RESEARCH DOSSIER
Fatigue, Symptom Variability, and Activity Pacing in Rehabilitation
Abstract. Fatigue and fluctuating symptoms shape the tolerability, continuity, and outcomes of rehabilitation, yet their mechanisms, measurement, and treatment vary across clinical conditions. This fundamental critical narrative monograph investigates the history and taxonomy of fatigue, subjective and performance fatigability, symptom dynamics, exertion-related deterioration, psychological and environmental contributors, activity pacing, and participation. It evaluates original psychometric studies, randomized interventions in multiple sclerosis and poststroke rehabilitation, research on chronic pain and ME/CFS, contemporary long COVID evidence, and current NICE, WHO, and CDC guidance. Three intervention traditions are distinguished: energy conservation, symptom-contingent energy management, and operant or quota-based activity planning. The scientific synthesis finds condition-specific effects of selected behavioral interventions, substantial variation across trial outcomes, persistent limitations in the comparative pacing evidence, and important safety implications of post-exertional malaise. Longitudinal observation, clinically meaningful phenotyping, attention to delayed responses, and outcome measures centered on sustainable participation are essential to stronger causal inference. This research offers a traceable evidence-based conceptual synthesis; it reports no original patient data, preregistered systematic review, or independent meta-analysis.
Keywords: rehabilitation psychology; fatigue; fatigability; symptom variability; activity pacing; energy conservation; post-exertional malaise; fatigue management; multiple sclerosis; poststroke fatigue; ME/CFS; long COVID; chronic pain; ecological momentary assessment; within-person analysis; participation; patient-reported outcomes; rehabilitation safety
Fatigue as a Rehabilitation Research Problem
Fatigue is a multidimensional experience of diminished physical or mental energy, increased effort, exhaustion, or reduced capacity to sustain activity. Rehabilitation research encounters fatigue as a symptom, an exposure that alters the delivery of care, a potential intervention outcome, and a determinant of participation. These functions are related, but they demand distinct measures and causal questions. A patient may report substantial fatigue despite preserving performance on a brief standardized task; another may demonstrate declining performance without describing intense subjective exhaustion. The scientific object of this dossier is the temporal organization of fatigue-related functioning: how symptoms fluctuate, how activities and environments contribute to those fluctuations, and how people and rehabilitation services allocate finite capacities across meaningful activities.
The crucial outcome is the relation between symptom burden and life activity. Fatigue severity asks how exhausted a person feels. Fatigue interference asks how much exhaustion limits work, household tasks, mobility, cognition, relationships, or recreation. Fatigability asks how performance or perceived effort changes under a defined task demand. Activity pacing refers to several strategies for distributing, altering, interrupting, or scheduling activities and rest; the meaning of the term depends on the clinical population and the intervention theory. Rehabilitation seeks clinically meaningful functioning and participation while accounting for symptom exacerbation, comorbid conditions, environmental constraints, and the person's priorities. The distinction between symptom intensity and activity interference has direct consequences for selecting outcomes, interpreting trials, and planning accessible care.
This monograph is a critical narrative synthesis of identifiable original studies, psychometric research, evidence reviews, and official clinical guidance. It examines historical development, conceptual boundaries, measurement, within-person fluctuations, competing mechanisms, randomized rehabilitation interventions, safety, and research design. It does not report a new patient cohort, registered systematic review, original pooled effect estimate, or individual treatment recommendations. Sources and quantitative observations are cited at the point of use; evidence grading follows design, comparator, measured endpoint, and applicability. The wider inquiry belongs to Rehabilitation Psychology and intersects Health Psychology and Behavioral Medicine, Clinical Psychology, Behavioral Neuroscience, and occupational therapy.
Historical Development of Fatigue Research
Early rehabilitation traditions centered functional capacity, workplace endurance, and recovery after injury. Over the late twentieth century, fatigue became an explicit patient-reported construct requiring standardized measurement, particularly in neurological and systemic illness. Krupp and colleagues (1989) (https://doi.org/10.1001/archneur.1989.00520460115022) introduced the Fatigue Severity Scale for research in multiple sclerosis and systemic lupus erythematosus. Their original study assessed 25 people with multiple sclerosis, 29 with lupus, and 20 healthy adults; these clinical groups provided initial psychometric evidence, not population norms for all rehabilitation patients. The scale emphasized the influence of fatigue on functioning, establishing an important historical bridge between felt exhaustion and disability.
