Tang et al., 2002). Based on Tang et al.‘s work, we will consider a change of ≥7° to be a meaningful improvement. Although IFM function contributes to foot and toe alignment, we have included hallux valgus as an exploratory measure and do not anticipate it will significantly change in an older adult population. This can be likened to the malalignment of an osteoarthritic knee that does not change with exercise intervention alone (Raghava et al., 2001).

A physical therapist researcher with 17 years of clinical experience will perform all foot measures.

Fall Rate will be determined by the number of falls per person in the 12-month study period. Falls will be recorded based on journal entries and will be cross-referenced with subjective reports at 2-week contacts. Researchers and participants will use the operational definition of a fall as an unexpected event in which a person comes to rest on the ground or lower level (Laurence, 2021; Guirguis-Blake et al., 2018).