Abstract
This research aims to address limitations of current hand grip measurement techniques by designing, building, and testing a more robust cylindrical dynamometer capable of measuring orthogonal force components while accounting for forces exerted by individual finger segments. The novel grip fixture consists of three static beams instrumented with strain gauges measuring orthogonal forces and eight attached miniature compensating load cells accounting for force contribution from finger segments. Following device fabrication and calibration, data was validated in a pilot study (n=6) where grip strengths were compared between the custom fixture and a Jamar dynamometer. Data suggests the custom dynamometer provides hand grip strength measurement comparable to that produced on a traditional dynamometer (p > 0.05). Future research studies using a similar device can guide therapy for patients with impaired hand strength by providing clinical professionals with a more complete profile of the forces exerted by hand segments.
Introduction
Grip strength measurement is utilized in both research and clinical settings due to its high correlation with physical and mental health (Peterson et al., 2017; Carson, 2018). Clinical research indicates that weaker grip strengths are related to decreased functional mobility, higher mortality rates, and an increased chance of cognitive decline (Carson, 2018; Vancampfort et al., 2019). In addition, grip strength readings are often used as a method for clinical professionals to evaluate the efficacy of different rehabilitation strategies for patients with upper extremity impairments (Sirajudeen et al., 2012; Lo et al., 2021).
Grip strength is most often quantified using dynamometers measuring total grip force in only one direction. The most commonly used device to assess grip strength is the Jamar dynamometer, which is recommended as the gold standard by the American Society of Hand Therapists (Hogrel, 2015). Handheld dynamometers are tools capable of measuring isometric grip force while a subject squeezes the two parallel handles of the device together with maximum force obtaining an overall grip force reading (Hogrel, 2015).
Despite the widespread acceptance of handheld dynamometers to record grip strength, researchers acknowledged that such devices may be limited in scope for certain patient populations (Hogrel, 2015; Tyler et al., 2005). For example, grip dynamometer readings for patients with rheumatoid arthritis are often reduced due to the patient’s limited range of motion in the joints, making it more difficult to squeeze with maximum force. Similarly, strength readings recorded from older adults or sick patients may reflect inaccurately low values due to difficulty handling the weight and rigidity of the device (Neumann, 2018). Tyler et al. (2005) highlights one limitation of the dynamometers, suggesting that the “incremental scale of the Jamar [dynamometer] is too large to detect small changes in strength.”
Another shortfall of dynamometers for research purposes is the inability to isolate the segment of the hand which exerted the maximum or minimum amount of force. Recently, some researchers have made steps to improve upon the tools and accuracy of grip force readings. The University of Wisconsin-Madison improved grip strength measurement by creating a multi-axis dynamometer capable of 1) measuring shear strain independent of location of the applied load and 2) accounting for orthogonal components of force (Irwin et al., 2013). Although the device provides a more accurate representation of the force across joints, it is limited to measuring the overall force exerted by all four fingers across each beam preventing a more precise force representation of individual finger segments.
One reason a custom device has not been built measuring individual finger force components is the difficulty creating small sensors with high enough resolution to localize finger segment forces. To remedy this situation, our lab previously designed, built, and tested custom-made miniature compensating load cells (Laird et al., 2020). The compensating load cell differs from traditional load cells in industry applications which are limiting due to their relatively large size. The custom load cell is small enough to measure forces exerted by individual finger segments independent of the location of applied force.
The purpose of this study was to create a grip force dynamometer capable of measuring grip force contributions of individual finger segments utilizing the miniature compensating load cells. This paper presents the methods used to create and test the novel, custom fixture capable of measuring orthogonal force components of individual finger segments.
Methods
Design and Development of Custom Fixture
Building the device
The novel prototype designed and produced for this study consists of three beams: two “segmented” beams with purpose-built miniature load cells to measure the individual finger segments; and, one “reference” beam that measures the opposing force exerted by the palm/thumb hand segment. All three beams were machined with two perpendicular pockets instrumented with strain gauges to obtain orthogonal components of force. The strain gauges were attached as close as possible to the neutral axis minimizing spurious measurements of bending stresses. The beam design and strain gauge placement can be seen in Figure 1 . The two segmented beams were designed to accommodate four of the miniature load cells in order for the forces exerted from the different segments of each of the four fingers to be detected. All eight load cells were individually instrumented with strain gauges arranged in a Wheatstone Bridge configuration to detect the level of force exerted by individual finger segments. The strain gauges are configured to measure the applied force independent of the force distribution or location (Laird et al., 2020). Figure 2 includes images of the load cell design and the strain gauge configuration.
