Overview:
- Conceptual framework for neurorehab plan of care:
- Healing via spontaneous recovery and through rehab
- Diagnosis, prognosis, intervention, assessment, outcomes
- Contrasting examples
- Review of "systems" analysis of this framework, as an optimization problem (from Module 1)
- Conventional Approaches to Diagnosis, Intervention and Outcomes Assessment
- Patient records (paper, electronic)
- Scales ("instruments" "forms" "measures") - by trained observer and/or self-report
- Objective (sensor-based) measures (not common, but examples from PT, speech, gait)
- Example: Prognosis process and dynamic prognosis model using fuzzy inference system (Module 4 in BIEN 269)
- Example: Stroke Neurorehab
- General Assessment and Intervention Tools
- Objectives
- Roles for assessment
- Types of measures
- Classification of sensors
- Challenge of timely assessments
- Process of making assessments
- Outcomes assessment approaches
- Neurorehab Assessment Scales
- Categories and examples
- Considerations for administering scales via telerehab tools
- Classification for Innovation: Assessment, intervention/therapy, activity monitoring (e.g., wearable)
- Overview of Intervention Approaches and Alternative Tools/Strategies
- Overview of roles for technology in current rehab practice
- Example of a rehab hospital with leadership in use of tech innovation for rehab
- Examples of "semi-automated" reahb therapy
- Possibilities for model assistances
- Example: UniTherapy Software for Assistive/Remote Therapy and Teleassessment
- Structural organization of package (Guest lecture by Xin "Tyre" Feng)
- Homework: Hands-on assessment of features in the UniTherapy package
Neurorehab Plan of Care after Trauma: Conceptual Framework
Traumatic events in life, such as neural trauma, are sometimes a part of life. They happen, and often require timely medical attention. As we noted in the section in Module 1 on the field of clinical rehabilitation, the three largest inpatient populations at the typical comprehensive rehabilitation hospital for adults all involve addressing the consequences of neural trauma: stroke, traumatic brain injury (TBI) and spinal cord injury (SCI, here's another good info page for SCI).
In this section we will overview the plan of care from an engineering systems perspective. Significant trauma, whether neural or other forms (e.g., musculoskeletal, cardiopulmonary), normally initiates two types of dynamic processes:
- a sequence of processes internal to the body that relate to minimizing the damage and then helping promote healing and "spontaneous" recovery; and
- a sequence of processes within the healthcare system that also relate to helping understand and minimize the damage and then working with internal healing processes to maximize the degree of recovery and independence of the person, through a plan of care that includes strategic interventions.
Both of these are areas of considerable scientific inquiry. Clearly there is still much to be learned about dynamic healing bioprocesses, as we discussed in the section in Module 1 on rehabilitation science (this is also a key focus in our BIEN 267 course on rehabilitative science). There is also much to be learned about how to best use limited resources to provide effective care, as we conceptualized in the section in Module 1 on optimizing clinical rehabilitative care. In this section, the clinical rehabilitative service delivery process was broke down into a collection of stages: diagnosis, prognosis, intervention, assessment, outcomes. These stages, classically covered in allied health practice guidelines and textbooks as key components of a plan of care (or "clinical algorithm"), were mapped via our "systems" perspective to an optimization problem (from Module 1). From this perspective (see diagram below), the plan of care and its implementation is a function of the prognosis, with the prognosis being a function of diagnostic information, desired outcomes and available resources.
Conventional assessment data that is gathered throughout the process of diagnosis, prognosis, intervention and outcomes assessment include patient records (paper, electronic) and scales (also called "instruments" or "forms" or "measures"). Typically these data come from trained practitioners, self-report, various biochemical tests, and imaging technologies. Objective (sensor-based) measures are not as common, but there are examples from areas such as performance on physical therapy equipment, speech performance, and gait analysis. The next section, on neuro-assessment, will develop this further.
- The Unitherapy project, presented to the class by Xin "Tyre" Feng, is an example of a novel tool in this area, and it and other new approaches at assessment are discussed in the Neuro-Innovation section.
