Final project update!

Long time no post......

Since we last posted, we got a LOT of new components and did a lot of initial testing:

(Instagram edition!)

SD/microSD shield + accelerometer 1 

SD breakout board + accelerometer 2 

Lilypad! We may or may not use it, but since this device has to be portable, we will probably turn to this eventually.

Attempting to combine the RCtime setup from Lab 2 with the accelerometer (it didn't work... we decided to use the analog pins on the Arduino in the end)

Dual accelerometer testing!

Dual accelerometer + piezoelectric sensor testing! We decided to try using the piezoelectric sensor as a switch for the time being, but we are considering other options to better fulfill the sensing needs of this device as well.

Insoles version 1!

Our device in progress

Winslow wrote all of this :( 

At least the quality of our blog photos is somewhat improving.

More importantly, though, we thought we should update you a little on where we're heading with this project and explain where this device fits in... in the context of the current market and such.

DEVICE NEED

We want to address a medical condition that is both a symptom of underlying disease and a problem in itself. Abnormal walking behavior, or gait, is a consequence of distal sensory and motor loss due to nerve damage, or neuropathy. Distal nerve damage is specifically called peripheral neuropathy, which, statistically speaking, is primarily attributable to conditions such as the following: 

     1. Diabetes mellitus (diabetes)- Progression of diabetes is marked by high blood sugar levels, which degrade the myelin sheaths of nerve cells over time; furthermore, diabetes causes chronic kidney disease and failure, which also contributes to irreversible peripheral nerve cell damage (Krishnan and Arun, 2009).
     2. Alcoholism- Chronic alcohol abuse is strongly correlated with the incidence of chronic liver disease, which has been found to be directly associated with distal sensory loss related to peripheral neuropathy (Chaudhry et al., 1999). 

Abnormal gait patterns are also seen in children and adults with autism-spectrum disorders (Shetreat-Klein et al., 2012). Even children who exhibit abnormal gait habits without any underlying medical condition eventually develop postural and gait deformities that introduce medical and, later, social problems; these cases are also difficult to identify because researchers found that children can consciously modify their walking behavior in clinical settings (Armand and Watelain, 2007; Pendharkar et al, 2008; and Wren, 2004).

Gait abnormality is a medical condition in and of itself, but due to the above associations, it can also be considered an indicator of serious underlying issues or future concerns that mandate early identification and treatment. Severe cases of damage due to the long-term effects of gait problems are major reasons for lower extremity amputations. In 2011, 65,700 lower-limb amputations were carried out on diabetes patients alone because of complications due to nerve damage (Centers for Disease Control and Prevention [CDC], 2011)! 

Beginning in 1989, medical literature reviews acknowledged that early identification and preventative treatment of patients at risk could reduce amputation rates in large clinical centers by as much as 85%; in the two decades since, several technologies have been presented in the literature as accurate quantitative diagnostic tools, but none has been widely implemented for clinical purposes (Bild et al.).

The common initial method of diagnosis for peripheral neuropathy is still a subjective pressure test using a monofilament rod and a vibration test using a tuning fork that require some significant existing damage for a positive result (National Institutes of Health [NIH], 2009). In other words, first they prod your foot in various places with a pin, and then they hold a tuning fork to it to see what frequencies you can feel. Subsequent screens involve nerve conduction studies or expensive blood work (NIH, 2009). 

What about detection methods for abnormal (but non-pathological) gait habits that could result in deformity later on? We found that no device that fulfills this need has been commercialized. There is a clear gap in both the market and the literature for a quantitative diagnostic tool that answers all of the concerns described above while being low-cost and convenient to use. A pressure-sensing device that monitors the sole of the foot could detect minor gait abnormalities that would help diagnose underlying neuropathy. It could also direct otherwise healthy patients to corrective therapy for harmful gait habits; this type of device could thus fill a current gap in diagnostics.

We figure we should should probably also give an idea of how large our target population is. 

If you add up the segments of the population with

• diabetes (both types 1 and 2), 
• children ages 10 to 15 (a major risk group for skeletal deformities and gait dysfunction, since this is the time they go through their final growth spurts), and 
• patients with chronic liver and kidney disease, alcoholism, and other related high-risk conditions for neuropathy,

and subtract accordingly to account for the estimated segments that have combinations of these conditions, we find that the target population could range between 50 and 60 million individuals.

