INTRODUCTION

Cushions are routinely prescribed to wheelchair users who are at risk for developing pressure ulcers.  Wheelchair cushions are intended to adequately support the buttocks by either enveloping the tissues in an attempt to minimize pressure gradients or redistributing forces away from tissues under bony prominences. Because many designs of wheelchair cushions exist, there is a need to estimate their performance in supporting the body. Over the years, assessing cushion performance has involved both human subjects (Burns & Betz, 1999; Ferrarin, Andreoni, & Pedotti, 2000; Garcia-Mendez, Pearlman, Cooper, & Boninger, 2012; Kernozek & Lewin, 1998) and bench testing (Pipkin & Sprigle, 2008; Sprigle, Dunlop, & Press, 2003) . Cochran published an early set of tests designed to measure cushion performance (Cochran & Palmieri, 1980). These represented several important constructs but most tests could only be applied to flat cushions. These studies used indenters which were rigid. The ISO standard for wheelchair cushion evaluation based on interface pressure mapping (IPM) has rigid and compliant models(ISO, 2007). While there is a high repeatability and sensitivity with rigid models, only interface pressures can be measured. In distinction, compliant models better reflect the deformations that occur under load and offer the potential to measure these deformations and internal pressures.  These models which have a rigid substructure and an elastomeric base were designed as an attempt to represent the human physiology better (ISO, 2007). Prior studies have used the compliant indenter to measure the interface pressure (Pipkin & Sprigle, 2008; Sprigle et al., 2003) . IPM is used to pick the appropriate cushion by measuring the pressure between the skin and cushion (Kernozek & Lewin, 1998; Takechi & Tokuhiro, 1998) . Unfortunately, the IPM has not been found reliable for standardized testing (Sprigle et al., 2003; Srivastava, Gupta, Taly, & Murali, 2009) . IPM measures only the normal pressure at the cushion interface and cannot be used to characterize internal loading of the tissue. Internal loads of the tissue are important because pressure ulcers may be created either superficially or from deep within the tissue (Bouten, Oomens, Baaijens, & Bader, 2003; Srivastava et al., 2009) . Apart from determining the internal loads, deformation also plays a key role in the formation of pressure sores specifically the ones that are defined as deep tissue injury by the National Pressure Ulcer Advisory Panel (Washington, DC). The shape of the buttock-cushion interface has been measured using humans (Brienza, Karg, & Brubaker, 1996; Kadaba, Ferguson-Pell, Palmieri, & Cochran, 1984; Sonenblum, Sprigle, Cathcart, & Winder, 2013) . These studies are useful for learning about how human characteristics impact deformation but are not well suited to objective measurement of cushion performance due to individual preference and morphology. Our aim is to develop and validate a compliant loading indenter capable of measuring both the internal load and deformation of the ersatz tissue.

DESIGN CRITERIA

The design of the indenter was based on anthropometric data; dimensions, overall shape and size were consistent with current ISO buttock models for wheelchair testing (ISO, 2007). It has an internal rigid substructure with ischial spacing of 11cm (4.33 inch) and bi trochanteric breadth of 36 cm (14.17 inch). This represents the major bony prominences in the buttocks: the ischial tuberosity and trochanter. The indenter was also designed symmetrically so measurements made on one side correlate with measurements made on the other side. The indenter has an outer elastomeric shell that mimics tissue. Two choices were considered for the outer shell - spherical and ellipsoid. The ellipsoid design was chosen because it represented the contour of a seated buttock better than a spherical shell.

