How do quantity, frequency and alignment of sensors affect the results of plantar pressure distribution measurement?

P.W. Brand coined the statement “Pressure is the critical quantity that determines the harm done by the force” meaning that not the forces acting on the foot primarily cause an overload, but the pressure which is defined as the force acting on a certain area. Therefore, pressure distribution measurements indicate how the foot is loaded during ground contact and in which locations high, possibly even damaging pressure values may occur and cause pain. The measurements may provide an objective evaluation of the foot function under dynamic loading conditions. Therefore, plantar pressure distribution measurements have been established in the last years as clinically accepted procedures for the functional examination of foot problems caused by diseases, injuries or deformities. Furthermore, they have developed into a diversely used research tool for detailed examinations of foot biomechanics.

A literature search, with a combination of respective keywords (Pedograph* OR pedobarograph* OR “plantar pressure” OR “foot pressure” OR baropodo*) provided 2121 entries in the Medline data bank or 3768 entries in Scopus (as per September 2016), reflecting the distribution and wide-ranging use of these examination methods. This development started in the 70s, picked up speed between 1990 and 2000 and reached a further boost in the last 10 years (figure 1). Unfortunately this literature search also shows a common problem:

So far, there has not been a uniform denomination for the measuring method, so that relevant articles will be found under different key words. In Germany the term plantar pressure distribution measurement is commonly used, whereas internationally the term pedobarography (or pedography) has prevailed, but there are also regionally specific terms such as baropodometry in Italy. A closer look at the publication titles reveals a wide range of applications that range from “classic” clinical topics such as diabetic and rheumatoid foot problems as well as foot and toe deformities (hallux valgus, pes planovalgus, pes cavus, etc.) to treatments with shoes and foot orthotics all the way to sports-related issues.

Sensor arrangement

The number of yearly entries in SCOPUS on the topic “plantar pressure distribution measurements” has continuously risen in the last four decades and reached a peak in the year 2014 with 323 entriesWith the increasing application of pressure distribution measurements as a research method more measurement systems became available. Thus, a potential user of a respective system should get informed sufficiently beforehand, in order to purchase a suitable device based on his individual requirements. One should be aware of some general differences that are due to using a pressure distribution platform as opposed to an in-shoe-system with measuring insoles.

Pressure distribution platforms are usually used for measurements in a lab situation, where they should be embedded in the floor or in a walkway in order to enable natural walking during the measurements. The application is usually limited to barefoot measurements, since pressure distribution patterns for measurements with shoes are influenced by the sole design of the shoes. That means that the foot function is examined independently from the shoe conditions. Usually only one footprint is taken per measurement (unless the more recent, larger platforms are used, that can record two to three ground contacts), so that repeated measurements have to be carried out subsequently in order to collect the desired number of trials.

On the other hand, there are in-shoe measurements with insoles that are also suitable for lab-independent measurements in the clinical environment (examination room, therapy area, clinic hallways) or even for field measurements for example in case of sports activities. They have the advantage that a larger number of steps can be stored and subsequently analyzed in one measurement. Thus the measurement between sole of the foot and the shoe or insole provides information on the effect of shoe design or modification or treatment with insoles.

For in-shoe measurements it should also be distinguished whether the complete sole is used as an active measuring surface or if individual sensors are used to describe the loading characteristics. In case of a sole measuring the entire foot contact area, a complete description of the load characteristics is possible.
With individual sensors on the other hand the sensors are arranged in a way that the regions are evaluated where the main loads are expected.

Schematic depiction of a measuring sole with a complete sole measuring surface (left) compared to a sole with individual sensors (example with 16 sensors). It illustrates that not all anatomic regions can be completely depicted with individual sensors. In extremely deformed feet this discrepancy possibly is even bigger, since the sensors cannot sufficiently record the main loading areas, for example in case of an extreme planovalgus foot or a club footIn case of a normal foot shape a more or less comparable ground reaction force can thus be evaluated or reconstructed by summarizing the information from the individual sensors. Therefore, the values of individual sensors are added resulting in a force curve that is similar to a double-peak curve of the vertical ground reaction force.