Prognosis prediction, a critical part of developing a plan of care, involves use of an interesting synthesis of evidence-based prediction, expert training and experience, and common sense. It is most commonly the job of the physiatrist (or rehab doctor), working with other clinicians, the patient and family. While it is challenging, it is based on clinical training and hopefully a pillar in knowledge of the relevant scientific literature.
- A doctoral student has created a modeling environment that was initially designed to try to extract expert knowledge of clinicians so as to try to predict prognosis as a function of interventions (input events) to the system. This model, developed by Wang and Winters, originally called Intelligent Telerehab Assistant-Prediction mode (ITA-Predict) with a Med-Predict module, is now called SoftBioME, and is used in the BIEN 269 (Modeling Rehabilitative Biosystems) course. This model uses fuzzy logic to develop differential equations of rehabilitative bioprocesses and practitioner experience. Inputs to the model include facts from a patient record, therapeutic sessions, medications, and context data. States typically represent degrees of impairment or physiologic variables. Outputs typically relate to predicted performance metrics, and outcomes are predictions of general measures to be maximized or minimized, such as independence.
Typically available resources are tied to a certain stage within the plan of care, with different resources available for inpatient services, outpatient services, homecare, etc. One of the reasons that neurotrauma dominates inpatient care is that it is widely recognized that the classic outpatient/homecare reimbursement model of roughly 3 interventions per week, often sufficient for musculoskeletal soft tissue and cardiopulmonary healing, is typically not enough for neurorehabilitation. More frequent interventions are needed. A conceptual example is displayed below that shows how interventions can affect outcomes, and how a "chronic" steady-state doesn't necessarily mean an optimal state.
Figure: Conceptual example of the scientific challenges associated with optimizing the rehabilitative plan for therapeutic interventions so as to maximize outcomes, using as an example a typical scenario representative of the neurorehabilitative process related to stroke. Therapeutic interventions (represented as vertical lines in top trace) occur every weekday for about a 3-week inpatient time, during which the degree of recovery (thick line) is presumed better than spontaneous recovery (fine line). This period is followed by roughly twice per week interventions during several months of outpatient/home therapy, in which the person is also encouraged to follow a home exercise program on his/her own. In this case, a gradual decrease in compliance is commonplace, and some of the recovery is lost, though the sustainable level is still above spontaneous. Evidence suggests that 2-3 sessions per week is ineffective for neurorehabilitative recovery. There is, however, considerable evidence that intensive therapy sessions of several weeks duration can be very beneficial for certain cases (represented by dashed lines). The scientific issue relates to optimizing the timing, intensity, and duration of interventions.
Background: Standard Intervention Tools
- Structure and General Forms of Interventions
- Applied Therapy
- Administered by (or directly supervised by) specially trained clinician
- Technology-assisted (e.g., exercise machines, bio-stim, speech software)
- Self-initiated, with health management support (e.g., nurse)
- Self-initiated (e.g., training as inpatient or outpatient, then implemented at home)
- Medication/injection therapy
- Administered by clinician (e.g., injection)
- Self-administered (perhaps with reminders, often implementing prescription by physician)
- Health management (both preventative and therapeutic)
- Access to clinician support and expertise
- Access to caregiver
- Self (e.g., diet, etc.)
