**3. Enriching the clinic-home loop**

### **3.1. Wearable sensor nodes**

There are several wearable sensors available off the shelve, as a preference, small wireless sensors are more adequate, to simplify placement and allow free range of movements with no restrictions. To communicate with the mobile apps, Bluetooth is commonly used, so no wires at all are needed.

We can also see devices that not only have EMG sensors for biofeedback, but also have accelerometers and other sensors to provide positioning feedback and correction, which, although do not tell the whole story in terms of what's happening internally, can be a much simpler usage scenario for the patients (**Figures 2** and **3**) (**Table 1**).

### **3.2. Serious gaming**

The serious gaming concept is a good description of what is modern biofeedback. Appealing and intuitive graphics are available to facilitate the knowledge about the intended movement or muscular contraction. The idea is to captivate the patient's attention and motivate them to keep doing the exercise with dedication and in higher levels of demand than normal. It is like a game against the self, where the patient wants to cross over his/her limits and comfort zone. This principle is part of the sEMG biofeedback tools used in the clinic and at home, where the objective is always to achieve the goals defined by the physiotherapist for each specific exercise, creating a more engaging experience [10].

**Figure 3.** Block diagram of the wearable sEMG sensor.

**Figure 2.** Example of a possible form factor.

Bridging the Clinic-Home Divide in Muscular Rehabilitation

http://dx.doi.org/10.5772/intechopen.76790

65

Range Up to ~10 m

ACC 14-bit resolution MAG 16-bit resolution

Size 28 × 70 × 12 mm

Weight 25 g

**Table 1.** Sensor characteristics.

Sensors:

**Communication Bluetooth classic or Bluetooth low energy**

EMG 12-bit resolution with a 3-μV signal noise

Battery 155 mA of 3.7 LiPo rechargeable (8-h battery life)

Nevertheless, considering that excessive exercise can be harmful, there is a real need for defining the exact parameters at which home training should be aimed for. Indeed, a home application enables autonomous training in a controlled manner, not only allowing patients to train at the right intensity and in a guided way but also providing them with the confidence and motivation required to adhere to and finish the plan. In our approach, the display includes a straightforward image showing where to place the sEMG sensors, as well as a bar chart, asking to push or relax the monitored muscles; at the same time, the counter counts down on the time for each repetition and on the number of series. This process will be repeated for each exercise prescribed, and, in the end, a final report shows a score based on the repetitions concluded, also aiming to motivate the patient. In addition, it is of utmost importance to give the user a number of tools to aid the exercise execution. Such is the case of an option that dynamically adjusts the exercise goals on the fly, taking into consideration the values that are

**Figure 2.** Example of a possible form factor.

home training app, changes will appear, adjusting the app to patients' exercise needs. With these solutions, the link between clinic and home training is straight and the knowledge regarding the activities performed by patients at home is direct, facilitating the configuration of follow-up physical therapy sessions and reducing the number of visits to the clinic

There are several wearable sensors available off the shelve, as a preference, small wireless sensors are more adequate, to simplify placement and allow free range of movements with no restrictions. To communicate with the mobile apps, Bluetooth is commonly used, so no wires

We can also see devices that not only have EMG sensors for biofeedback, but also have accelerometers and other sensors to provide positioning feedback and correction, which, although do not tell the whole story in terms of what's happening internally, can be a much simpler

The serious gaming concept is a good description of what is modern biofeedback. Appealing and intuitive graphics are available to facilitate the knowledge about the intended movement or muscular contraction. The idea is to captivate the patient's attention and motivate them to keep doing the exercise with dedication and in higher levels of demand than normal. It is like a game against the self, where the patient wants to cross over his/her limits and comfort zone. This principle is part of the sEMG biofeedback tools used in the clinic and at home, where the objective is always to achieve the goals defined by the physiotherapist for each specific

