Remote care nearby

Introduction

The Western world needs to find ways of coping with the increasing number of elderly people and people with chronic diseases who need care for long periods of time. As chronic diseases cannot be cured, treatment aims to relieve symptoms and related complaints. Interventions like physiotherapy and multidisciplinary rehabilitation are therefore becoming more important. Various studies have shown that these treatments are on average effective, although for 30–50% of the patients these treatment do not have the desired effects. ^1–3 This is disappointing for the patients as well as the professionals.

Different reasons for these disappointing results have been given. Patients with chronic disease are often cared for by multiple professionals. As a consequence, communication between these professionals, and between the patient and the professionals, may be less than ideal, which affects the quality of care. ^4,5 Furthermore, treatments are often insufficiently tailored to the individual patient ⁶ and also insufficiently aimed at the patient's self management. ⁷ In addition, patients often experience problems with translating the skills learned from health professionals to their own home and work situation with the result that they stop doing their essential exercises. ⁸ This set of interacting reasons is in line with the chronic care model of Wagner, ⁹ which assumes that an informed activated patient and a prepared, proactive practice team are two essential parties needed to produce improved outcomes.

Besides the increasing number of people who need care for long periods and the quality of treatment, the costs of health care also need to be considered. Costs of care are high and are expected to further increase unless health policies are changed. ¹⁰ More care needs to be delivered, with a higher quality, better tailored to the needs of the individual patient without costing more money or personal effort.

There are high expectations of technology, especially Information and Communication Technology (ICT), to make health care more effective and efficient. As a result there have been many initiatives related to ICT and care. As we have previously pointed out, ¹¹ these can be divided into two main categories:

The use of ICT to improve current care processes by making information more accessible. Examples of these applications are electronic medical records and applications that improve logistics;

Methods

If a telemedicine application is not accepted by its users, the patients and professionals, it will never be implemented successfully in everyday care. Generally speaking, a telemedicine application has a better chance of being accepted if the users can easily handle the application and if the application fulfils the clinical needs of both patients and professionals. This requires three key matters to be dealt with:

Clinical content. Telemedicine services for patients with chronic disorders that aim to increase the patient's level of functioning need to be able to monitor aspects of the patient's functioning, as well as to provide the patient with adequate feedback about this;

OPEN IN VIEWER

Figure 1

Three essential elements in the development of telemedicine systems

Clinical content

The clinical content element consists of two parts: the application needs to assess the patient's functioning in an objective way and it needs to provide feedback to the patient in order to enable him or her to change in response.

Assessing the patient's health status or functional level

Many people with chronic disorders develop a vicious circle as described in various neurophysiological ^17,18 and cognitive behavioural models. ^19,20 All models hypothesize that patients change their physical functioning as a consequence of their disease. However the way that people change differs, and there is insufficient evidence about who develops a chronic disorder and via which mechanism. In addition, there is insufficient research evidence about the functional status of patients during everyday life, which is necessary to define a good starting point for a telemedicine service. Promising parameters, related to the functional status of patients with chronic diseases, are general physical activity patterns and muscle activation patterns.

The average general activity pattern measured by a 3-D accelerometer worn by patients with various chronic diseases for seven consecutive days is shown in Figure 2. This figure indicates that various patient groups show differences with respect to controls of a much stronger decline in activities over the day. For a telemedicine application this means that the activity pattern over the day is important in obtaining insight about physical functioning.

OPEN IN VIEWER

Figure 2

Activity levels in the morning, afternoon and evening expressed as a ratio with respect to the morning level for patients with chronic lower back pain (CLBP), chronic fatigue syndrome (CVS), chronic obstructive pulmonary disease (COPD) and a healthy control group (Controls)

Considering the muscle activation patterns, various experimental studies (e.g. van der Hulst et al. ²¹ ) have shown that patients with chronic pain have higher muscle activity, i.e. they show problems with relaxation of their muscles during everyday tasks, like walking. Thus a telemedicine application that aims to normalize muscle activation patterns needs to be able to quantify the muscle activation patterns and especially the level of muscle relaxation during everyday tasks.

Feedback

Once we are able to quantify the functioning of patients in their daily environment, the next step is to give the patients insight into their functioning by providing adequate feedback. Feedback, first defined by Rosenblueth et al., ²² is essential in the case of learning, no matter what is being learned. Feedback can be divided into intrinsic feedback and extrinsic feedback. Intrinsic feedback refers to the sensory-perceptual information of an individual which is available when an activity is being performed. Extrinsic (augmented) feedback is feedback which is provided on top of intrinsic feedback, mainly by external sources such as a therapist or a technical device. In the case of augmented feedback for the treatment of patients with chronic diseases, the feedback has three different aims. ^23,24 The three aims are:

Reflecting/awareness. Patients are given a reflection on their own functioning with the aim of making them aware or increasing their awareness of the inadequate aspects of it;

Motivating. After people have become aware of the inadequate aspects of their functioning they must be motivated to change. This function of feedback derives from psychological oriented theories like the Social Cognition Theory, ²⁵ Theory of Planned Behaviour ²⁶ and the Trans Theoretical Model ²⁷ which state that changing functioning or behaviour is a dynamic process. It is the consequence of the interaction between the person, the behaviour and the environment, as well as a staged process;

Changing. This function is subdivided into two different aspects: (1) starting from the staged process of change, intentions need to be transformed into concrete actions to change and once changed, adequate functioning needs to be maintained; (2) to be able to change it is important that patients have adequate motor skills to perform adequately, i.e. motor skills learning. ^28,29

An overview of these different functions of feedback is shown in Figure 3.