The population would use our device at home or wherever their usual daily schedule takes them, since that is the type of monitoring we are trying to achieve. Generally speaking, they would not require others to help them in the use of the device, since they simply turn it on and off with the push of a button.

ALTERNATIVE DEVICES AND TECHNOLOGIES

What about current alternative devices? If there are alternative devices out there, why are clinics still using pins and tuning forks? Sometimes such simple diagnoses are "good enough" for a quick clinical screen, but what about the children with potentially harmful gait habits but no neuropathy or other pathologies? How can we tell if their habits cause potential harm in the long run (indicated by causing significant changes in the pressure distribution in their feet), especially if they tend to consciously change their behaviors when they are in the clinic?

We found three major competing devices and technologies:

1. TekScan F-Scan system. This is a system very similar in concept to the device we have in mind: TekScan offers both stationary and wearable (portable) versions of their system, and these generate high-resolution pressure data for the soles of the wearer's feet. For reference, one insert contains 960 sensors. These system packages are priced on the order of $4,995 each. TekScan specializes in both the pressure/force-sensing sensors themselves and the application of these sensors to consumer products. 
2. Sensor Products, Inc. (SPI) pressure-sensitive film. This is a precision-engineered film 4-8 mm thick that develops gradients of color based on pressure distribution. TekScan has a similar system that they also offer for diagnostic use. SPI markets themselves as "tactile pressure experts," with collaborations with companies such as Fujifilm to develop a variety of precision films (their core technology) for pressure analysis.
   The problem with the film products is that they give no data with respect to time, which we are interested in. We only see the final color distribution (which is mapped to a pressure distribution using the accompanying software and scanning package). If there were unexpected bumps or other accidents that caused a "spike" of pressure on the film somewhere, we would have a hard time telling if it was due to their actual gait or if it is just an irrelevant artifact.
   Furthermore, these films have problems with temperature and humidity sensitivity: SPI's film in particular has a maximum temperature cap of about 95 degrees Fahrenheit and a humidity cap of 90%. However when people wear shoes for extended periods of time, the local temperature in the shoe can reach up to 115 degrees and experience 100% humidity. Special coatings or casings that do not interfere with pressure exertion on the film itself would be required for these films.
3. RsScan Lab, Ltd. This company develops systems similar to TekScan's piezoelectric crystal-based product line, with more emphasis on a very high density of sensors on flat portable mats. Their 2m x 1m mat boasts over 16,000 sensors. The cost of these devices is consequently higher, on the order of $8,000 to $15,000 for a mat and the accompanying hardware and software. RsScan Lat differs from TekScan in the breadth of their offerings: currently their primary product is the Footscan system, and they do not manufacture and distribute sensors the way TekScan does. They do offer a downstream product: they manufacture custom inserts for users who provide them with Footscan data. 
4. MIT Media lab. In 2002, a group at the Media Lab released a paper detailing a pressure and orientation sensor shoe very similar to the insert system we are developing. However, the final product was hardware-heavy (it had a large hardware component sticking out the back of the shoe), and we did not like the fact that the wearer could not wear his/her shoes of choice. Furthermore, it would be inconvenient to wear the shoes they developed in public while the user goes about their daily activities, which is necessary for the kind of monitoring that we have in mind. 

Essentially, the biggest hindrances presented by the options available now are wearability and price.

In terms of desired outcome, our insert will be very thin--definitely less than 1 cm thick--and seek to be as unnoticeable as possible to both the wearer and people around them. Then data could be recorded as they go about their usual routine (or, if desired, during a specific athletic activity, etc.) and more accurately reflect their behavior than a brief walking or running test in a clinic (this is especially useful for diagnosing children, in this case). We want these inserts to be low-cost enough when scaled-up so that clinics could have many pairs in various sizes at their disposal (we're aiming for <$100/pair in the end).

Sorry this post has been so long so far! We just have a lot of progress to catch you up on, haha...

Reverse Engineering

While we're at it (sorry!), here is a breakdown of one of the major competing devices in the market today:

F-SCAN SYSTEM BY TEKSCAN, INC.