Nevertheless, considering that excessive exercise can be harmful, there is a real need for defining the exact parameters at which home training should be aimed for. Indeed, a home application enables autonomous training in a controlled manner, not only allowing patients to train at the right intensity and in a guided way but also providing them with the confidence and motivation required to adhere to and finish the plan. In our approach, the display includes a straightforward image showing where to place the sEMG sensors, as well as a bar chart, asking to push or relax the monitored muscles; at the same time, the counter counts down on the time for each repetition and on the number of series. This process will be repeated for each exercise prescribed, and, in the end, a final report shows a score based on the repetitions concluded, also aiming to motivate the patient. In addition, it is of utmost importance to give the user a number of tools to aid the exercise execution. Such is the case of an option that dynamically adjusts the exercise goals on the fly, taking into consideration the values that are

whenever possible.

64 Biofeedback

at all are needed.

**3.2. Serious gaming**

**3.1. Wearable sensor nodes**

**3. Enriching the clinic-home loop**

usage scenario for the patients (**Figures 2** and **3**) (**Table 1**).

exercise, creating a more engaging experience [10].

**Figure 3.** Block diagram of the wearable sEMG sensor.


**Table 1.** Sensor characteristics.

being monitored during the execution. Other possibilities include sending a written message to the physiotherapist through a field on the final report, which will be received in real time. The physiotherapist can then change the prescription remotely, adapting the exercises to the current patient state and needs.

**Figure 4** depicts an example as to what an exercise would look like, with clear, achievable goals. The patient can visually grasp on the video what movements he/she needs to perform, what is expected from him/her in terms of sets and repetitions and whether he/she is contracting/relaxing the right muscles.

In **Figure 5**, an example of a simplified report can be seen. The idea is that the patient can get an immediate gratification by seeing a very simple number that should depict how well he/she was able to achieve goals (there is a scale between 0 and 100 inside the star). There is also information below, as to how each exercise has contributed to the final score.

The HIT TARGET represents the time the patient was able to keep up with the goals within the exercise, and the COMPLETE how many repetitions he/she was able to complete. Using a ponderation between these values, and all of the executed exercises, a final score can be shown.

**3.3. Monitoring dashboard**

**Figure 5.** Final report.

available.

scription to go towards the patient's needs

home session, as depicted in **Figure 7**.

assists the physiotherapist to quickly check whom to address:

be able to execute them during the prescribed period;

**1.** Green—user is active and has been executing the exercises on time;

The access to a monitoring dashboard is crucial in this clinic-home process. It is the way to establish a permanent contact with the patient even if several weeks pass since the last visit to the clinic, allowing access to all the home training data and to perform changes on the pre-

Bridging the Clinic-Home Divide in Muscular Rehabilitation

http://dx.doi.org/10.5772/intechopen.76790

67

In **Figure 6**, we can see an overview of all the patients assigned to a given physiotherapist, and what their status are, with regard to their homework compliance. The traffic light colour code

**2.** Yellow—user is active, but is lagging behind on the exercise execution, so he/she may not

**3.** Red—it is impossible for the user to execute the remaining exercises during the time

This not only allows the physiotherapist to have the next session more tailored to the needs of the patient but also tweaks the exercises that the patient needs to execute during the next

**Figure 4.** Example of instructional video.

Bridging the Clinic-Home Divide in Muscular Rehabilitation http://dx.doi.org/10.5772/intechopen.76790 67

**Figure 5.** Final report.

being monitored during the execution. Other possibilities include sending a written message to the physiotherapist through a field on the final report, which will be received in real time. The physiotherapist can then change the prescription remotely, adapting the exercises to the

**Figure 4** depicts an example as to what an exercise would look like, with clear, achievable goals. The patient can visually grasp on the video what movements he/she needs to perform, what is expected from him/her in terms of sets and repetitions and whether he/she is contract-

In **Figure 5**, an example of a simplified report can be seen. The idea is that the patient can get an immediate gratification by seeing a very simple number that should depict how well he/she was able to achieve goals (there is a scale between 0 and 100 inside the star). There is

The HIT TARGET represents the time the patient was able to keep up with the goals within the exercise, and the COMPLETE how many repetitions he/she was able to complete. Using a ponderation between these values, and all of the executed exercises, a final score

also information below, as to how each exercise has contributed to the final score.

current patient state and needs.