Chalder and colleagues (1993) (https://doi.org/10.1016/0022-3999(93)90081-P) developed a fatigue questionnaire initially using responses from 274 new general-practice registrations, while Schwartz, Jandorf, and Krupp (1993) (https://doi.org/10.1016/0022-3999(93)90104-N) further developed measurement of fatigue dimensions. This expansion created a methodological problem still present in contemporary intervention research: instruments differ in their recall periods, item content, populations, scoring, and target constructs. A numerical decrease on one scale cannot be assumed equivalent to the same proportional decrease on another. Later item-response methods and computerized adaptive testing in the PROMIS program improved measurement precision, as demonstrated in the original PROMIS development report (Cella et al., 2010) (https://doi.org/10.1016/j.jclinepi.2010.04.011) and the rehabilitation-focused fatigue bank evaluation by Lai and colleagues (2011) (https://doi.org/10.1016/j.apmr.2010.08.033).
Occupational therapy developed energy conservation as an approach to task redesign and resource allocation; neurologic rehabilitation investigated cognitive and behavioral contributors to persistent fatigue; and ME/CFS research elevated post-exertional symptom exacerbation as a central safety consideration. These traditions developed partially overlapping vocabularies. In multiple sclerosis, a structured energy conservation course was evaluated by Mathiowetz and colleagues (2005) (https://doi.org/10.1191/1352458505ms1198oa). In poststroke rehabilitation, Zedlitz and colleagues (2012) (https://doi.org/10.1161/STROKEAHA.111.632117) compared cognitive therapy alone with cognitive therapy augmented by graded activity. Later studies treated fatigue management as a complex behavioral intervention, incorporating beliefs, scheduling, participation, and patient choice. Its historical evolution therefore spans psychometrics, neurology, rehabilitation science, occupational therapy, and disability-centered accounts of everyday life.
An important contemporary development concerns the separation of symptom-contingent energy management from fixed-increment exercise progression. The 2021 NICE ME/CFS guideline (https://www.nice.org.uk/guidance/ng206/chapter/recommendations) places the person's fluctuating energy limit and risk of post-exertional malaise at the center of activity decisions. Its rejection of fixed incremental increases for ME/CFS contrasts with trials of carefully selected poststroke patients in which graded activity can be helpful. These guidelines concern different clinical mechanisms and populations; the same intervention label cannot authorize cross-diagnostic transfer. The 2025 WHO post-COVID rehabilitation guidance (https://www.who.int/teams/health-care-readiness/post-covid-19-condition) further differentiates post-exertional symptom exacerbation, exertional oxygen desaturation, cardiac impairment, and fatigue without symptom exacerbation when considering training.
Definition and Taxonomy of Fatigue
Subjective fatigue is an internal report of exhaustion, diminished energy, or heightened effort. Performance fatigability concerns measurable changes in output as activity continues under a specified task, including reduced force, slower cognitive responses, or changes in accuracy. Perceived fatigability describes rising effort or fatigue ratings under a controlled activity demand. The two latter constructs require task and time definitions: a short cognitive test after sleep loss and a prolonged physical endurance test do not measure interchangeable phenomena. Functional consequences, such as missed work or reduced community participation, add another level of observation influenced by accessibility and external support.
Physical fatigue and cognitive fatigue frequently coexist. Physical fatigue may be expressed as reduced capacity for walking, lifting, or maintaining a posture; cognitive fatigue may emerge during reading, conversation, attention-demanding work, or planning. Emotional, social, and sensory activity can also influence overall energy demands. The NICE ME/CFS framework (https://www.nice.org.uk/guidance/ng206/chapter/recommendations) explicitly incorporates cognitive, physical, emotional, and social activities into energy management. This comprehensive accounting is useful in rehabilitation because therapeutic tasks compete for the same person's time and tolerable activity exposure. It is also a clinical boundary: the evidence from ME/CFS cannot be presumed to apply with equal strength to every fatigue syndrome.
Fatigue severity, fatigue impact, fatigue interference, energy limitation, and exercise intolerance belong to different construct families. Sleepiness denotes propensity to fall asleep and may accompany fatigue without sharing its principal mechanism. General weakness, motor impairment, apathy, depressed mood, breathlessness, orthostatic intolerance, and pain may limit activity through partially distinct pathways. Clinical evaluation must identify which problems are experienced and observed rather than interpreting every difficulty as one disease process. The NICE stroke rehabilitation guideline (2023) (https://www.nice.org.uk/guidance/ng236/chapter/Recommendations) identifies the Fatigue Severity Scale, Fatigue Assessment Scale, and Modified Fatigue Impact Scale as possible instruments for poststroke assessment, reflecting the need to select validated tools for an explicit clinical context.
