INTELLIGENT DEVICE TO MONITOR AND REMIND PATIENTS WITH FOOTWEAR, WALKING AIDS, BRACES, OR ORTHOTICS

[0001] Not Applicable

[0002] Not Applicable

[0003] This invention generally relates to body movement monitoring systems, and specifically to an ambulatory system which monitors whether the user is wearing and/or using his or her prescribed footwear, walking aid, braces (including knee and ankle braces) or other orthotics. The present invention 1) identifies the parameters related to the user's postures and movements (e.g., standing, walking, and running) and classifies the user's posture and foot loading condition using a body-worn sensor; 2) identifies whether the body-worn sensor is worn by the user; 3) detects whether the user is using any other tagged devices (such as prescribed footwear, walking aids, braces and orthotics) during foot loading conditions (e.g., walking or standing longer than 10 seconds) or specific postures; 4) in one embodiment, alerts the subject and/or the subject's caregivers if the user ambulates without a specified tagged device (e.g., without using the prescribed footwear or other walking aid); and 5) identifies the user's posture during time periods over which specified footwear are being worn or carried by the user.

[0004] Objective, novel, and cost-effective approaches for assessing and improving patient adherence to treatment are in high demand. Of particular interest are methods for quantifying athe patient's adherence to (and preference for) wearing therapeutic and prescribed footwear and offloading devices (collectively, “therapeutic footwear.”) An exemplary use of such technology is for monitoring diabetic patients who have developed neuropathic foot ulcers and who must wear therapeutic footwear as the most important part of their healing regimen. In the field of diabetic foot ulcer prevention, knowledge of patient preference for specific types of footwear will aid developers improve the design of customized footwear, thus improving healing

[0005] Patient adherence to wearing their therapeutic footwear can significantly enhance and improve healing. Although laboratory studies using total contact insoles and rocker-sole shoes have demonstrated modest reductions in pressure, these clinical footwear trials are inconclusive. Within 12 months, 26-42% of the studied patients had re-ulcerated. These results are likely due to lack of patient adherence to footwear. For example, despite taking a significant number portion of their daily steps at home (some studies have estimated this fraction to be over 50%), patients view their homes as “safe zone” where they typically do not wear prescribed footwear. As a result, high-risk patients wear their therapeutic footwear not as often as they should (some studies estimate that such patients wear the required footwear only 15-28% of the time).

[0006] What exacerbates the problem faced by such patients is that the methods used studies quantify patient adherence likely underestimate the patients' non-adherence to a significant degree. Adherence is typically measured through face-to-face interviews with a small number of patients. Previous investigations have shown that patients under-report “sensitive” conditions in face-to-face interviews versus telephone or paper surveys.

[0007] The present invention consists of a system for monitoring, assessing and improving patient adherence to any prescribed footwear.

[0008] Applied to diabetic care, the present invention can monitor patient adherence to instructions provided by caregivers—specifically for using a cast boot or other prescribed offloading devices and therapeutic footwear during walking, and/or foot loading condition. The invention can also remind and/or alert the patient if he or she forgets or neglects to use the prescribed footwear during standing and walking or other predefined movements. The invented system can promote the use of therapeutic footwear, socks or removable cast walkers (RCWs).

[0009] Foot ulceration is one of the most common precursors to lower extremity amputations among diabetics [3-5]. Ulcerations are pivotal events in limb loss for two important reasons: they provide a direct pathway for infection [6] and also cause progressive tissue necrosis and poor wound healing in the presence of ischemia. In diabetic adults, foot infections rarely develop in the absence of a wound. In this population, ulcers are the most common type of wound [6]. Foot ulcer, therefore, plays a central role in the causal pathway to lower extremity amputation [7].

[0010] Clearly, an effective treatment of foot ulcerations is critical to any plan for amputation prevention. Lowering of pressure, shear, and repetitive injury to the sole of the foot are fundamental to diabetic ulcer care.

[0011] Total Contact Casting (TCC) is considered the “ideal” gold standard to heal diabetic foot ulcers [8-12]. TCCs have been shown to reduce pressure at the site of ulceration by 84-92% [13]. Despite a significant body of clinical and laboratory work indicating their efficacy and safety, total contact casts are not widely used due to several practical barriers limiting their adoption by the general medical community. They are difficult to apply, expensive, and place significant demands on resources from busy clinics. Additionally, should the foot ulcer require re-examination, the TCC must be removed and then fully reapplied. Due to these disadvantages, an effective alternative to TCCs is needed.