OPEN IN VIEWER

Figure 3

Schematic overview of the functions of feedback to produce improved functioning of patients with chronic diseases, taking into account psychological theories as well as the theories of motor skill learning

In conventional treatment to improve the physical function of a patient there is usually face-to-face contact between the professional and the patients, where the professional provides direct feedback by talking and showing the patient how to move properly. In telemedicine treatment the challenge is to replace some or all of the face-to-face feedback by feedback provided by technology. This requires choices to be made. ³⁰ The options include:

Content of feedback. Three choices have to be made about the content of the feedback.

Will it be feedback on results (FR) or feedback on performance (FP)? FR gives information about the performance with respect to the goal, whereas FP gives specific information about how the task is being performed and is therefore much richer in its information. Research into these two different concepts has mainly focused on learning motor skills and shows that for simple tasks both forms of feedback are effective, ^31–33 although some studies show that FP is more effective than FR. ^34–36

Will the feedback be positive or negative? For simple tasks negative feedback is more effective than positive feedback. Ashby and O'Brien ³⁷ showed that in more complex information integration tasks, both forms of feedback are effective, but that the combination of the two is most effective.

When providing feedback there is always a comparison of the patient's functioning with some form of norm. The patient's functioning can be compared with their functioning at another moment, i.e. a temporal comparison. ³⁸ Alternatively a comparison can be made with a group of other people, i.e. a social comparison. ³⁹ Which form of comparison is most effective for improving functioning is not yet known. Proper et al. ⁴⁰ have shown that social comparison has a positive influence on awareness but not on the actual performance.

Modality of feedback. In conventional care extrinsic feedback provided by the health-care professional is rich in modality. A professional can talk, show and let the patient feel how to move in a proper way. In technology-based feedback, choices need to be made about which modality to use and there is also a choice between visual, auditory and sensory feedback. Which modality is most effective has mostly been researched in young health people and to a lesser extent in patients with stroke, ^24,36 but the evidence has not been consistent. As well as the modality, the apparatus for providing feedback needs to be considered, e.g. a device such as a mobile phone, an object in the surroundings of the patient such as a light, a computer or even a robot. The choice of the apparatus will probably depend on personal preferences and the context in which the telemedicine application will be provided. For example, at work a patient may prefer a different apparatus than at home. Another important aspect is what information to add to the feedback to make it more personalized. Such information may concern location, time or the weather.

Timing of feedback. The third aspect of feedback that needs to be considered is the timing. Two choices can be made: (1) feedback at predefined times which can be further subdivided into feedback during the performance of an activity (concurrent feedback) or after the activity has been performed (terminal feedback). Also the frequency of feedback needs to be considered, for instance continuously, after a certain amount of activity or after every few seconds; (2) feedback when certain events occur, for instance when the muscle relaxation is below a certain threshold. Which timing strategies are effective needs to be investigated and in this research not only the effects of timing on clinical outcome but also the preferences of patients and the chance of developing dependency on the system needs to be considered. For instance, Voerman et al. ⁴¹ investigated the effect of different timing intervals on muscle relaxation and concluded that feedback every 5 seconds is not ideal and may even be irritating to the patient.

Design

As many as 75% of telemedicine applications do not survive the project phase and can be considered a failure. ¹⁶ One important reason is that the different stakeholders involved have different ideas about the telemedicine application. This leads to miscommunication and developments that do not fit the needs of the users. The applications are not then accepted by the users. ⁴² Better insight into aspects that are related to acceptance of technology is important. It is known that acceptance is influenced by technical aspects of the system such as ease of use, individual characteristics like age, as well as the surroundings of the individual. ⁴³ However, these aspects are too generic to use as a starting point in developing telemedicine applications. Thus, there is a lack of knowledge about factors that influence the acceptance of telemedicine applications. There is also a need for good definition of user requirements. The methods used most often for this purpose are interviews and focus groups.