- Ties Between Intervention and Assessment
- Close ties (e.g., assessments embedded into intervention) are idea
- Optimize recovery and outcomes by using feedback of assessments to refine interventions
Early Diagnosis and Management Process for Stroke
- Background: Basic Mechanisms of a Stroke
- Ischemic (~ 70%)
- Hemmoragic (~30%)
- Advantages of Rapid Response - Minimizing Damage
- Thrombolysis with tPA (tissue plasminogen activator), if within ~ 3 hours of acute ischemic stroke
- Use of a quick CT or (or perhaps MRI / MR angiography in future) to make sure its ischemic
- Aspirin within 48 hours (mild benefit, low risk and cost)
- Early Medical Management: Refining Initial Diagnostic Assessment
and Documentation
- Stroke etiology and areas of brain involved
- Identification
of symptoms and basic functioning
- Imaging tools such as CT, MRI, MR angiography, ultrasound (type, location(s))
- Functional Assessment: Types/severity of neurological deficits
- Arm and leg function
- Check for dysphasia (problem with swallowing)
- Types/severity of comorbid diseases
- Any complications and abnormal health patterns
- Monitor for changes in clinical status over time
- Estimate (e.g., through interviews) functional status prior
to stroke
- Medical Management for All Stages
- Basic information on rehab therapy from www.stroke.org
- Nutrition/hydration/sleep/rest
- Stabilize medical condition
- Take steps to prevent recurrent stroke
- Minimize risk factors (these include hypertension, smoking,
diabetes, high serum cholesterol, heavy alcohol consumption, overweight)
- Medications include oral anticoagulants, aspirin
- Possibly surgery (not common)
- Secondary complications include:
- Deep vein thrombosis (DVT)
- Dysphasia (problem with swallowing) and aspiration
- Skin breakdown
- Bladder/bowel function
- Prevention of urinary tract infections
- Seizures/falls Acute care
monitoring will include:
- Spasticity/contractures (and treatment options)
- Shoulder injury (positioning
- Screening and Planning for Rehab (typically 1-4 days after trauma)
- Alternative choices for therapy (if resources permit):
- Needs specific rehab services
- Needs comprehensive rehab
- Needs further recuperation before rehab decision
- Too incapacitated for rehab
- Assessment
tools
- Baseline, during, at discharge, perhaps at later outpatient/homecare stages
- Scales
- Neurological deficit (NIH Stroke Scale)
- ADL abilities: FIM, Barthel
- Motor function/skills (Fugl-Meyer,
MAS, AMPS, Jebson, ...)
- Mental status (many), depression (many)
- Balance (Berg), Mobility (Rivermead)
- Speech/language aphasia (many)
- Activity log (MAL for amount/quality of use)
- Quality of life/health status, satisfaction,
...
- Imaging (internal estimates of structure, function)
- Common Functional Impairments in Stroke (each case in unique)
- Over 75% are hemiplegic (i.e., affecting mostly one side of body)
- Neurological domains
- motor (face, arm, leg)
- sensory (loss of sensations, change in sensation (altered sensitivity, numbness/tingling), loss of perception
- vision (monocular visual loss, left-sided neglect, etc.)
- language (dysphasia - disturbances include comprehension, naming, repetition, fluency, reading, writing)
- Muscle weakness (force, power deficit)
- Sensory reflex deficit
- Spasticity/spasms
- Posture/movement asymmetry & balance/gait
- Poor gross/fine coordination
- Often stages (related to spontaneous recovery, neural plasticity):
- Flaccadic/weak then spastic/stiff
- Synergy patterns evolve
- Typically lower extremity recovers earlier and better than upper
- Whole Body/Legs
- Weakness and/or spasticity big issues
- Standing - asymmetric body weight support, posture
- Gait - slow & asymmetric
- Arm
- Shoulder dysfunction (e.g., subluxation)
- Reaching and arm positioning
- Tendency for flexor synergy (flexed elbow &wrist, thus hand in front of chest)
- Grasping/manipulation: Tendency for weak wrist/finger extensors
- Pragmatic
considerations
- Emotional status, motivation, communication level,
endurance, tolerance for rehab
- Social/environmental: presence of caregiver, living
situation, family, ...
- Resources and available rehab programs
- Inpatient
- comprehensive rehab hospital
- acute care hospital
- Nursing homes
- Outpatient rehab (typically several hours for
3-4 visits/week
- Home rehab (typically some therapy, nursing
visits
- Other
- Prognosis
- Based on evidence and expert experience
- Often stages for spontaneous healing
- May be several prognosis, e.g. with/without rehab (spontaneous recovery vs spontaneous plus active rehabilitative intervention)
- Goal-Setting and Plan for Intervention
- Most interventions are like "impulses" to the dynamic healing process
- Key forms of intervention
- Therapy (PT, OT, speech):
- Considerations: type, timing, intensity,
duration
- maximize "functional
outcomes"
- controversy: more focus on returning key ADL functions
for independent living, or on (longer-term) skill development?