Beam design in CAD (left). Pockets machined in the beams contain strain gauges to measure shear stress (right).
Load cell design in CAD (left). Leaves of the load cell are retrofitted with strain gauges wired in a Wheatstone Bridge configuration to detect force (right).
Assembling the prototype
A 3D-printed cylindrical outer casing serves as the interface between the load cells and the hand. Figure 3 depicts the completed device assembly. As shown in the figure, the two beams instrumented with four load cells are aligned to detect the force from each individual finger. The force measurement obtained from individual load cells sum to that obtained for its respective beam. The third beam does not contain load cells but instead provides a reference measurement for overall force exertion.
Assembled grip fixture (left). Device in use (right).
Calibrating the device
The beams and each individual load cell were calibrated using LabVIEW (National Instruments; Austin, TX) software, known weights, and an analog strain conditioning system (Vishay Micro-Measurements; Raleigh, NC). Beams were clamped horizontally and known weights (2.3 kg, 4.5 kg, 6.8 kg, 9.1 kg) were hung from the same location along the long axis of the beam. The output voltage was recorded, and calibration equations were plotted to be used in the LabVIEW program. Figure 4 provides a representative calibration plot taken from a single load cell.
Calibration equation generated for one of the eight load cells and is representative of all eight calibration plots.
Signals acquired from the pockets were constant across all four load cells for a certain weight, consistent with the theory that shear strain can be measured independent of location of the applied load. This theory was tested by applying a known weight (4.5 kg) across different distances from the center of the second pocket (2.5 cm, 5.1 cm, 7.6 cm, 10.2 cm, 12.7 cm) and confirming that the output remained the same, as portrayed in Figure 5 . The device was also tested to confirm that the sum of the voltages acquired from the load cells equal those acquired from the strain gauges in the pockets of that beam. The resulting graph can be seen in Figure 6 . Finally, Figure 7 confirms that the sum squared of the output signals acquired from the two pockets of a beam equals the applied known weight when the beam is rotated along its axis.
Graph confirming shear strain measured on beams is independent of application of applied force.
Graph confirming that the sum of the signals acquired from the load cells equals those acquired from the pockets of its respective beam.
Graph confirming the sum squared of the output signals acquired from the two pockets of a beam when rotated by varying angles equals the applied known weight.
Validation of Device Accuracy: Pilot Overall Grip Study
To verify that the custom dynamometer reflects accurate grip force readings, a pilot study was conducted in which grip forces measured using the custom dynamometer were compared to those from the Jamar dynamometer.
Approach
In a laboratory setting, the grip strengths of participants were recorded using (1) the custom dynamometer developed in this study and (2) the Jamar dynamometer. To mimic the cylindrical grip of the custom dynamometer, a cylindrical handle of the same diameter (7.6 cm) was 3D printed and affixed to the handle of the Jamar dynamometer for testing, as seen in Figure 8 .
Jamar dynamometer.
Participants
A total of 6 males, ages 21-26 years, participated in the study. The participants signed consent forms approved by the University’s Institutional Review Board (IRB). None of the participants reported any current injuries, illnesses, or pains that affected hand strength or hand function. Anthropometric data for the hand were recorded for each participant. These measurements included hand breadth, hand circumference, hand length, and length of palm. Table 1 summarizes the anthropometric data collected for each participant.
Descriptive statistics (mean ± sd, [min-max]) of anthropometric hand measurements of participants.
Experimental protocol
Each participant was informed about the purpose of the study prior to data collection. Hand measurements were then taken and recorded. Participants were brought to the testing area where they were familiarized with the two dynamometers (Custom and Jamar). When instructed by the experimenter, the participant gripped the dynamometer with maximum force for approximately 5 seconds. Participants were instructed to keep their elbow at a 90° angle and maintain neutral wrist posture. A 2-minute rest period was provided between each trial to eliminate any muscle fatigue. The order of dynamometers was randomized per participant to eliminate any order effect. This procedure was repeated for a series of four trials on each dynamometer.
Data analysis
The readings obtained from the Jamar test were averaged over all four trials for each participant. The readings acquired from the custom dynamometer were analyzed in two ways: (1) by summing the signals from all eight load cells plus the resultant forces obtained from the pockets of the reference beam and (2) summing the resultant signals obtained from all three beams. These two readings were compared to ensure similarity thereby confirming that the load cells were functioning properly. Readings from the custom dynamometer were then also averaged across the set of 4 trials for comparison with the Jamar readings.