However, individual sensor systems are limited in their relevance since in case of extreme foot deformities highly loaded foot regions may not or not completely be represented (picture 2). In case of diabetic feet local peak pressure points might not be detected in case of measuring soles with individual sensors that do not cover the complete foot sole.

If the pressure peak is located between the sensors, no pressure will be indicated at all, depending on the size of the space between sensors. If the pressure peak lies at the edge of the sensor, it can only partly detect the pressure and thus indicate a value that is too low. Especially regions at risk  for ulceration could be overlooked during measurements and thus not sufficiently protected against overloading with appropriate treatment. 

Sensor and measuring frequency

Independent of the sensor arrangement there are critical differences resulting from the technical specifications of the measurement systems such as overall size, the active sensor area, the arrangement and size of the individual sensors as well as the sensitivity and the measuring range. The size of the sensor area determines how difficult or how simple it is to record the complete foot print without changing the approach. With smaller platforms there is accordingly less margin of error in case of larger feet, so that the approach should be measured exactly.

Apart from the total size of the measurement area an important aspect is the size of the individual sensors or the spatial resolution. In order to guarantee a sufficiently detailed representation of plantar pressure distribution beneath the human foot, a sensor frequency of 4 sensors per cm2 is necessary (Davis et al., 1996). This corresponds to a squared sensor with a size of 5x5 mm. With larger sensors there is a risk of underestimating the actual pressure beneath an anatomical structure.

However, a higher resolution of 9 sensors per cm2 did not provide additional information even for the smaller feet of children (Rosenbaum & Lorei, 2003), so that the resolution of 4 sensors per cm2 appears sufficient. In case of a lower sensor spatial resolution, i.e. larger sensors there is the problem with detecting lower pressure values for local pressure peaks  due to an averaging across the sensor surface (picture 3). Therefore, care should be taken when comparing pressure values that were recorded with different measurement systems.

The values are system-specific and cannot or can only partly be compared between different measurement systems.
Furthermore, a lower sampling frequency or scan rate can lead to an underestimating of actual peak pressure values. If the pressure only momentarily reaches a peak value, the sensor has to be scanned briefly in a corresponding time interval, in order to detect this pressure. During a longer measuring interval it is also detected over a period of time and thus a lower pressure value is indicated. For the load changes occurring during walking, sampling frequencies of 50 Hz (50 samples per second) should be sufficient, but during sports like fast running or jumping and landing this can be too slow to detect rapidly changing loads.

The sensor frequency influences the depiction of the maximum pressure value: If a locally limited pressure peak, for example beneath a metatarsal head (here symbolically depicted with a black circle) bears down on a little sensor of 5x5 mm almost full-facedly, the actual pressure value of 8 is depicted. In case of a double 4-fold sensor size, it is averaged over the pressure ­values of the values of the 4 previously depicted sensors, reducing the pressure to 5. If you ­further increase the sensor, the mean value is further reduced to 2, so to speak to a fourth of  the initial value, so that the actual effective pressure continues to be underestimated with increasing sensor size and reduced spatial resolutionThis should help to understand that also thresholds for avoiding damages to the foot are system-specific and are not generally valid for all measurement systems. This has to be taken into consideration mainly for the treatment of feet with sensitivity disorders/neuropathy such as Diabetes or the Charcot-Marie-Tooth disease, because of the unreliable feedback from patients. That is why no absolute values have been defined as targets for an effective treatment with therapeutic shoes and/or orthotics of
diabetic feet, but a relative degree of unloading is required according to The German Diabetic Society which requires a 30-per cent pressure reduction by soft-cushioning and that shall be proven with the help of pressure distribution measurements.

Some advice for carrying out measurements

Different methods have been proposed for approaching the measurement platform: The “full-gait” or the “mid-gait” methods record ground contact at normal, constant gait speed for which an approach of at least 4 – 5 steps is necessary. Furthermore, after the measurement some steps should be allowed in order to avoid slowing down during the measurements.