- Medication (ranging from oral medications to Botox injections)
- Education and support for self-care
Rehab Assessment Tools
While there are many reasons for assessment, most relate in some way to estimation of health status of the client or to better understanding of the client's context (e.g., life goals and key activities, environment, personal support system). Assessment tools have taken added significance in recent years because of the move towards evidence-based practice. The former is of primary interest here.
Types of measures for estimating health status include:
- Measures of body biochemistry and physiological status
- Measures of degree of impairment and body components of function
- Measures of disability and functional performance,
- Measures of a global, integrative nature such as degree of independence
The Figure below provides a classification scheme for the types of sensors that can be relevant.
Figure. Classification scheme for types of sensors of interest in rehabilitation. Along the left is the sampling frequency (“med” implies rates from about 10-1000 total samples/sec, “low” is under 10 samples/sec), and an indication of frequency of collection (“often” it generally refers to continuous monitoring). “Med” applications require significant planning for data reduction and storage unless “events” can be reliably obtained on-the-fly, or a mechanism exists for direct wireless data transfer to the environment (e.g., Bluetooth). Data from “low” and “single measurement” applications can generally be stored on a mobile device, and if necessary transferred when convenient. Wellness sensors are included within physiologic measurement. ECG: electrocardiogram, EMG: electromyogram, HR: heart rate, SPO2, EADLs (electronic aids to daily living).
Such measurements may be at various stages in the rehabilitative process, from diagnosis through outcomes. Depending on the situation, the initial diagnosis may be quite obvious or may involve expert synthesis of a number of the above measures. Once a diagnosis has been, assessments serve a variety of purposes, including baseline measurements, assessments during the rehabilitative process, assessments at hospital discharge, assessments of health status at periodic intervals, etc.
In rehabilitation, assessments normally can be viewed as occurring at a slice of time. The reason is that change in rehabilitation tends to occur on the order of weeks or months, in contrast to assessment time periods which are typically on the order of seconds to under an hour. Short-term changes such as adaptive learning of tasks or transient tissue viscoelasticity may themselves be part of controlled measurements of signal properties, but are not really rehabilitative change. Thus a key challenge is to determine when to make measurements. Another key challenge is what to measure, and why. Because of the focus on evidence-based practice, another common aim in clinical research studies is to use measurements to help wean out change that is due to natural healing mechanisms and that specifically due to interventions.
The concept of timely assessment was raised in Module 3 (Telerehabilitation), and indeed was part of our motivation for telehealth tools. Optimizing times of assessments can involve a balancing act between the best times for sensitively detecting change and practical considerations such as the convenience of the client and practitioners. A potential advantage of sensor-based measures is the opportunity for minimally-intrusive measurement. The target of such measures may be physiologic (e.g., heart rate, respiratory rate), performance (e.g., embedded in game), and/or activity (e.g., home ADL indicators).
The process of making assessments tends to fall into one of four categories:
- Observer assessment
- Expert observer with specialized training
- structured scales
- Patient chart data (e.g., mobility, ROM, tone, strength, coordination, balance, gait, sensory loss, visual deficits)
- Observer with limited training
- Self-report
- structured format (e.g., questionnaire, scales)
- open-ended format (e.g., notes, open-ended questions)
- Performance measurement (sensor-assisted), to estimate external capabilities
- Use of technology with sensor/apparatus (e.g., physiologic, performance)
- Use of imaging tools to estimate internal structure and/or function
- Anatomical structure through spatial images (e.g., x-rays, CT)
- Physiologic function (e.g., fMRI)
Observer and self assessment commonly use scales to make measurements. The motivation for scales is that they provide a scoring metric. The challenge is that psychometrics/rater scales are inherently subjective. It is for this reason that considerable effort has gone into clinical research methodology, as we presented in Module 1 we defined the key concepts of reliability, validity, sensitivity and bias. One common approach is to develop scales that integrate a collection of scored sub-measures, typically summing to a total score that is then the target of reliability/validity studies.