Symptom Variability as a Scientific Construct
Symptom variability means patterned changes in a symptom within and between individuals over time. Researchers must specify its timescale, amplitude, timing, and context. A person's fatigue can fluctuate within a morning, between weekdays and weekends, after a therapy appointment, across menstrual or sleep cycles, or over months of recovery and relapse. A single retrospective score compresses that dynamic information. Daily variability can include predictable diurnal patterns, delayed reactions after exertion, episodic flares, gradual recovery, and irregular changes associated with new illness or treatment.
The research unit is often a person-time observation rather than a single person-level score. Within-person associations ask whether fatigue is higher than that person's usual level following a given exposure; between-person associations ask whether people with greater average activity also have greater average fatigue. The two associations can differ in magnitude and direction. A highly active individual may exhibit lower long-term fatigue than a less active individual even when unusually intense activity predicts temporary worsening within that same individual's day. Aggregated cross-sectional correlations cannot resolve this divergence. Repeated measurement, temporal ordering, and context-sensitive modeling are therefore central to the subject.
The timing of observation is decisive when post-exertional worsening occurs. CDC guidance (2024) (https://www.cdc.gov/me-cfs/hcp/clinical-care/treating-the-most-disruptive-symptoms-first-and-preventing-worsening-of-symptoms.html) describes post-exertional malaise in ME/CFS as symptom worsening after physical or mental exertion, commonly appearing 12–48 hours afterward and lasting days or longer. Recording only symptoms immediately after a training session can miss the adverse response that determines future participation. By contrast, other conditions may produce transient exertion-related fatigue with recovery on a different timetable. Scientifically responsible monitoring fits the hypothesized latency and recovery interval to the syndrome under investigation.
Activity Pacing: Three Research Traditions
The first tradition is energy conservation. Its object is reducing unnecessary energetic costs through reorganizing the task, modifying the environment, prioritizing activities, using assistive equipment, changing body positions, or alternating demands. It often appears in occupational therapy and in rehabilitation of people living with neurological conditions. A relevant trial endpoint might be fatigue impact or perceived capacity to carry out daily routines, rather than increased total step count. The original six-week randomized energy conservation study in multiple sclerosis by Mathiowetz et al. (2005) (https://pubmed.ncbi.nlm.nih.gov/16193899/) represents this tradition.
The second tradition is symptom-contingent or energy-envelope pacing. Here the target is prevention or reduction of post-exertional deterioration through alignment of physical, cognitive, emotional, and social activities with an individual's variable tolerance. Plans may incorporate breaks before symptoms intensify, flexible downscaling during flares, and attention to delayed responses. The person defines meaningful goals and supplies indispensable knowledge about symptom patterns. NICE describes ME/CFS energy management as a long-term self-management strategy requiring specialist support and flexible adjustments; it states that activity is not automatically increased. NICE (2021) (https://www.nice.org.uk/guidance/ng206/chapter/recommendations). The approach represents risk-sensitive allocation under fluctuating limits, and assessment must include potential harms of exceeding those limits.
The third tradition is operant or quota-based pacing within behaviorally oriented chronic pain treatment. Its mechanisms include preplanning activities, activity-rest cycling, avoiding extreme oscillations between unusually high and low activity, and stabilizing routines so that functioning is less controlled by momentary pain ratings. In selected conditions this approach can be combined with progressively revised functional goals, provided the clinical response permits it. The 2026 systematic review by Antcliff and colleagues (https://pubmed.ncbi.nlm.nih.gov/42085181/) deliberately excluded pacing programs grounded in energy-envelope or symptom-contingent approaches; its nine included studies involved 873 participants, with heterogeneous interventions and very low confidence in effectiveness evidence. This exclusion demonstrates that the apparent unity of the word pacing can conceal distinct intervention families. Pooling these approaches as though they were one standardized treatment produces a misleading inference.
Function, Participation, and Meaningful Activity
Functioning is not exhausted by the number of tasks completed in a therapy room. A person may preserve participation through accessibility, flexible work arrangements, task delegation, mobility aids, and rest opportunities even while the underlying fatigue remains substantial. Conversely, an intervention may lower a symptom score without restoring employment, educational access, or personally valued social roles. Research outcomes should therefore include fatigue intensity, interference, safety, feasibility, autonomy, and participation as distinct variables. The parent Rehabilitation Psychology Science monograph (https://psychologybybogdanova.com/science/rehabilitation-psychology/) situates these within the broader study of disability, environment, and rehabilitation agency.