Tekscan has created a variety of force sensors and pressure mapping applications including the F-Scan System, an in-shoe plantar pressure analysis device. This system offers both wireless and wired recording of pressure mapping of the feet. Data is logged into a separate compartment worn at the waist or transmitted via a RF signal to a computer with a receiver. This pressure mapping data is then analyzed and visually shown on custom computer software developed by Tekscan. 

We could not get a physical device to take apart because these systems cost approximately $4,995 each. Instead, we looked for patents related to the device to get an idea of what was inside.

This technology is referenced in US 7,758,523 B2 patent created by Kineteks Corporation for a shoe insert apparatus that can function remotely. Transmitters are placed in the lower portion of the sole insert in a cavity located under the arch of the foot. The holding cavity could be molded into an arch support, which would then contain a circuit board. This circuit board contains a battery for power, the transmitter for sending RF signals to a receiver, a receiver to receive RF signals from a remote transmitter or from the transmitter located on the same circuit board, a sampling circuit to test the signals from the force sensors, and a conversion circuit to convert the signals to be suitable to be sent through a RF signal. Data logging capability through integrated memory may also be included in the final design. 

Force sensors are placed almost symmetrically along the transverse and longitudinal axes of the foot with respect to the transmitter cavity. Additional sensors are placed with particular attention to the metatarsophalageal joint supports and the first toe.

Figure 1. Patent and product images for competing device from TekScan. Left: Depiction of the top view of the device stipulated in Patent US 7,758,523 B21. As mentioned it depicts the position of the cavity insert and the force sensors to be used with the device. Middle: Depiction of the F-Scan System data analysis for a case study of diabetic ulcer treatment using this system2. Right: Data analysis of the movement of the foot coupled with pressure distribution mapping of the RSscan system3. 

This technology seems to be integrated not only into Tekscan's F-Scan System but also in RSscan International's RSscan products. They both offer very similar functionality, but RSscan demonstrates a more advanced user interface with additional data analysis capabilities. In addition, the F-Scan System uses their own force sensors that they have developed and are selling separately as well. The F-Scan System seems to fit into TekScan's business strategy quite well: with many of their systems, they are  essentially developing applications for their own sensors. They have a market advantage in many respects by doing so, too, since they control the quality and supply of the systems' main components--a clear form of vertical integration. While we don't have that kind of capability, we find through the online reviews of numerous F-Scan users (easily found via Google) that it is considered to be feature-heavy to a fault in some cases: it is designed to satisfy a wide range of needs, but this in turn raises the unit cost, and physicians in podiatry clinics remark that they only use the most basic functions. If we could develop a device that focuses only on all of the most commonly used features of products like the F-Scan, we could significantly reduce the cost to physicians (and therefore, generally, patients).

Our device will use a similar technology with a wired system using force sensors at similar locations emphasizing the first toe, the metatarsophalageal joint supports, and the heel. This pressure mapping data is routed to a central device worn at the hip and processed into a data string to be incorporated into a text file on a SD card. These processed signals are then uploaded onto a computer to be analyzed for both pressure distribution mapping with respect to time and foot orientation as the foot moves through space with respect to time. We thought that the incorporation of the circuitry into the arch would be nice, but that would mean we also need alternative versions for, say, people who are flat-footed, for comfort reasons. Thus we decided to just incorporate all of the circuitry into the belt-clipped portion of our system.

In addition, cost is a major parameter that we want to address in our own device: having such a product priced at $4,995 makes this kind of system rather impractical for smaller-scale clinical use. As such, we found that it is mostly reserved for athletes at the Olympic level and large podiatry clinics. We would like to design one with lower-cost but high-performance sensors that would allow many pairs to be available at a time, even in small clinical settings.

References:

Armand, S., and Watelain, E. (2007). Linking clinical measurements and kinematics gait patterns of toe walking using fuzzy decision trees. Gait and Posture, 25 (3):p.475-484.

Arlett, J. (n.d.). Case study: Diabetic ulcer treated with f-scan. Retrieved from  http://www.tekscan.com/diabetic-ulcer-charcot-foot

Bild, D.E., Selby, J.V., Sinnock, P., Browner, W.S., Braveman, P., Showstack, J.A. (1989). Lower-Extremity Amputation in People with Diabetes: Epidemiology and Prevention. Diabetes Care, 12, 24-31.