METHOD

Compliant Cushion Loading Indenter (CCLI)

 

CCLI was built based on the anthropometric data of a person that has 36cm (14.17-inch) distance in between the trochanters and 11cm (4.33-inch) spacing between the ischial tuberosities(IT).The model has a rigid internal substructure with an outer elastomeric shell. The internal rigid substructure is symmetrical and made out of wood. It consists of two IT’s, two trochanters and a backing plate. The IT is made of wood. It is a cylindrical structure 65mm (2.56-inch) long, 50.8mm (2-inch) diameter with a 50.8mm (2-inch) curve at the end to accommodate transducers. The trochanter is rectangular: 28.58mm (1.12-inch) by 50.8mm (2-inch) wide and 25mm (0.9-inch) long. The IT and trochanters are attached to a back plate made of plywood 19.05mm (0.75-inch) thick. The back plate has extra holes drilled on the right (ultrasound) side for an array of ultrasound transducers (Figure 1). Transducers were placed on both sides. On one side, pressure transducers were placed at the medial and lateral prominences. On the other side, ultrasound transducers were placed as shown in Figure 2. The ultrasound transducers are designated by their distance from the midline. In order to enhance the reflected ultrasonic signal at off-normal incidence angles, the midline of the outer shell was textured with acrylic beads embedded near the surface in a dense array. A 25.4mm (1-inch) wide, 304.8mm (12-inch) long and 2mm (0.07-inch) thick strip was made. It was filled with 1.59 mm (0.06 in) acrylic beads and elastomer with density of 256 beads per square inch. The elastomer was made using Dragon Skin FX Pro (Smooth On Inc., Easton PA) with 90% thinner such that ASTM D575 test would yield a material stiffness of 3.14 Nmm-1 .304.8mm (12 inch) long cylindrical samples with diameter of 25.4 mm (1 inch) were also made to measure the speed of sound in the elastomer which was found to be 980m/s. The model was made such that the offset between the IT of the substructure and the outer shell is 20mm (0.78-inch). The entire CCLI is attached to a loading rig (Zwick / Roell Z005, Kennesaw GA).

Instrumentation and Control

Pressure transducers STS TD10 (STS Sensors, Danbury, CT) are placed in the IT and trochanter. Before testing, the sensors were calibrated against a range of known pressure values. The pressure transducers are connected to a bridge circuit which is in-turn connected to a SCXI-1000 Chassis (National Instruments, Austin, TX). The voltages from SCXI chassis is collected by a PCI-6034E data acquisition card (National Instruments, Austin, TX). The ultrasound sensors used are 5MHz, 12.8mm (0.5-inch) unfocused transducers (V309-SU, Olympus Inc, Waltham, MA) at every location except trochanter. At the trochanter, a 5MHz, 12.8mm (0.5-inch) focused transducer with focal length of 25.4mm (1-inch) (V309-SU-1.00, Olympus Inc, Waltham, MA) is used. The ultrasound transducers are connected to multiplexer whose input comes from a pitch catch receiver (DPR 300, Imaginant Inc, Pittsford, NY). The signal is switched by controlling relays via USB (1017_1 – PhidgetsInterfaceKit 0/0/8, Phidgets Inc, Alberta, Canada) which toggles a 12VDC signal to a switch (Magnecraft 782XBXM4L-12D, Schneider Electric, Des Plaines, IL).The ultrasound signal is acquired using a NI-USB 5132 DAQ (National Instruments, Austin, TX) and is averaged 128 times to improve the signal to noise ratio. The control and data acquisition algorithm is implemented in LabVIEW (National Instruments, Austin, TX).

Procedures

Two cushions were selected a HR50 foam (Hibco Plastics, Inc., Yadkinville, NC) and a Jay 3 (Sunrise Medical, West Midlands, UK) to assess the variability and sensitivity of the indenter. To insure a difference in performance a simple (flat HR50, PDAC HCPCS coding: E2601) and a complex (Jay3, PDAC HCPCS coding: E2624) cushion were chosen. Two loads were applied; 44kg and 53kg representing the average upper body mass of a 70kg and 83kg person respectively. Ten measurements were made on each cushion for each load. Before every load, an unloaded measure was also taken. The cushions were preconditioned before the test according to ISO 16840-2 protocol. The ambient conditions were maintained at 23°C ± 2°C (73.4°F ± 3.6°F) and relative humidity at 50% ± 5% according to ISO protocol.

Before the cushion tests, unloaded pressure at the medial and lateral prominences and elastomer thicknesses at various locations were measured. The CCLI was lowered and the desired weight was applied for 180s. After 120s, pressure, base to model height (from loading the rig) and elastomer thickness at various locations were recorded. The cushion was unloaded for 120s between measurements. The Jay cushion was kneaded back into its original position after every measurement.