Choice of activities matters because energy expenditure has opportunity costs. Time allocated to clinical exercises, travel, and medical appointments may reduce capacity for caregiving, paid work, study, or relationships. A program may be technically achievable and socially undesirable if it consumes the person's entire daily activity budget. Rehabilitation decision-making should establish whose goals count and what outcomes a person values. Fatigue research gains explanatory depth by recording the activity foregone as well as the activity performed, together with the contextual resources that make it possible.
This perspective also changes how clinicians understand adherence. When an individual reduces an activity following a delayed crash, the behavior may represent an informed protective adjustment. When an individual limits activity because of fear of pain despite the absence of a medically relevant exertional deterioration pattern, a different behavioral mechanism may be involved. Distinguishing such situations requires longitudinal history, clinical judgment, symptom monitoring, and informed collaboration. Behavioral labels alone cannot determine the underlying mechanism or appropriate intervention.
Multilevel Mechanisms of Fatigue and Activity Limitation
Fatigue is produced and maintained within interacting physiological, cognitive, psychological, behavioral, and environmental systems. The scientific task is to identify mechanisms supported in specific populations and to avoid inferring a single biological pathway from the shared language of exhaustion. Neurological damage after stroke, neuroinflammatory disease, disturbed sleep, dysautonomia, pain, medication effects, nutritional impairment, emotional distress, deconditioning, and prolonged systemic disease may all be associated with fatigue. Their causal roles vary across individuals and clinical conditions. Multiple contributing factors can be present simultaneously, and their relevance can change during a rehabilitation course.
Mechanistic research should separate the generator of fatigue from processes that amplify its functional impact. For example, a disrupted sleep schedule may increase subjective tiredness; limited transportation can multiply the effort required for an appointment; executive dysfunction may prolong planning tasks; and a high-demand work environment may force activity into a pattern that exceeds available resources. A consequence at the level of social participation can emerge from the interaction of several modest limitations. This framework accommodates medical and psychological assessment without making one type of explanation replace another. The strongest inference follows when measured changes in a proposed mediator precede and statistically account for changes in the outcome, and when alternative pathways have been tested.
### Biological and Neurological Pathways
A neurological diagnosis establishes a disease context but does not fully specify fatigue mechanism. In multiple sclerosis, lesion burden, inflammatory processes, sleep problems, disability, depressive symptoms, and treatment effects have been studied as possible contributors; their contributions differ across samples. After stroke, fatigue may persist beyond a period of obvious motor recovery and can coexist with sleep disorders or mood difficulties. Rehabilitation studies therefore measure fatigue directly rather than treating neurological improvement as a sufficient proxy. In research on post-COVID condition, potential exertional oxygen desaturation, autonomic symptoms, post-exertional exacerbation, and cardiac impairment create additional safety-relevant subgroups. The WHO's 2025 clinical management recommendations (https://www.who.int/teams/health-care-readiness/post-covid-19-condition) require evaluation and management of exertional desaturation and cardiac impairment before physical exercise training in adults with post-COVID condition.
Biomarkers have different inferential roles: identifying a comorbidity, describing physiological response, or testing a mechanistic hypothesis. Heart rate, actigraphy, oxygen saturation, sleep measures, and cardiopulmonary testing can each contribute to a research question. No single measurement replaces the person's account of fatigue or delayed symptom worsening. An activity monitor records movement imperfectly; it does not measure cognitive exertion, sensory load, or the physiological cost of equivalent tasks for different people. A low step count can reflect environmental accessibility, illness severity, deliberate pacing, or lack of opportunity. Strong studies specify the causal estimand and the measurement validity of each marker before drawing conclusions.
### Cognitive Effort and Attention
Sustained attention, divided attention, memory retrieval, and executive control create forms of demand that are poorly represented by physical activity logs. A person who appears physically inactive may spend much of the day navigating administrative tasks, communicating with services, or concentrating during therapy. Cognitive fatigue can therefore contribute to a delayed deterioration in daily functioning even when step counts remain unchanged. In designing a repeated-measures study, investigators can pair brief fatigue ratings with task characteristics, subjective cognitive effort, task performance, and recovery intervals. Such pairing can distinguish baseline disability from change under ongoing demand and can reveal whether apparently similar activities have different costs.
Cognitive appraisal can influence perceived effort, goals, help-seeking, and tolerance for uncertainty. These relationships invite experimental work, but their discovery does not establish that fatigue itself is generated by beliefs. A well-designed mechanism study would evaluate both symptom-related physiological features and cognition in the same temporal model while preserving the validity of the underlying medical diagnosis. The potential effect of cognitive therapy on fatigue is an intervention question; its observed benefit in a selected trial does not retroactively prove that cognitive processes caused every participant's initial fatigue.