[0012] The present invention enhances patient adherence by reminding the user and/or alerting caregivers if the patient moves without using the cast. This alarm function is useful to a significant number of patients who tend to remove treatment devices from their feet [14]. Although standard RCWs reduce peak pressures as effectively as TCCs [13], in both descriptive studies and randomized clinical trials TCCs achieve higher healing rates. A logical explanation for this is patient non-compliance to treatment.

[0013] In the fields of elderly care, physical medicine, and rehabilitation, the present invention can be used to measure adherence and remind patients to use various prescribed braces, orthotics, prostheses, and ambulatory aid devices during walking, movement or prescribed posture.

[0014] In clinical research and studies, the present invention provides valuable insight into the factors affecting patient compliance with prescribed footwear, walking aid devices, orthotics, and braces. Additionally, the present invention can help quantify the relationship between the usage of prescribed footwear, walking aids and orthotics with the user's state of health.

[0015] The present invention can also be used to gather data on the type of footwear being used and the duration of time spent in each body posture. This invention will provide insight into exactly what shoes are being worn for a specific activity by individuals. Such data enables the prescribing clinician to use more effective therapeutic or treatment strategies based on the patient's preferences for footwear and activity demands. It may also help manufacturers in their design and development of footwear.

[0016] In addition, the invented technology can provide key information about the user's activity pattern. For example, this invention can identify, for a given subject, the physical condition leading to least adherence with the caregiver's instructions. Such information, in turn, may lead to new regimens for improving patient adherence. By providing an alarm during undesired foot loading conditions, moreover, patient adherence may be improved significantly, thus minimizing complications during the treatment period. The alarm can be adjusted by the clinician. For example, the clinician can set the alarm when the subject's walking falls outside of a predefined range of parameters (e.g., 5 steps walking or 10 seconds standing) without wearing the prescribed footwear or offloading device.

[0017] The present invention consists of a body movement monitoring system (see FIG. 1), including a sensor unit 103 comprising a multitude of accelerometers. The sensor unit may be either carried by the user 101 or attached to the user's body. The accelerometers may be mono- or multi-axial. If multi-axial, the accelerometers measure accelerations in perpendicular directions. The accelerometer data is used to detect the user's foot loading condition, posture (e.g., standing and walking) as well as his or her gait parameters.

[0018] The system and algorithms used in the present invention are robust and can be used independent of the placement of the sensor on the body. The system may provide information and alerts to cell phones and other aid devices such as autobiofeedback or reminder systems.

[0019] The system includes one or more identifiable Radio Frequency Identification (RFID) tags 104 incorporated in or attached to said user's prescribed footwear 102, brace, orthotics or other walking aid device (FIG. 1). If the sensor detects the RFID tag to be out of range—as would be the case if the user 101 does not wear the footwear 102 incorporating the tag 104—and at the same time detects a foot loading condition (e.g., walking), it reminds the user 101 and/or alerts the clinical center (e.g., via radio-frequency (RF) communication).

[0020] The system includes an auditory and/or vibration alarm that can be used to alert the user 101 or the user's caregiver. Once the system detects a foot loading condition at the same time as an out-of-range RFID tag, it activates an alarm and/or a reminder system after the user 101 exceeds a threshold number of walking steps (e.g., 5 steps). The walking steps can be identified using the signals from the accelerometer(s), processed by algorithms developed to measure the walking period.

[0021] Some or all of the required analysis may be carried out on-board the sensor unit. In some cases, software-based algorithms, developed as part of the present invention, are integrated within the processor circuits performing the analyses. These algorithms can be used for real-time interventions.

[0022] The system also includes one or more data storage systems configured to store (log) signals recorded by said accelerometer(s), or the information derived by one of said processor circuits, or both, as well as the RFID data. One or more of said data storage systems may be housed within the sensor unit 103.

[0023] Additionally, the system allows identifying the period in which the sensor 103 is not worn at the chest level by assessing user's 101 respiration fluctuation. The sensor 101 may optionally include a selection switch for the user's 101 preferred mode of biofeedback (e.g., auditory/vibration; similar to sound/vibrate options available on cellular phones).

[0024] The algorithm used by the presentation invention consists of the several steps described below:

[0025] I. Gait Analysis and the Identification of Walking Periods

[0026] The first step performs a gait-analysis function, removing from the recorded data periods associated with postural transitions and lying. This algorithm is based on previous work described in reference [15].

[0027] Next, to attenuate the trunk rotational artifacts, the algorithm estimates the magnitude of tangential acceleration—i.e., the time-varying norm of the vertical and frontal accelerometers' signals. This allows for the suppression of the rotational components of trunk acceleration. The gravitational component may be removed, and the frontal acceleration signal may be reduced through wavelet transform with a filter—e.g., mother wavelet: db4, scales 2-5.