On the other hand, the “first-step” method was described and recommended as a reliable method for patients with pronounced pain or gait disorders in order to obtain enough valid measurements without overburdening them. For this method the subject is standing at a short distance in front of the measurement platform, so that a complete ground contact can be recorded with the first step. Of course, such a roll-over process cannot be compared to full-gait due to the lower speed and the subject still accelerating. Therefore, lower pressure values were reported with the first-step method (Morlock & Mittlmaier 1992). In case of comparing repeated measurements the same method should be used in order to achieve comparable data.

As an alternative method, measurements with the second or third step on the sensor area are used and recommended. Even though all methods are described as valid (Bus et al. 2005), they should not be considered comparable. The second-step method has been recommended especially for the examination of diabetic patients, since it caused the least number of repeated measurements and thus minimized a burden for this patient group (Bus et al. 2005).

The aim is to obtain sufficient repeated measurements for further evaluation, in order to consider the inherent variability of individual gait cycles and still to get a realistic picture of plantar loading patterns of an individual patient. Usually five repeated measurements are carried out, however a minimum of three measurements have been recommended before. With five measurements per appointment a good reliability could be proven for the emed platform in a one-week period (Gurney et al. 2008). However, the reliability values are not generally valid and transferable to other measuring systems, instead they should be considered as valid for the specific measuring system and the described measurement and evaluation protocol.

With in-shoe-measurements the number of repeated measurements plays a lesser role, because during the measurements in the shoe all steps are recorded during the measurement process. So a sufficiently long measuring distance (minimum 10 meters) provides enough measurements during fluid walking after a few initial steps and allows for registering 5 to 10 steps. Furthermore, it should be ensured that no steps are recorded during the braking at the end of the walking distance or when turning around for the way back, since these steps would not show comparable loading characteristics.

The reliability of repeated measurements supports the good applicability of pedobarography in daily clinical practice, when patients often are examined in repeated appointments (i.e. before and after an intervention, a rehabilitation program or surgery). This way, differences between repeated assessments can be interpreted as showing actual clinical differences and are not caused by lacking repeatability of measurements.

Finally it should be noted that the gait speed should be considered during measurements. Various examinations revealed that the load pattern changes with increasing speed with higher pressure values in most regions, but also an increased medial loading in the forefoot can be detected (Rosenbaum et al. 1994). Even  though the importance of the speed of gait is generally acknowledged, it does not need to be prescribed or dictated by using a metronome, since this might cause an unnatural gait pattern.

However, the gait speed should be kept constant during an appointment which is usually easily achieved after a “warm-up” phase and is sufficiently controlled by measuring the approach distance. Marked variations in the gait speed during the measurements should therefore be avoided. In repeated visits a comparable gait speed can be aspired as long as the current pain situation does not contradict that aim. Alternatively, changes in gait speed could indicate that the patient’s situation has improved or gotten worse.  If measurements with different speeds are compared, the potential pressure differences should be considered.

Summary and outlook

Pedobarography nowadays is a well-established method and different measurement systems are available with technical solutions that can be selected according to the specific research questions and interests. Thus, the application of plantar pressure distribution measurements is not limited to specialized centers, but they are increasingly distributed in clinics and medical practices that are focusing on the foot, as well as in prosthetics and orthotics or pedorthic companies. This way, additional functionally important information is provided to many patients and their practitioners exceeding the normal clinical and radiographic data.

Yet, the interpretation of the results still requires a certain amount of experience and expertise, in order to view and filter the amount of information and to draw the possible conclusions for influencing or changing the further treatment strategy. The existing body of literature can help as a background, but there seems to be no way around an intensive and regular analysis of technical options and limitations. Future research projects should help to identify, substantiate and validate the most informative analyses and parameters, in order to help the less experienced user in selecting a suitable approach.

Address of the author:

Dieter Rosenbaum
Human Performance Lab
Descente Korea LTD
Descente Shoes R&D Center
162, Myeongjigukje 6-ro
Gangseo-gu, Busan, Republic of Korea