Data Analysis

Repeated measures ANOVA will be used to evaluate the effect of load and cushion on pressure and deformation. The pressure data will be a difference of loaded and unloaded pressure. The deformation is calculated as the difference of the unloaded and loaded elastomer thickness. A typical result of elastomer thickness at any location is found by the following steps. Find the first ultrasound reflection. This corresponds to a time t. The speed of sound in elastomer is known to be 980m/s. Then elastomer thickness (d) can be calculated by the formula, d (in mm) is equal to 980 (speed of sound in elastomer in m/s) times t in microseconds divided by 2 (since it takes twice the time for the signal to travel both back and forth) times 1000 (this will convert the microseconds to milliseconds and the result will be in mm). When calculated, d is equal to 0.49 times t (Equation 1).

d = 0.49t                                                              (Equation 1)

RESULTS

The pressure, deformation and model’s height to the base data is tabulated in Table 1. The unloaded pressure and elastomer thickness data had a coefficient of variance less than 2%. Repeated measures ANOVA showed significant difference in cushion and load for all cases except, the load in lateral prominence under deformation which had an outlier (removing the outlier showed significant difference) and cushion at 129mm under deformation which showed no significant difference across cushions. There was a significant difference due to cushion and load in pressure and model’s height from the base.

DISCUSSION

The data suggests a high level of repeatability across loaded and unloaded conditions for different cushions and loads. There is also sensitivity with respect to the interface and the load in most cases.  The unloaded pressure had a coefficient of variation under 2%. The unloaded elastomer thickness COV was under 1%. Since the data is repeatable, unloaded data may be taken only once before the study. The highest pressure 99.8 mmHg was seen on the foam cushion at higher loads on medial prominence whereas; the lowest pressure on medial prominence of 60.3 mmHg was seen in the Jay cushion. The highest pressure on the lateral prominence of 58.2 mmHg was seen on the Jay cushion at the highest load; lowest pressure on the lateral prominence was seen on the HR50 foam cushion. This is because, in the foam, the medial prominence is loaded first then the lateral prominence is loaded. But in Jay, the medial prominence is loaded on a viscoelastic fluid while the lateral prominence shares the load. This is proved by the elastomer thickness under the lateral prominence which was higher in Jay compared to HR50. This also validates the indenter sensitivity, and variability under load.

The pressure difference between medial prominence and lateral prominence was larger in HR50 compared to Jay. The deformation in HR50 had a smaller range than the corresponding deformation under Jay cushion. This is reflected in the standard deviation of the deformation in Table 1. Deformations in Jay show more of a viscoelastic nature than HR50, which was expected and precisely why these cushions were chosen. The 129mm location is the second highest deformation zone in the Jay cushion and highest deformation zone in the HR50 (Table 1).  The highest COV in deformation was in HR50 lateral prominence under low load and the lowest was below the 22mm location under load. This can be attributed to the low shear modulus of the elastomeric material, which bulges to the side under load. The ultrasound transducers are fully capable of measuring the thickness as stated earlier; the COV was less than 2%.

Study Limitations

This study used only two cushions and only 2 loads, so the results may not be generalized. The reason for choosing so was because the indenter had to be validated and assessed for sensitivity, repeatability and variability. Also the location of the transducers will affect the result, placing them elsewhere would provide different results than those shown here. Further study is needed to evaluate the effect of the indenter on different cushions.

CONCLUSION

This paper presents a method to design and develop a compliant cushion loading indenter (CCLI) capable of measuring internal loads and deformation of the buttocks. The CCLI was validated using two cushions and two loads and found to be sensitive and repeatable. A few minor, yet important issues like changes in the unloaded load and deformation were investigated and found to be repeatable. In future studies, unloaded data can be taken before the study. The aim of the study was to validate the model hence only two cushions were chosen. CCLI can be used by clinicians and manufactures as a tool to evaluate wheelchair cushions based on internal load and deformation.