[0028] An algorithm using peak detection estimates the number of gait steps and cadence using the vertical accelerometer signal. A walking segment is defined where at least three successive peaks beyond a threshold are recognized at a pre-defined interval.

[0029] In addition, the system identifies right and left gait steps using the lateral accelerometer signal. During the recognized walking period, the lateral accelerometer output is integrated to obtain the lateral velocity. Comparison of the time location of the positive and negative peaks indicates the right-left sequence of the steps. The information from the step detection algorithm and the left and right step detection algorithm are used to identify left and right gait cycle times (right gait cycle time (n)=right step (n)+left step (n+1); left gait cycle time (n)=left step (n)+right step (n+1); here ‘n’ denotes the index of the identified step). This timing information is crucial to the detection of ‘turning.’ A turn is detected if the difference between the right gait cycle time and the left gait cycle time at time ‘n’ was exceeds a threshold.

[0030] II. Use of Respiration Fluctuation to Identify Whether the User has Worn the Sensor Unit.

[0031] The system includes algorithms implemented in the sensor unit 103 to measure and monitor the user's 101 adherence in wearing the body-worn motion sensor 103. These algorithms detect fluctuations in the acceleration data recorded by the sensor 103 due to user's 101 respiration. At rest, the fluctuation of the acceleration data due to respiration is identified by the algorithms (typically the standard deviation >0.005 g, FIG. 3b). When the sensor 103 is placed on a table, the acceleration fluctuation is close to zero and only white noise is recorded (typically the standard deviation <0.0005 g, FIG. 3c and FIG. 3d).

[0032] III. Using RFID Tags to Identify Whether the User has Worn the Prescribed Footwear

[0033] The system includes an RFID tag reader (e.g., an RF unit) in the sensor unit 103 and one or more RFID tags 104 inside the prescribed footwear. The algorithm implemented by the processor on-board the sensor unit 103 first characterizes the type of foot loading condition 201 occurring at any instant of time. On this basis, the algorithm identifies periods of the walking 203 and/or standing 202 by categorizing the foot-loading condition. Identification of walking and its period is explained in Section I above.

[0034] If a foot-loading condition satisfying predefined criteria is identified (e.g., a duration more than 10 seconds for standing 204 or more than 10 steps during walking 205, the sensor unit scans the RFID tag(s) 104 (FIG. 2, 207). If the RFID tag(s) 104 is (are) out of range (FIG. 2, 208), the sensor unit sounds an alarm and/or stores the time instant at which this condition is first detected within the on-board memory of the sensor unit 103. The sensor unit 103 continues to rescan the presence of RFID tag(s) 104 in a predefined interval until the time when it either recognizes its (their) presence, or the user's 101 posture changes to a non-foot loading condition (e.g., sitting, lying, etc.)

[0035] Among the advantages of the present invention are the following:

• • [0036] a. RFID technology is inexpensive;
  • [0037] b. since the RFID tag 104 is light and small, it can be unobtrusively integrated into most types of shoes, insoles, or socks. This is unlike current instrumented shoes requiring room for bulky circuit boards;
  • [0038] c. the system identifies the adherence of the user 101 in wearing the sensor by measuring respiration fluctuation. Therefore, unlike other approaches, the system identifies whether the sensor unit 103 is being carried by the user 101;
  • [0039] d. the system will activate the RF component only in predefined activities, thereby minimizing the power requirements of the system, and enhancing battery life.

[0040] IV. Using the Optional Accelerometer in the Tagged Device(s) to Increase the Accuracy of the Device(s) in Evaluating Adherence to Prescribed Footwear

[0041] The tagged device 102 may optionally include an accelerometer used to detect the user's 101 locomotion. In one embodiment of the invention, the user's 101 locomotion is identified using a peak-detection algorithm. If, during a predefined interval (approximately six seconds), at least three successive peaks exceeding a predefined threshold (approximately 0.1 g) are detected by the optional accelerometer, that interval is assumed to include locomotion by the user 101. Otherwise, the interval is assumed to not contain any locomotion (e.g., ‘no locomotion’). In one embodiment, this information may be recorded in an embedded memory within the tagged device as a coded value of “1” for locomotion and “0” for “no locomotion.” When the sensor unit identifies locomotion simultaneously with detecting the presence of the tagged device within a predefined distance of the sensor, the content of the memory on board the sensor unit 103 is checked to verify whether a locomotion event was also detected using the tagged device. If no-locomotion is recorded or detected by the tagged device, it is assumed that the tagged device has not been worn.