Results
The overall grip force readings from the custom dynamometer were calculated (1) when accounting for the signals acquired from the load cells and (2) when only accounting for the signals acquired from the beams’ pockets. Figure 9 graphically depicts data from both dynamometers. The mean values of the grip force for the custom dynamometer when accounting for the load cells signals, the custom dynamometer when only accounting for the beams, and the Jamar dynamometer were 29.1 kg, 31.2 kg, and 29.8 kg, respectively. This is shown in Figure 10 . Both force readings obtained from the custom dynamometer were determined to be statistically similar (p=0.795) thereby verifying the load cells’ functionality. To validate the accuracy of the custom dynamometer’s overall force readings, the data was compared to readings obtained from the Jamar test (the industry gold standard). Again, grip forces were statistically similar (p=0.974). P-values obtained from Tukey post-hoc tests are outlined in Table 2 .
Graphical representation of data recorded from grip force tests.
Mean grip force readings.
P-Values using Tukey pairwise comparisons and 95% confidence.
Discussion
The objective of this paper was to detail the design and development of a custom hand grip dynamometer that measures orthogonal components of force, while also accounting for individual finger segment contributions. The purpose of the pilot study was to validate the accuracy of the device. Findings confirm that the dynamometer provides accurate grip force readings by comparing results to the industry standard Jamar test.
A previous study completed by Irwin, et al. (2013) details the development of a multi-axis dynamometer that quantifies orthogonal components of force. Calibration results from the 2013 study agree with data from the present study. Both dynamometers allow for shear strain to be measured independent of the location of the applied load. As discussed by Irwin et. al (2013), this characteristic allows for accurate grip measurement “in circumstances where subjects may tend to grip differently, e.g., for a range of different hand sizes, or following injury or disease.” The dynamometer developed in this study differs from the one in the previous study in that it is instrumented to measure individual finger segment forces using miniature compensating load cells. Due to the difficulty in developing small enough load cells to be used in the prototype, the grip span for this device was fixed at 7.6 cm. According to Irwin et al., the average force exerted for this grip span was approximately 21.4 kg, compared to the 29.1 kg force obtained from this study. The difference between these results can be attributed to the small sample size and/or the fact that the current study only evaluated grip strength of males, whereas half of Irwin et. al’s participant base was female.
Although mean grip strength measurements were comparable between dynamometers (p=0.974), some variability was observed between measurements. One source of variability may be maintenance of the participant’s arm posture. Each participant was instructed to maintain a 90° elbow angle and neutral wrist posture during testing; however, there was no physical constraints implemented to enforce this protocol (e.g., cast to maintain certain wrist posture; clamp/brace to maintain elbow angles). The lack of physical constraints could have inadvertently resulted in differing postures between the two dynamometers, thereby affecting the force readings. It’s also unclear if and to what extent the participant was pushing or pulling on the custom dynamometer using forces from their forearm or body weight.
Additional variability may be due to the participant being required to support the weight of the Jamar dynamometer in their hand, while the custom dynamometer was affixed to a table for support. Finally, the lack of immediate feedback from the custom dynamometer could also present a source of variability. The fact that the Jamar instantly displays the grip force exerted could cause incentive to squeeze with more force.
Future research initiatives will focus on ways to develop even smaller load cells to be used for such an application. Smaller sensors will provide the opportunity for handle adjustability and make it possible to test on a wider range of hand sizes (i.e., children). Researchers will also explore ways to develop an improved prototype that leverages the use of externally positioned sensors.
A long-term research goal involves developing a parametric hand model relating grip strength and hand shape. Existing parametric hand models are configurable only by overall body dimensions or hand length or breadth. However, additional hand dimensions (e.g., relative finger length and thickness) are important variables when analyzing hand functionality and strength. Having the ability to account for small changes in strength contributions from individual finger segments will allow researchers to develop a more robust parametric model to be used in ergonomic analysis.
Another future goal is to conduct full scale research studies utilizing similar dynamometers to study the grip strength of patient groups experiencing hand impairments. These participants may include people with osteoarthritis, carpal tunnel syndrome, rheumatoid arthritis, and/or stroke. Findings from future studies may provide data to create and guide therapy capable of improving hand function for this percentage of the population. Ideally, more detailed data localizing hand impairment will allow data guiding clinical professionals to tailor hand therapy strategies to best accommodate the needs of different patient populations. Ultimately, findings will help researchers understand the nature of force production as it relates to coordination for different disorders resulting in neurological and biomechanical deficits in the hand.