Models of Telehealth Encounters, or Tele-Encounters

In the previous section on Universal Access it was noted that there are many types of access barriers, and many possible strategies for achieving access. This section focuses on strategies for breaking down the barrier of distance. Here the focus is more on the effectiveness of the "tele-encounter" between the client and another person and/or device. We start by classifying telehealth services in various ways, with a focus on the "process" associated with the encounter.

Classification Model for Telehealth

Process Models for Telerehab (Systems Perspective)

• Different goals require different solutions (participants, forms of tele-encounter)
  • Many different processes are viable (see list below)

• General possibilities:

• Hub-spoke community telemedicine (or teleconsultation), between conference rooms at different hospitals or clinics, typically to give a practitioner, and possibly a patient, access to specialized expertise,
• Telehomecare, between a consumer (and perhaps caregiver) in a home and a practitioner at a remote site, typically a telenurse who also serves as the case manager,
• Minmally obtrusive telemonitoring, where information is obtained, perhaps stored, and transmitted to a central site where it is monitored,
• Home teletherapy, where therapeutic interventions in the home are either administered remotely or telecoached and telesupported from a distant location.

Community Telemedicine (Medical Teleconsultation) Model:

• Infrastructure: Proactively planned rooms for group videoconferencing via H.320 (occationally H.323), electronic record-sharing
• Typical room has a large conference table, special lighting, multiple cameras (and often monitors), sound system. Communication is by audiovisual conferencing, sometimes augmented by an image from a document camera (e.g., medical image) or computer application (e.g., Powerpoint).
• Forms of Tele-Encounter: Typically used for tele-consutations, i.e. access to specialized expertise via "hub-spoke" model.
  • "Hubs" typically on the grounds of major medical centers"Spokes" typically near the town center of a rural community. Typically there are clincians on both sides of a call, sometimes with patient also present. Desired outcome of tele-encounter: successful consultation (e.g., diagnosis, plan of care, outcomes assessment)
• Technology Usability & Access:
• usability: often high (for those without disabilities); communication culture changes when multimodal interface is available, for better or worse.
• access: low-to-moderate (must get to and use a specialized room)
• Local Example:
  • Zablocki VA in Milwaukee ("hub") with Iron Mountain VA ("spoke") for rehab "televisit" for areas such as prosthetic prescription or follow-up.

TeleHomecare Model:

• Infrastructure: Typically a "practitioner site" (e.g., nurse) and a "patient site" (that may also include a caregiver)
• Often H.324 videophones are used (or IP). The practitioner site often includes a computer workstation (e.g., for inspection of data, records). The practitioner site is simpler, including a monitor (small screen or their own TV) and a controller (like "remote" or phone keypad).
• Forms of Tele-Encounter: Similar in aims and scope to a "home visit," with typical tasks including:
  • interactive tele-support/education & general monitoring via interview/observation (using videophone)
  • cooperative telemonitoring/teleassessment (e.g., often including sensor-based health status indicators such as scopes, vital signs)
• Technology Usability & Access:
• Technology usability: interface is simple to use, but poorer quality video and audio, and limited access to health technologiesAccess: potentially very high (key questions on protocol that first ned to be addressed: Scheduled or self-initiated calls? Who initiates? Who has access to whom?)
• Local Example:
  • VA telehomecare, with videophones sent home with certain SCI clients.

Telemonitoring:

• Infrastructure: Mixture of sensor-based technologies and videoconferencing.
• Forms of Tele-Encounter: As unobtrusive as possible, may involve sensing of physiologic and/or task performance measures (possibly with one-way video).
• Technology Usability & Access:
  • usability: if well planned (e.g., simple, targeted interface), potentially high
  • access: if well planned (e.g., wireless/wearable), potentially high
• Local Example: Aspects of our ongoing ITA project (discussed later in class)

Teletherapy:

• Infrastructure: Two-way information transfer between telepractitioner & client terminals model.
• Forms of Tele-Encounter: Psych, speech-language, occupational therapy, physical therapy, medications; tele-coaching (e.g., nurse?)
• Technology Usability & Access:
• usability: depends on application, could be challenging
• access: potentially very high (vs current situation)
• Local Example: VA aspects of psych? Plans for UniTherapy project.