Outcomes assessment in particular has taken center stage in rehabilitation. We define the following:
- Outcomes measurement: Systematic observation of outcome indicators
- Outcomes monitoring: Repeated measurement over time of outcome indicators in a manner permitting causal inferences about source of observations.
- Outcomes management: Use of information and knowledge gained from outcome monitoring to achieve optimal patient outcomes through improved clinical decision making and service delivery.
Categories of Neurorehab Scales (with published validity/reliability
studies)
- Acute Assessment (e.g., for Admission/Screening):
- Consciousness/cognition (e.g., Glasgow
Coma Scale, Mini-Mental
State Exam (MMSE))
- Stroke deficit (e.g., NIH
Stroke Scale, Canadian
Neurological Scale)
- Global disability (e.g., Rankin
Scale)
- ADL/outcomes (e.g., Barthel, FIM, SIS-
see below)
- Health outcomes, physical & mental (e.g., Health
Survey SF-36, on web)
- Screening for rehab adherence (cognition, motivation, depression)
- Rehab Admission/Monitoring/Outcomes:
- General scales such as ADLs (Bathel, FIM) or various QOL (quality
of life)
- Targeted functional assessment scales such as for balance, mobility,
language/speech, dysphagia, hand function, cognition, depression,
continence
- See also summary at http://www.strokecenter.org/trials/scales/
Example - Scales Used in Telerehab & Human Performance Lab (e.g.,
past MS research project of Adenine Stanislaus), related to Neurorehab:
- Functional Impairment:
- Fugl-Meyer Assessment: A systematic suite
of tests using a 3-point ordinal scale that quantify
motor recovery stages based on the scales of Brunnstrom and
Twitchell
(ontogenetic
concept
of motor recovery). In addition to motor recovery
(100 points), balance (14 points), sensation, range of motion
(44 points), sensation (24 points) and pain (44 points) are
also assessed (total maximum score is 226). Movement is examined
in and out of synergies. It
is widely
used for
research
studies. We tend to use the 66-point upper extremity portion
of the assessment.
- General ADL/Independence:
- Functional Independence Measure (FIM): An
18-item test using a seven level ordinal scale
that targets functional assessment
and independence. Roughly 2/3 of items target
motor function, 1/3 cognitive function. Documentation consists
of observing
and
recording what a person actually does. It can
be completed in approximately
15 minutes. Includes a very large national database,
with strong federal buy-in (e.g., NIDRR funding, participation
by VA hospitals).
- Barthel Index: A widely used 100-point assessment
of independence in ten daily activities (10 points for
feeding,
5 for bathing, 5 for grooming, 10 for dressing, 10 for
bowels, 10 for bladder, 10 for toilet use,
15 for transfers,
15 for mobility,
10 for stairs), originally designed for use with people
with neuromuscular or musculoskeletal
disorders.
It is
normally
completed within 5-10 minutes.
- Daily Activity:
- Motor Activity Log (MAL): This is a "real world"
measure of 30 different functional tasks, scored by self-report
during a structured interview in terms of "how often" and
"how well" they are performed (both on 0-5 scales in 0.5
increments), typically applied to the previous week. Developed
by the
group responsible for constraint-induced movement therapy.
It can be
completed in roughly 30 min.
- Targeted Functional Performance:
- Nine-Hole Peg Test: The 9 hole
peg test is a simple timed test of fine motor coordination, involving
placing dowels (9 mm in diameter and 32 mm long) in 9 holes. Subjects
are
scored on the amount of time it takes to place and remove all 9
pegs. Two scores are collected, one for each hand. Takes several
minutes.