### Sleep, Pain, Mood, and Medication Burden
Sleep duration, continuity, insomnia, sleep apnea, circadian disruption, and the timing of rest can influence next-day fatigue. Pain may disturb sleep, increase the perceived cost of movement, and reduce engagement in social activities. Depression can include diminished energy or motivation, while anxiety can increase cognitive workload and interfere with restorative periods. Some prescribed medicines can produce sedation, dizziness, or cognitive slowing; other therapies can reduce fatigue by treating underlying illness. Rehabilitation assessment has to differentiate these contributors without reducing all symptoms to a single psychiatric or behavioral description.
The 2023 National Clinical Guideline for Stroke (https://www.strokeguideline.org/chapter/motor-recovery-and-physical-effects-of-stroke/) recommends periodic review of fatigue and consideration of depression, anxiety, sleep disorders, pain, and other factors that exacerbate it. The guideline's clinical approach illustrates a general principle of differential evaluation: a symptom can be important in its own right and also demand investigation of coexisting, potentially treatable disorders. Changes in medication belong under appropriately qualified prescribing oversight. Research publications should record concurrent treatments because these can confound measured rehabilitation effects.
Psychometric Assessment and Construct Validity
Fatigue measures differ in target construct and validation history. The Fatigue Severity Scale (https://pubmed.ncbi.nlm.nih.gov/2803071/) contains statements about the functional effect of fatigue and was first tested in selected neurological and systemic disease samples. The Chalder Fatigue Questionnaire (https://pubmed.ncbi.nlm.nih.gov/8463991/) emerged from a different general-practice research context. The PROMIS Fatigue item bank (https://pubmed.ncbi.nlm.nih.gov/21958919/) uses item-response modeling and supports multiple administration formats, including computerized adaptive testing. The Modified Fatigue Impact Scale, Fatigue Assessment Scale, and other disease-specific scales may emphasize physical, cognitive, or psychosocial impact differently. Researchers must examine item wording and scoring before defining a primary endpoint.
A scale can display high internal consistency while failing to capture fluctuations or exertional worsening. Reliability refers to the precision and stability appropriate to a measurement goal; validity concerns whether the evidence justifies the interpretation of scores. A validated measure in one disease population may have different item functioning in another, and a measure developed in a single language may require cross-cultural evaluation. Ceiling and floor effects, missingness, response shift, and the ability to detect meaningful change matter when interpreting long-term rehabilitation studies. A short scale administered only before and after an intervention might detect overall improvement while missing repeated crashes during treatment.
NICE's 2023 poststroke assessment recommendations (https://www.nice.org.uk/guidance/ng236/chapter/Recommendations) identify three candidate scales rather than asserting a universal winner. The guideline's supporting evidence review stresses feasibility and psychometric suitability for the poststroke population. It also distinguishes the role of standardized measurement in establishing a trial baseline from evidence that a specific intervention works. Scale selection should be declared before outcome data are examined, and clinical interpretation should consider whether observed mean change exceeds measurement error and represents functional benefit.
Ecological Momentary Assessment and Diaries
Ecological momentary assessment samples symptoms and contexts during ordinary daily life, reducing dependence on a single retrospective summary. A diary can document the kind and duration of activity, physical position, cognitive effort, perceived energy, sleep, rest, environmental conditions, and the timing of subsequent symptom worsening. Repeated ratings support analysis of person-specific patterns and broader statistical comparisons. The CDC's ME/CFS pacing resource (https://www.cdc.gov/me-cfs/pdfs/toolkit/managing-pem_508.pdf) illustrates the clinical reasoning behind activity and symptom logs: mapping exposure, delayed symptoms, and individual tolerance before adjusting activity.
Study design requires a balance between temporal resolution and participant burden. Frequent prompts can interrupt rest and become an additional source of exertion, especially in severe illness. A small number of strategically timed observations, accessible interfaces, optional assistance, and flexible recording periods may capture meaningful transitions more ethically than rigid high-frequency sampling. Researchers should report completion rates by symptom severity and investigate whether missing observations cluster during crashes. A dataset dominated by relatively well days can conceal precisely the episodes that matter most.