Of course, a single session may switch between several process models.

Scientific Roots for a Conceptual Model for Telerehabilitation

This section develops a conceptual foundation for telerehabilitation that is based on considerations of how telehealth approaches fit into optimizing a rehabilitative plan of care. Its premise is that an optimal design is only as good as its weakest link, and that any of the three processes identified in the Figure below may be this limit. Thus, the proposed scientific framework assumes a model in which all three are part of an integrated system that is to be studied.

Figure: Conceptual model of telerehabilitation within a consumer-centered rehabilitative plan of care. Central to this model is the concept that user-centered telerehabilitation tools can lower the barriers of distance and time. This approach relaxes constraints for timely assessment and interventions that are part of conventional rehabilitative care, and thus the optimization problem. Three interwoven areas are critical to a scientific understanding of the challenge: (1) the science of rehabilitative change that occurs as a function of set of interventions; (2) the science of optimizing human-technology interfaces so as to insure that they are effective and convenient for the user, given the user’s abilities and the collection of tasks to be performed; and (3) the behavioral science of effective interaction that includes tele-encounters and an appropriate support infrastructure that helps motivate and maintain suitable lifestyle modification and promotes compliance with implementing optimal interventions and assessments.

More on each of the three areas follows.

1. Optimizing the Therapeutic Plan of Care: Rehabilitation Science

Scientifically, the field of rehabilitation exists because biosystems (e.g., cells, tissues, organs, persons) are inherently adaptive. We talked about this in Module 1, in the section on rehabilitative science. Clinical rehabilitation is rooted in the premise that carefully planned and delivered therapeutic interventions enhance patient outcomes. The causal relationship between physiologic processes, interventions and outcomes is complex; healing dynamics are a function of the type of injury or impairment (e.g., musculoskeletal, neuromotor, cardiopulmonary), type and magnitude of tissue injury or disease, age of the person, timing and intensity of interventions, etc. While some dynamic change can be attributed to spontaneous healing process, an optimal recovery requires a timely sequence of interventions of appropriate intensity and duration (e.g., exercise therapy sessions, administration of pharmaceutical agents). In general, the timeframe of most interventions is on the order of seconds to hours, with some interventions performed or facilitated by providers (who are typically reimbursed in 15-min segments), and others the responsibility of caregivers or self-care (e.g., home-based exercise, taking medications). Optimal sequencing for these interventions requires informed decisions, which in turn require timely assessments. We addressed some of these issues in Module 1, in the section on rehabilitation as an optimization problem.

With telerehabilitation, the “system model” changes, i.e. constraints are relaxed by new opportunities for more timely assessment and intervention. No longer need there be an acute stage (during which therapeutic services are provided) and a chronic phase. This model in turn allows a more “consumer-centered” management, and opens up new alternatives for optimizing the therapeutic intervention strategy.

U.S. systems changes during the last ten years of clinical rehabilitation practice have not supported this need for more optimally timed and scheduled assessments. Contact hours between practitioners and patients have dropped dramatically. Inpatient days after a traumatic event, such as stroke, have been reduced by roughly 50 percent. Thus, there is currently a greater dependence on outpatient and home health care, which rarely is reimbursed for more than several months. The modest reimbursable services from practitioners (e.g., assessments, interventions) are generally limited to 2-3 times per week. When the patient has completed this portion of the rehabilitation process, the person is classified as “chronic.” During this “chronic” time period there is an implicit assumption that the dynamic physiologic system doesn’t change, despite a time of isolation (and often therapeutic limbo). Except for occasional samples of the person’s state through occasional checkup visits, this is the case until there is a degree of change in pathology or impairment that requires re-admission (e.g., emergency room, inpatient). The system then responds to this new (or recurring) problem, which is often a secondary complication such as a pressure ulcer or pneumonia.