- Jebson-Taylor Hand Function Test: Timed performance
of seven test items designed to represent various
aspects of hand function, using common activities such
as writing, simulated
feeding, holding objects, turning cards or pages
as in reading, etc. The dimension used to measure each
function is the length
of time taken to complete each of the tasks. Performed
in 10-15 minutes for both hands.
- Wolf Motor Function Test: A lab-based test
focusing on arm function that involves 15 timed
measures and 2 force-based measures which progress in
complexity
from engaging
individual joints to use of the total arm. For
the 15 timed tests, an ordinal score elated to the quality
of movement
is also scored.
All are goal-directed, and several are functional
(e.g., raising a can to the mouth). It can be completed
in roughly
30 min.
Overview of Opportunities for Innovation, and Survey of Ongoing R&D
- Classification for Innovation:
- Assessment
- intervention/therapy
- activity monitoring (e.g., wearable)
- Possibilities for technologies addressing multiple areas
- Roles for Technical Tools in Interventions
- Injections and surgery by physicians
- Technologies for incisions, surgery, EMG, injections, etc.
- Medications (e.g., baclofen, dantrium), motor point block (Botox, phenol), pharmacologic agents (e.g., amphetamines)
- Examples of Augments for Conventional Allied Health Therapy
- Physical therapy equipment
- Externally-powered isokinetic/isotonic exercise
- Occupational therapy equipment
- Computer interfaces / assistive technologies
- Speech/language therapy
- Therapeutic software to practice language understanding, speech, communication
- Example of a Rehab Clinic Trying Innovative Technologies: Sister Kenny Rehab Institute
"Semi-Automated" Robotic Therapy
- "Robots and Therapy" info from stroke site
- "Robotic" therapy
- Discussed more fully in Dr. Michelle Johnson's class.
- MIT-Manus clinical trials, updates
- InMotion products
- InMotion2 - planar shoulder-arm robotic
- Mostly tracking tasks, somewhat like games
- Stanford / Palo Alto VA
- MIME (Mirror Image Motion Enabler)
- 3D robot for Rehab Therapy (e.g., summary paper)
- Passive and active assist, repeatable movement patterns
- E-stim/FES therapy (muscle, nerve, brain areas) & EMG biofeedback
- Computer-Assisted Motivating Therapy (CAMR) approaches
- Palanca, joysticks, TheraJoy, TheraDrive, etc.
- "Virtual-reality" therapy
- Driving simulators
- Ranging/exercise protocols (e.g., Rutgers)
- Gait / walk-over protocols
- Neurocognitive
Example: UniTherapy Software for Assistive/Remote Therapy and Teleassessment
- Structural organization of package
- Telerehab capabilities: Telepractitioner and Patient Terminals
- Accessibility features (e.g., mapping to user-space via range-of-motion)
- Systems Identification Manager
- Tracking Manager
- Impedance Manager
Innovations in Mobile Systems and Intelligent Agents
- Roles for Technical Tools in Activity Monitoring / Mobile Assessment
- Technical Motivation
- wireless technologies
- wearable sensor and computing technologies
- intelligent systems
- Practical Motivation
- New rehab classification model includes "activity" and "participation"
- Many rehab "scales" are trying to estimate daily activity - better would be to measure directly
- Frequency of activity
- Quality of activity (e.g., quality of movement)
- Desire for timely assessments
- Key functions:
- Activity documentation
- Task-assistance
- Decision-support
- Greatest challenge: intelligent management of information
- Intelligent agents - observes (senses) on the user’s behalf
- Is empowered to do something once triggered, subject to rules that are under the user’s control or supervision.
- Works mostly in the background, performing simple tasks (e.g., event recognition causing data collection) that require minimal attentional resources, and
- Negotiate with other agents (e.g., storage media), with lower-level decision-making capabilities.
- Intelligent assistants - interacts with user to help the user complete a specific type of task
- Ex: reminders related to a certain assessment or therapeutic protocol, performance monitoring to change settings during exercise tasks
- Functions as an assistive technology.
- Rules are typically driven by a combination of both a user-event (e.g., input to PocketPC screen) and a sensor-event.
- Possibilities
|
|