66 Biofeedback

ing/relaxing the right muscles.

**Figure 4.** Example of instructional video.

can be shown.

### **3.3. Monitoring dashboard**

The access to a monitoring dashboard is crucial in this clinic-home process. It is the way to establish a permanent contact with the patient even if several weeks pass since the last visit to the clinic, allowing access to all the home training data and to perform changes on the prescription to go towards the patient's needs

In **Figure 6**, we can see an overview of all the patients assigned to a given physiotherapist, and what their status are, with regard to their homework compliance. The traffic light colour code assists the physiotherapist to quickly check whom to address:


This not only allows the physiotherapist to have the next session more tailored to the needs of the patient but also tweaks the exercises that the patient needs to execute during the next home session, as depicted in **Figure 7**.


To answer these needs, a cloud-based infrastructure with easily scalable and deployable services needs to be used. With this in mind, in case there are higher surges of access, the system can easily deploy new instances, or throttle their capacity up, without impacting the system's

Bridging the Clinic-Home Divide in Muscular Rehabilitation

http://dx.doi.org/10.5772/intechopen.76790

69

There are currently several commercial offers to build this kind of robust architectures that do not rely on single servers (either physically or on a single geographical location). For the sake of this discussion, Amazons' AWS system is depicted, as an example or suggestion as to how such an architecture can be deployed. As it can be seen, it relies on distributed static assets throughout the globe, and load balanced requests to readily available, and scalable applica-

In this day and age, there have been several regulatory concerns about the patient's privacy, first with the US Health Insurance Portability and Accountability Act (HIPAA) from 1996, and with the General Data Protection Regulation (GDPR) from 2018. For these systems to work, some unique patient information is usually required to be stored on a database, such as their name, email, age, or a password so that they can access their own information (right

Since these systems can collect huge amounts of clinical data, it is also advisable that in case there is a data breach, there is a pseudonymization of the most relevant information as this reduces the risks to associate the data with the subjects [11]. This process is depicted in **Figure 9**,

availability, enabling continuous delivery.

of access).

**Figure 7.** Homework configuration.

tion servers (depicted earlier as Amazon Elasticbeanstalk).

**Figure 6.** Dashboard overview.

The most important parameters that can be used to adjust the difficulty of the rehabilitation process are as follows:


#### **3.4. Cloud-based infrastructure**

Such a system is rooted in a distributed infrastructure that supports this plethora of equipment, the data being collected, and different user roles management (**Figure 8**). This infrastructure needs to be reliable, secure, scalable, and high performance.

These needs come from the fact that having patients at home using these systems needs to be much more scalable than having single systems on the clinics being shared among the several professionals, as the quality of service will be under heavier scrutiny by a broader audience. To answer these needs, a cloud-based infrastructure with easily scalable and deployable services needs to be used. With this in mind, in case there are higher surges of access, the system can easily deploy new instances, or throttle their capacity up, without impacting the system's availability, enabling continuous delivery.

There are currently several commercial offers to build this kind of robust architectures that do not rely on single servers (either physically or on a single geographical location). For the sake of this discussion, Amazons' AWS system is depicted, as an example or suggestion as to how such an architecture can be deployed. As it can be seen, it relies on distributed static assets throughout the globe, and load balanced requests to readily available, and scalable application servers (depicted earlier as Amazon Elasticbeanstalk).

In this day and age, there have been several regulatory concerns about the patient's privacy, first with the US Health Insurance Portability and Accountability Act (HIPAA) from 1996, and with the General Data Protection Regulation (GDPR) from 2018. For these systems to work, some unique patient information is usually required to be stored on a database, such as their name, email, age, or a password so that they can access their own information (right of access).