[0042] FIG. 1 is a schematic of the proposed system. The sensory unit is shown on the subject's chest in this figure. Generally, however, the sensory unit 103 may be either carried out by the user 101 or attached to his or her body.

[0043] FIG. 2 is a flowchart of the algorithms used to identify the foot loading conditions, walking periods, and for computing the subject's spatiotemporal parameters of gait.

[0044] FIG. 3 illustrates an example of acceleration data (part a), including when the sensor was worn (part b); and when the sensor was placed on a table (parts c-d).

• [0045] [1] D. G. Armstrong, P. L. Abu-Rumman, B. P. Nixon, and A. J. Boulton, “Continuous Activity Monitoring in Persons at High Risk for Diabetes-Related Lower-Extremity Amputation,” Journal of the American Podiatric Medical Association, vol. 91, pp. 451-455, 2001.
• [0046] [2] D. G. Armstrong, L. A. Lavery, S. Wu, and A. J. Boulton, “Evaluation of Removable and Irremovable Cast Walkers in the Healing of Diabetic Foot Wounds: a Randomized Controlled Trial,” Diabetes Care, vol. 28, pp. 551-4, 2005.
• [0047] [3] S. C. Wu, J. L. Jensen, A. K. Weber, D. E. Robinson, and D. G. Armstrong, “Use of Pressure Offloading Devices in Diabetic Foot Ulcers: Do We Practice What We Preach?” Diabetes Care, Aug. 11, 2008.
• [0048] [4] S. C. Wu, R. T. Crews, and D. G. Armstrong, “The pivotal role of offloading in the management of neuropathic foot ulceration,” Curr Diab Rep, vol. 5, pp. 423-9, December 2005.
• [0049] [5] D. G. Armstrong, “Variability in Activity May Precede Diabetic Foot Ulceration,” Diabetes Care, pp. 3028-3029, 2004 2004.
• [0050] [6] P. W. Brand, “The diabetic foot,” in Diabetes mellitus, theory and practice, 3rd ed, Ellenberg M Rifkin H, Ed. New York: Medical Examination Publishing, 1983, pp. 803-828.
• [0051] [7] R. E. Pecoraro, G. E. Reiber, and E. M. Burgess, “Pathways to diabetic limb amputation: basis for prevention,” Diabetes Care, vol. 13, pp. 513-521, 1990.
• [0052] [8] American Diabetes Association, “Consensus Development Conference on Diabetic Foot Wound Care,” Diabetes Care, vol. 22, p. 1354, 1999.
• [0053] [9] W. Coleman, P. W. Brand, and J. A. Birke, “The total contact cast, a therapy for plantar ulceration on insensitive feet,” J Am Podiatr Med Assoc, vol. 74, pp. 548-552, 1984.
• [0054] [10] P. A. Helm, S. C. Walker, and G. Pulliam, “Total contact casting in diabetic patients with neuropathic foot ulcerations,” Arch Phys Med Rehabil, vol. 65, pp. 691-693, 1984.
• [0055] [11] S. C. Walker, P. A. Helm, and G. Pulliam, “Chronic diabetic neuropathic foot ulcerations and total contact casting: healing effectiveness and outcome probability (abstract).” Arch Phys Med Rehabil, vol. 66, p. 574, 1985.
• [0056] [12] D. R. Sinacore, M. J. Mueller, and J. E. Diamond, “Diabetic plantar ulcers treated by total contact casting,” Phys Ther, vol. 67, pp. 1543-1547, 1987.
• [0057] [13] L. A. Lavery, S. A. Vela, D. C. Lavery, and T. L. Quebedeaux, “Reducing dynamic foot pressures in high-risk diabetic subjects with foot ulcerations. A comparison of treatments,” Diabetes Care, vol. 19, pp. 818-21, 1996.
• [0058] [14] D. G. Armstrong, L. A. Lavery, H. R. Kimbriel, B. P. Nixon, and A. J. Boulton, “Activity Patterns of Patients With Diabetic Foot Ulceration: Patients with active ulceration may not adhere to a standard pressure off-loading regimen,” Diabetes Care, vol. 26, pp. 2595-2597, September 2003.
• [0059] [15] B. Najafi, A. Vaziri, and A. R. Boloori, “An Ambulatory System for Measuring and Monitoring Physical Activity and Risk of Falling and for Automatic Fall Detection; EFS ID: 2314530; Application number: 60979557,” US-Patent, Ed., 2007.
• [0060] [16] B. Najafi, J. Wrobel, and D. Armstrong, “A Novel Ambulatory Device For Continuous 24-H Monitoring Of Physical Activity In Daily Life,” in North American Congress on Biomechanics (NACOB), Michigan, 2008.