Diary observations remain subject to error. Participants may forget an event, interpret fatigue scales differently over time, or modify their behavior because they are being observed. Smartphone recording can favor people with stable digital access and sufficient cognitive endurance. Passive devices reduce some reporting demands but introduce their own uncertainties about nonwear, unusual activity, device placement, and whether the recorded behavior matches the scientific construct. Triangulation of patient report, sensor measurements, and task-specific performance improves interpretation when their distinct meanings are preserved.
Within-Person Statistical Models
A rigorous design specifies the timing of exposure, mediator, and outcome. Researchers might model next-day fatigue as a function of activity today while controlling for previous fatigue, sleep, illness activity, weekday, and individual baseline differences. Multilevel models can separate within-person deviations from between-person average activity. Distributed-lag methods can represent delayed worsening rather than forcing all responses into the same-day interval. Dynamic structural equation models and time-series approaches can evaluate feedback, where fatigue affects tomorrow's activity and today's activity affects subsequent fatigue. The statistical sophistication is useful only when the measurement schedule captures the physiological and behavioral processes of interest.
Confounding by indication and reverse causation are unusually important. A person may rest more because fatigue is worsening, making a naive analysis associate rest with more severe fatigue. The same problem applies to prescriptions for energy management: patients with the greatest limitations may receive the most intensive pacing support. Randomized assignment, repeated baseline periods, careful mediator measurement, and sensitivity analyses offer stronger evidence than simple observational correlations. Findings from an individual N-of-1 series may explain a person's temporal pattern but will require replication and transportability assessment before informing broader clinical policy.
Temporal variability can be summarized through within-person standard deviation, autocorrelation, event frequency, recovery duration, and transitions between activity states. Each summary loses some information; the choice should be linked to a scientific question. For delayed post-exertional episodes, the interval between an exposure and peak worsening is relevant. For sustained fatigue interference, cumulative time unable to pursue chosen activities may be more meaningful. For treatment safety, the rate and severity of prolonged setbacks matter. Studies should prespecify which patterns count as benefit or harm rather than selecting favorable metrics after data are observed.
Post-Exertional Malaise and Symptom Exacerbation
Post-exertional malaise (PEM) refers to a characteristic worsening of symptoms after exertion, often delayed and lasting longer than ordinary post-activity tiredness. It is a central feature in contemporary ME/CFS evaluation and is reported in a subset of people with post-COVID condition. Cotler and colleagues (2018) (https://doi.org/10.3390/diagnostics8030066) investigated a brief questionnaire for assessing PEM, reflecting the need for instruments designed for this phenotype. A standard fatigue severity score does not by itself establish PEM; assessment includes triggers, timing, intensity, accompanying symptoms, recovery course, and clinically relevant alternatives.
The implications for rehabilitation are substantial. A patient can complete an exercise session with acceptable immediate symptoms and experience major deterioration later. Under such conditions, an intervention's safety must be evaluated across the appropriate follow-up window. NICE recommends energy management based on fluctuating energy limits and specifically rejects fixed-increment exercise programs for ME/CFS. NICE NG206 (https://www.nice.org.uk/guidance/ng206/chapter/recommendations). The WHO's post-COVID recommendations likewise call for pacing-oriented energy conservation when post-exertional symptom exacerbation is present; exercise training for other symptom domains is conditioned on assessment of the relevant safety risks. WHO (2025) (https://www.who.int/teams/health-care-readiness/post-covid-19-condition).
The limits of current evidence also matter. A 2023 scoping review by Sanal-Hayes and colleagues (https://doi.org/10.1186/s12967-023-04587-5) included 17 heterogeneous studies of pacing in ME/CFS; some studies reported benefits, others no effect, and some detrimental effects relative to controls. The authors emphasized variable definitions, outcome choices, and methodological quality. Guidelines can justifiably favor harm prevention while direct comparative evidence about which specific pacing components yield the best long-term outcomes remains uncertain. Safety-based guidance and treatment-effect certainty address related but distinct scientific questions.
Environmental and Social Determinants of Energy Expenditure
Activity cost depends on physical and institutional environments. Travel to a clinic may require lengthy transfers, standing in queues, sensory exposure, and cognitive navigation. A remote appointment can reduce transportation burden but may require screen concentration, stable connectivity, and privacy. Household work can be divided among family members or concentrated on a person with disability. Employers can provide flexible scheduling, adjustable expectations, remote access, or predictable breaks. These arrangements alter the functional consequences of fatigue even if a symptom scale remains unchanged.