**Figure 7.** Homework configuration.

The most important parameters that can be used to adjust the difficulty of the rehabilitation

**1.** Variance—this allows the mobile app to adjust the difficulty of the exercise, depending on how capable the patient is feeling that day. The 'easier' the Variance, the more leeway

**2.** Frequency—the frequency parameters allow the physiotherapist to define how often the patient executes the exercises, and how they are structured within the sessions. Tweaking these values can increase the recovery speed, but need to be treaded carefully, as it can lead

**3.** Exercise type—if the patient is more and more comfortable with the type of exercises he/ she has been given, it is a good idea to vary them, so that the muscle memory does not only

Such a system is rooted in a distributed infrastructure that supports this plethora of equipment, the data being collected, and different user roles management (**Figure 8**). This infra-

These needs come from the fact that having patients at home using these systems needs to be much more scalable than having single systems on the clinics being shared among the several professionals, as the quality of service will be under heavier scrutiny by a broader audience.

structure needs to be reliable, secure, scalable, and high performance.

process are as follows:

**Figure 6.** Dashboard overview.

to injuries or lack of motivation;

**3.4. Cloud-based infrastructure**

work with a given, very specific, motion.

there is;

68 Biofeedback

Since these systems can collect huge amounts of clinical data, it is also advisable that in case there is a data breach, there is a pseudonymization of the most relevant information as this reduces the risks to associate the data with the subjects [11]. This process is depicted in **Figure 9**,

**4. Conclusions**

**Figure 10.** Example of a tokenization process.

ments can be demonstrated.

High recurrence rates and high dropout rates lead to high economic and social costs in the treatment of musculoskeletal disorders (MSDs). Different reimbursement systems for the physical therapy treatments can be found within different countries in Europe, where private insurance or state health systems cover the cost and inefficiency of MSD treatments. In the US, the cost burden is shared between patient and insurers, where patient's co-pay at an average 50% per session. In other countries, as China, there are almost no physiotherapists; treatments of MSD are directed towards general practitioners who are overwhelmed with a large number of patients and tend to treat the pain and not address the root cause. Inefficiency and difficulty to assess the quality of treatments for MSD pain is a concern to the insurers and health-care systems, which are searching for more consistent diagnostic and treatment approaches, along with new models for reimbursement, in which reduced costs and objective outcomes of treat-

Bridging the Clinic-Home Divide in Muscular Rehabilitation

http://dx.doi.org/10.5772/intechopen.76790

71

When discussing with insurers such as Achmea in Holland, they recognize that while physiotherapy represents around 7% of their cost expenditures, quality physiotherapy care ranks as the number one reason a patient will select an insurance plan or change providers. Therefore, insurers are acutely aware of the need to increase patients' actual or perceived physical therapy care and motivation to resolve underlying pains. With no objective information of physiotherapy treatments, the quality of the service and the consistency of treatment outcomes are difficult to assess. Moreover, unscrupulous practitioners can bill unnecessary treatments, leading to high costs and a strong effort from the insurances to fight fraud. This problem was stated in the front page of NY Times: 'Physical therapy has become a Medicare gold mine. Medicare paid physical therapists working in offices \$1.8 billion in 2012 alone, the 10th-highest field among 74 specialties' [12]. In this context, a huge insurance reform is coming where the capitation of reimbursement will be done in the model 'pay per treatment outcome' instead of 'pay per session'. Physiotherapy clinics are then going to be pressed to lower overall costs of

**Figure 8.** Cloud-based architecture.

where the data stored on the repository use a pseudonymization to refer back to the data on the database.

This kind of mitigation should be used not only for the storage of sensitive data but also in the way the data are exchanged between the several systems that these micro-services, cloud-based systems require, so that man in the middle attacks are thwarted, by using tokenization (**Figure 10**), which is a way to substitute sensitive data, such as an email and password pair with a token that has no meaning or representation outside of the system.

**Figure 9.** Physical separation of sensitive data.

**Figure 10.** Example of a tokenization process.