Center for Chronic Disease Prevention and Health Promotion. (2011) Diabetes Fact Sheet: national estimates and general information diabetes and prediabetes in the United States, 2011. Atlanta, GA: Centers for Disease Control and Prevention.

Chaudhry, V., Corse, A.M., O'Brian, R., Cornblath,D.R., Klein, A.S., and Thuluvath, P.J. (1999). Autonomic and Peripheral (sensorimotor) Neuropathy in Chronic Liver Disease: A Clinical and Electrophysiologic Study. Hepatology, 29.6, 1698-703.

 Clinics: Orthopedic surgeons, orthopedic technicians, podiatry. (2012, May 04). Retrieved from http://www.rsscan.com/

Collings, T. D., & Younker, J. R. United States Patent, (2010). Remote sensing shoe insert apparatus, method and system (US 7,758,523 B2)

Elliott, M.T., Ma, X., and Brett, P.N. (2009). A smart sensing platform for the classification of ambulatory patterns. Proceedings of the Institution of Mechanical Engineers, part H: Journal of Engineering in Medicne. DOI: 10.1243/09544119JEIM523

Krishnhan, K. and Arun, M. (2009). Neurological complications of chronic kidney disease. Nature Reviews Neurology, 5, 542-551.

Low, D.C. and Dixon, S.J. (2010). Footscan pressure insoles: Accuracy and reliability of force and pressure measurements in running. Gait and Posture, 32 (4): 664-666.

Morris, S.J. and Paradiso, J.A. A compact wearable sensor package for clinical gait monitoring. Offspring Vol. 1, No. 1, pp. 7-15.

National Institutes of Health. National Institute of Diabetes and Digestive and Kidney Diseases. (2009). Diabetic Neuropathies: The Nerve Damage of Diabetes. National Diabetes InformationClearinghouse (NDIC), Accessed 1 May 2012. <http://diabetes.niddk.nih.gov/dm/pubs/neuropathies/>.

National Institutes of Health. National Institute of Diabetes and Digestive and Kidney Diseases. (2011).

Chronic Kidney Disease and Kidney Failure: Fact Sheet. Research Portfolio Online Reporting Tools. Accessed 1 May 2012. <http://report.nih.gov/nihfactsheets/ViewFactSheet.aspx?csid=34>

National Institutes of Health. National Institute on Alcohol Abuse and Alcoholism (2012). Fact sheet for the general public. Accessed 1 May 2012. <http://www.niaaa.nih.gov/FAQs/General-English/Pages/default.aspx>

Pendharkar, G., Lai, D.T.H., and Begg, R.K. (2008). Detecting Idiopathic Toe-Walking gait pattern form normal gait pattern using heel accelerometry data and Support Vector Machines. 30^th Annual International IEEE EMBS Conference, 4920-4923.

RsScan Lab, Ltd. Footscan^R Advanced Pressure Measurement Systems: Systems. Retrieved May 1, 2012, from: http://www.rsscan.co.uk/systems.php

Sensor Products, Inc. (2012). Fujifilm Prescale Tactile Pressure Indicating Sensor Film. Retrieved April 22, 2012, from: http://www.sensorprod.com/static/lamination.php

Shetreat-Klein, M., Shinnar, S., and Rapin, I. (2012). Abnormalities of joint mobility and gait in children with autism spectrum disorders. Brain and Development, doi:10.1016/j.braindev.2012.02.005.

TekScan, Inc. (2007). PreScale^TM Digital Analysis Systems. Retrieved April 18, 2012, from: http://www.tekscan.com/prescale-digital-analysis-system

United States Census Bureau. (2010). 2010 Census Data. Accessed 1 May 2012. < http://2010.census.gov/2010 census/data/>

Valmassy, R. (1996)  Clinical Biomechanics of the Lower Extremities, Missouri  Mosby. Pp 249, 250.

Wren, T. (2004). Gastrocnemiussoleus lengths in cerebral palsy equinus gait-differences between children with and without static contracture and effects of gastrocnemius recession. Journal of Biomechanics, 37(9): 1321-7.