To the engineer, this is a sampled data system, as shown by the switches in the Figure below, in which the switch is eventually “left open.” When such feedback is not available, the best that is possible is an open-loop, feed-forward control; the worst is no control at all. However, the control space is reduced since some of the interventions require a practitioner. Thus, the current algorithm, and its implementation, appears suboptimal in terms of health-related outcomes. It also appears suboptimal when outcome is defined to weigh costs, benefits and other factors – for instance, when benefits are only sampled sporadically, cost decisions are also likely suboptimal.

2. Optimizing Access to Usable Human-Technology Interfaces

The second component of the model is the human-technology interface, which in telerehabilitation is often a two-way human-technology-human interface. Tele-encounters are goal-directed, with specific types of tasks that must be accomplished by persons with certain abilities. Thus, the scientific foundation behind optimizing human-technology interfaces does not involve technology; rather, it concentrates on the study of human performance and product usability. In Module 2 we provided terminology for human factors engineering, ergonomics, usability, accessibility, universal design that is relevant to this discussion. Human factors science focuses on designing system interfaces to optimize the user’s ability to accomplish tasks successfully and without error, within a reasonable amount of time. It is an applied science with roots in understanding how people use tools. Central is the concept of usability – the extent to which a product can be used by specified users to active goals in an effective, efficient and satisfying manner. Choices to be optimized include selection of appropriate sensory modalities, cognitive/memory loads and motor apparatus. The human factors field includes various models of human interaction and information processing. Practical components of a user-oriented design also include ease of learning and user performance, retention, accessibility, and reliability. Without serious consideration given to user-centered optimization of such interfaces, students may see only the technology, and not recognize and understand the principles underlying the design of effective tele-encounters.

Figure: Human factors dynamics of a goal-directed, task-oriented tele-encounter. Each person brings to the tele-encounter a set of sensory channels for receiving information, motor capabilities for sending signals, cognitive and memory capabilities for processing information, communication skills, job skills, personal preferences, etc. Four modes of interaction are identified: video (one- or two-way), audio (typically two-way), data transfer, and tactile (typically two-way). Issues impacting on the effectiveness of the tele-encounter include available bandwidth, experience of the users with the equipment and with each other, the goals of the task(s), and the usability and accessibility of the interface equipment. S: sensory; M: motor; B: uni- or bi-causal manipulation.

The Figure above provides a process-oriented summary of some of the concerns at these interfaces. The rationale comes from the concept that for optimally timed assessment and intervention to become a reality, the tele-encounter must be convenient and safe for the consumer/patient. Always traveling to a hospital or clinic does not satisfy this criterion. Furthermore, as emphasized at a workshop on future homecare technologies, technological advances are expected to lead to a more consumer-centered healthcare system where home access to services is improved. In particular, much of this access will come from wireless communication, and through mobile technologies.

When I was involved with the RERC on Telerehabilitation, we studied various process models for tele-encounters, including:

• Tele-support (e.g., supportive tele-visits by nurses)
• Tele-monitoring (minimally intrusive, often involving sensors)
• Tele-assessment (active remote assessment, e.g., scoring)
• Tele-therapy (actual interactive therapy)
• Tele-coaching (support and instruction for therapy)

All have their place. Yet in conceptualizing these, it is important to recognize that videoconferencing is just one possibility for telerehabilitation. Telerehabilitation is really about the use of tools and strategies to minimize barriers of distance and time for two key rehabilitative processes: assessment and therapeutic intervention. One of the aims in designing human-technology interfaces is to design technology that proactively assists users in successfully completing tasks implementing these processes, using the most convenient user interfaces, and re-designing these interfaces to minimize attentional resources and level of effort. There are many unexplored possibilities.