Energy conservation in this context includes technology, assistive devices, accessible architecture, and redistribution of responsibilities. The rehabilitation question extends from individual coping to the design of social participation. A person's apparent improvement might reflect better accommodation rather than physiological recovery; both changes can be valuable, but they answer different research questions. Investigators should record contextual resources, avoid interpreting low participation as a simple failure of motivation, and report outcomes that matter to the individual.
Measurement Ethics, Equity, and Accessibility
People with severe fatigue, cognitive impairment, limited language access, or unstable housing may be systematically excluded from trials that require frequent travel and prolonged questionnaires. The resulting sample can overrepresent patients able to tolerate demanding rehabilitation protocols. Research reports should disclose recruitment setting, severity distribution, exclusions, incomplete follow-up, and accessibility adaptations. Functional outcomes should account for whether the measure assumes access to paid work, private transport, devices, or caregiver support.
The burden of research itself belongs in the ethical assessment. When the study requires a participant to perform a fixed exertional test, the protocol must be matched to the relevant clinical safety profile. Participants with delayed post-exertional deterioration require monitoring that includes possible later worsening and procedures for responding to adverse events. Data from sensors and diaries raise privacy concerns because they can reveal home routines, disability status, work patterns, and social contacts. Consent, retention limits, secure processing, and accessible withdrawal procedures are part of sound rehabilitation science.
Experimental Evidence in Multiple Sclerosis
Multiple sclerosis has contributed an unusually developed literature on fatigue management because fatigue interferes with daily life while motor and other neurological impairments may vary independently. The research question is commonly whether an educational, cognitive-behavioral, or energy conservation intervention improves self-reported fatigue, self-efficacy, participation, or quality of life relative to an appropriate comparator. As with all complex rehabilitation interventions, treatment theory, active ingredients, comparator intensity, follow-up duration, and outcome definition determine what an observed difference means.
Mathiowetz and colleagues (2005) (https://doi.org/10.1191/1352458505ms1198oa) conducted a randomized controlled trial of a six-week energy conservation course for people with multiple sclerosis. This original work established a concrete intervention model emphasizing reorganization of daily tasks and personal energy resources. The randomized design improves causal identification compared with uncontrolled educational programs. However, the study is an evaluation of an intervention package rather than an isolated test of every individual pacing technique, and transportability to other conditions requires mechanistic and clinical justification.
The FACETS trial reported by Thomas and colleagues (2013) (https://doi.org/10.1136/jnnp-2012-303816) randomized 164 participants with multiple sclerosis and significant fatigue to a six-session group program plus current local practice or current local practice alone. Primary outcome data were available for 146 participants. Fatigue self-efficacy favored FACETS at two follow-up assessments, while fatigue severity improved at the later follow-up; the abstract reported a mean severity difference of −0.36, 95% CI −0.63 to −0.08, at that follow-up. The trial did not detect differences in disease-specific quality of life or quality-adjusted life-years. This separation of endpoints illustrates how an intervention can improve coping confidence and selected symptom scores while broader participation outcomes remain unchanged. Participant and facilitator blinding is intrinsically difficult in an educational group program, and the comparator does not isolate each component of the program.
A later multicenter study of FACETS+ by Gay and colleagues (2024) (https://doi.org/10.1136/jnnp-2023-331537) examined a CBT-based intervention with booster sessions in relapsing-remitting multiple sclerosis. The PubMed report describes a statistically significant 12-month between-group difference on the Modified Fatigue Impact Scale, adjusted mean difference 7.89 points (95% CI 1.26 to 14.52), favoring the intervention. This finding adds evidence concerning persistence of effects but does not establish that each component is necessary, nor does it establish equivalence for every MS phenotype. Its open-label design and population restrictions should be considered when interpreting mechanisms and external validity.
Other trials demonstrate that the treatment effect depends on the observation period. A multicenter CBT trial in multiple sclerosis (2017) (https://pubmed.ncbi.nlm.nih.gov/28528567/) randomized 91 severely fatigued participants and found a short-term fatigue reduction relative to a nurse consultation comparator at 16 weeks, but the between-group fatigue effect had diminished by the one-year assessment. The study did not find clinically relevant participation effects. Sustained symptom management may therefore require maintenance strategies, environmental accommodation, or other supports; this remains a hypothesis to investigate, rather than a conclusion that symptom improvement lacks value.
### Delivery Format and Access
Treatment access constitutes an additional outcome. A 2023 noninferiority trial (https://pubmed.ncbi.nlm.nih.gov/37489562/) allocated 166 people with multiple sclerosis to blended or face-to-face CBT. The reported difference in fatigue at 20 weeks was consistent with the prespecified noninferiority criterion. Such studies address mode of delivery rather than the comparative value of CBT against no structured intervention. They are relevant to people for whom travel consumes substantial energy. Noninferiority designs demand close attention to their margin, adherence, missing data, and whether all participants could reasonably access the digital components.