A more modern technique with potential for telemedicine is scenario-based requirements engineering ⁴⁴ (see Figure 4). A scenario is a chronological description of a certain event (or series of events) that take place when using an application and can be used as the starting point for developing a common vision among all stakeholders. Scenarios help to define the people, their roles and activities, as well as the functions of the telemedicine application and the context in which it will be used. Using this scenario as a starting point, functional and non-functional specifications can be defined from which technical specifications can be derived and a prototype can be developed. The prototype needs to be tested with users at an early stage to explore whether the application fits the requirements. ¹⁶

OPEN IN VIEWER

Figure 4

Overview of a scenario-based method to define user requirements

Outcome

Once the clinical content and the design criteria have been defined, a telemedicine application can be developed that meets the requirements of the users and that is able to monitor and give feedback on the functional status of the patient. The next step is to demonstrate the results when the system is used in everyday practice. Outcome measures usually focus on the technical feasibility of the system and on user satisfaction. ⁴⁵ The results show that in general the applications are technically stable in clinical practice and that patients and professionals have a positive attitude towards using them. This type of evaluation is however not sufficient to justify subsequent large scale implementation. For this, a broader evaluation is necessary. DeChant et al. ⁴⁶ have proposed a staged approach for the evaluation of telemedicine applications, starting early in the development phase. The first phase evaluation is aimed at investigating whether the requirements of the patients are met and whether the application is technically stable. At this stage the user group is small, because the results are mainly used to develop a more mature application.

After the first phase evaluation, a mature application evaluation (stage 2) focuses on clinical outcome. If these results are positive, evaluation becomes even more extensive, using multiple endpoints such as effectiveness, efficiency or accessibility of care (stage 3). The last evaluation phase (stage 4) focuses on external validity and investigates whether the application is as effective when implemented in places other than where it was developed.

Example

The three elements of the method are illustrated for the myofeedback-based teletreatment service for the treatment of patients with chronic neck and shoulder pain.

Design

Based on the clinical content, a scenario was developed describing the people, their roles and activities as well as the functions of the application and the context in which it was being used. Based on this scenario, the functional requirements were defined. Important criteria were that muscle relaxation needs to be monitored and feedback needs to be given when the level of muscle relaxation is insufficient. It was a requirement that the system could be used at the workplace, that people could put on and take off the system by themselves, that the system could be worn under the clothes and that the feedback would be private. Patients and professionals preferred to have teleconsultations as a part replacement for the face-to-face contacts. These requirements resulted in the system shown in Figure 5.

OPEN IN VIEWER

Figure 5

The Body Area Network (BAN) of the ambulant myofeedback treatment. The right-hand image is copyright MobiHealth B.V.

Muscle relaxation is measured with dry electrodes that are incorporated into a garment that people wear under their clothes during everyday life. The garment is connected to a device that vibrates quietly when the level of muscle relaxation is insufficient. ⁵⁰ The patient data are sent wirelessly to a personal digital assistant (PDA) and on this PDA patients can see their own muscle activation data (i.e. this provides continuous feedback on performance). From the PDA, the data are transmitted to a web server which is accessible to the professionals. Once a week there is a consultation by telephone between the patient and the professional to discuss progress and make appointments for the following week. Treatment starts and ends with a face-to-face contact and lasts for four weeks.

Discussion

The present paper describes a method involving three key elements for the realization of successful telemedicine applications. The method is illustrated for a specific application, the myofeedback-based teletreatment service for patients with chronic neck and shoulder pain.

It appears that parameters reflecting physical functioning can give good insight into the health status of a patient and can therefore be used as starting points for telemedicine interventions. Parameters that have good potential in this respect are those related to general physical activity and to muscle activation. However, experiments also show that different disorders are reflected in these parameters in different ways. In addition, there has been little research investigating the functional status of patients in everyday life in an objective manner. This is needed for several chronic disorders to define good starting points for telemedicine services.

Once the functional status of a patient can be assessed objectively in everyday life the next step is to provide feedback to the patient in order to enable the patient to change. The challenge is to develop feedback strategies in telemedicine systems that can at least partly replace face-to-face contact with the professional. However, research into effective feedback strategies for changing functional status is still in its infancy. Research so far has only been performed for very simple tasks, healthy subjects or with a focus on motor relearning in patients with neurological disorders. There have been no studies that focus on the effects of feedback strategies on functional status or everyday tasks.

Concerning the design of telemedicine applications, one can conclude that many telemedicine applications are designed according to the technology, with the consequence that they do not fit the requirements of the users and therefore fail to become accepted. Miscommunication between the various stakeholders, lack of insight into factors related to acceptance of telemedicine applications, as well as methods and techniques to define the user requirements all represent important barriers. Scenario-based requirements analysis is a promising new method which can overcome these potential barriers. Because users are often unfamiliar with telemedicine services, they have difficulty in defining their needs and desires. Scenarios are also considered to be of potential benefit in formulating a good set of requirements. ⁵⁴

Outcome evaluation of telemedicine services in everyday clinical practice is crucial, but so far, the work has mainly been focused on the evaluation of technical performance of the system and satisfaction of the user. According to DeChant et al. ⁴⁶ this is the way that evaluation should be conducted for early prototypes. However, large scale clinical evaluation studies with multiple end points are needed before telemedicine services can be implemented routinely. ⁴⁵ Such studies are scarce.

Notwithstanding the above, addressing these three elements does not guarantee successful implementation. Other aspects that need to be considered are

Financing, since satisfactory financial arrangements for telemedicine are uncommon.