The table above classifies the many modes of sensor-based information of possible use in telehealth. Notice the inclusion of measures of functional performance, which at present tend not to be part of telerehabilitation measurement, but are arguably the most important type of measure. Often sampling rates are low enough so as to take up little bandwidth. For instance, the frequency content of voluntary movements is typically under about 6 Hz, and even hand tremor is only 8-12 Hz. Thus most biomechanical signals (motion, force) can be collected at rates such as 50 samples per sec, easily satisfying the Nyquist criteria for sampling, as long as impacts between the body and environment are not of interest. Many vital sign measurements are just single values. Even signals like ECG and EMG, often sent at about 500 samples/sec, are not that bandwidth-entensive when compared to video. There are several systems on the market that telemonitor multi-lead ECGs. Activity monitoring needn't require continuous monitoring, as usually data is stored by a small wearable data logger, and transmitted when technically convenient. Ultimately activity tele-monitoring of what people really do could replace measures that estimate independence such as the FIM.

Finally, technology interfaces that are convenient and assistive to the user are less likely to be rejected or abandoned. As we move toward consumer-centered healthcare, such considerations point to a challenge that can dominate over scientific considerations of rehabilitative bioprocess and subtle features of human-technology interface design: the immense challenge of consumer adherence.

3. Optimizing Behavioral Motivation and Compliance

In order for outcome optimization to occur, persons engaging in rehabilitation activities must be compliant with established plans. Compliance, or lack of compliance, with healthcare recommendations present complex challenges. In most cases, compliance with recommended treatment regimes is necessary to maintain or improve health, as well as to manage symptoms of illness or disability. It is a complex behavioral process, influenced by the environments in which people live, healthcare providers practice, and where care is delivered.

Compliance for maintaining a rehabilitation treatment regime, modifying lifestyle, and taking medications varies from person to person, within individuals, over time, and under different circumstances. Other factors that impact compliance include specific recommendations, complexity of prescribed regime, and degree of disruption in normal routines that occur if the treatment plan is implemented. The cost/benefit ratio also must be considered; that is, what does the person stand to gain if lifestyle changes are made?

It would be imprudent to address compliance without considering motivation. Many factors have been implicated to affect motivation, including patient wants, beliefs, perceived benefits, costs, family factors, cultural factors, age, mental and psychological status, severity of illness, co-morbidities, financial status, and rehabilitation environment and personnel, to name a few. Therefore, if the premise that motivation is key to compliance is accepted, interventions should be focused on increasing wants, beliefs, and rewards, while addressing cultural, family, and individual factors, and paying close attention to perceived barriers. That is:

• What does the person want to achieve? What are the goals?
• What does the individual believe about his/her ability to succeed? Does he/she believe that engaging in the prescribed plan will yield the desired outcomes?
• What are the perceived rewards if he/she engages in the program? Are the rewards realistic?
• What supports are in place?
• What are the perceived barriers?

Congruent with this concept of motivation is the notion of self-efficacy. In essence, self-efficacy is conceptualized as one’s judgment about his/her ability to organize and carry out a course of action to accomplish a specific goal. It is both appraised and enhanced by four mechanisms: (1) performance achievement, (2) verbal encouragement by a credible source, (3) role modeling, and (4) physiologic feedback (e.g., pain or fatigue with a specific activity). If these concepts truly have an effect on compliance, telerehabilitation is a logical approach to positively impact compliance, and therefore outcomes.

In social learning theory, interactions among humans are emphasized as the major source of information about themselves and the physical world. This is relevant to our conceptual framework in that telerehabilitation can enhance channels for interaction, and ultimately may improve compliance, behaviors, and outcomes. Telerehabilitation offers an approach that allows healthcare providers to interact with patients, and to monitor compliance and progress on a daily basis. These interactions might include such activities as telemonitoring of activities and vital signs, telecoaching, telesupport, and telereporting of progress.

Insuring regular contact between the healthcare provider and the individual involved in rehabilitation also incorporates a sense of accountability for both parties involved. In addition, providing the healthcare provider and patient with information about compliance and progress serves both individuals. It allows healthcare providers to make adjustments in therapies in a timely fashion, and establishes a mechanism for tracking one’s status on the recovery trajectory.

Possible reading materials:

from Winters et al., Emerging and Accessible Telecommunications, Information and Healthcare Technologies, RESNA Press, 2002): Chapter 11 (Winters), Chapter 12 (Shapcott), Chapter 13 (Tran et al).

|