Online programs can lower transportation demands but can increase screen load and pose usability difficulties. Their evidence is also affected by the distinction between guided therapy and self-directed psychoeducation, as well as variations in local healthcare provision. Research should report the severity and accessibility characteristics of participants who did not enroll or complete treatment. Feasibility should be measured by a person's ability to use an intervention without losing capacity for other valued activities.
Evidence in Poststroke Rehabilitation
Poststroke fatigue can persist despite improving neurological function and may influence therapy adherence, independent living, and community reintegration. The strongest available experimental findings concern selected rehabilitation populations rather than all stroke survivors. The COGRAT trial by Zedlitz and colleagues (2012) (https://doi.org/10.1161/STROKEAHA.111.632117) randomized 83 participants more than four months after stroke, comparing cognitive therapy alone with cognitive therapy plus graded activity training in an assessor-blind design. Both groups improved, and the combined approach produced larger gains on selected fatigue improvement and endurance criteria. The authors reported clinical improvement on one fatigue subscale in 58% of the combined-intervention group versus 24% of the cognitive-therapy-only group. The trial supports further investigation of tailored graded activity within this selected poststroke context; it provides no justification for automatic generalization to conditions marked by post-exertional malaise.
A pilot randomized study by Nguyen and colleagues (2019) (https://doi.org/10.1080/09602011.2017.1326945) examined CBT for poststroke fatigue and sleep disturbance with blinded outcome assessment. At four months, the abstract reported a fatigue difference favoring CBT of 1.92 points on the Fatigue Severity Scale short version (95% CI 0.24 to 3.60). The reported improvements in sleep and mood underscore that fatigue interventions can target coexisting modifiable contributors. Pilot designs mainly inform feasibility and preliminary signals; larger independent trials are needed to determine robust treatment effects, subgroup differences, and durability.
Clinical guidance has moved toward structured identification and repeated assessment. The 2023 NICE stroke rehabilitation recommendations (https://www.nice.org.uk/guidance/ng236/chapter/Recommendations) suggest standardized fatigue assessment early in rehabilitation and at six-month review, with suitable measures selected for context. The National Clinical Guideline for Stroke (https://www.strokeguideline.org/chapter/motor-recovery-and-physical-effects-of-stroke/) considers education, pacing and prioritization, rest, goal setting, environmental compensation, and cautious individualized exercise among several possible measures. Recommendations are plural because interventions address different contributors and patient goals; a guideline menu should not be interpreted as evidence that every option has an equivalent causal effect.
Chronic Pain and Operant Activity Pacing
Chronic pain research has investigated oscillations between activity bursts during relatively good periods and extended inactivity following pain exacerbations. Operant pacing seeks consistent, meaningful activity through planning, scheduled pauses, and a shift away from exclusive reliance on momentary pain levels. Its rationale is behavioral: a predictable schedule may reduce extremes, support gradual functional engagement where tolerated, and alter learned associations between symptoms and behavior. These are proposed pathways requiring experimental isolation; a favorable outcome on physical functioning does not itself reveal which mechanism caused the change.
Antcliff and colleagues (2026) (https://doi.org/10.1097/AJP.0000000000001393) systematically reviewed operant-approach activity pacing interventions in adults with chronic pain. The review identified nine studies comprising ten interventions and 873 participants. Common components included preplanning, activity-rest cycling, and alternating activities or positions. Findings leaned toward improved function, with mixed pain and fatigue outcomes. The authors rated confidence in the evidence very low and described limitations created by heterogeneous designs and a vote-counting synthesis. This is an important corrective to strong claims that activity pacing has one settled, generalizable treatment effect. It also emphasizes that operant pacing trials were intentionally separated from energy-envelope and symptom-contingent approaches.
Chronic pain can coexist with post-exertional worsening, dysautonomia, or other medical conditions requiring different activity constraints. A participant's response should therefore be assessed before interpreting increased movement as improvement. The meaningful outcome may involve fewer disabling flares, greater stability of chosen activities, or less disruption to work, regardless of whether total daily movement increases. A trial designed only to maximize activity volume risks overlooking the distinction between sustainable participation and short-lived performance gains.
ME/CFS, Pacing, and the Problem of Causal Evidence
Research on ME/CFS has brought explicit attention to the delayed consequences of exertion, prolonged recovery periods, and variable day-to-day capacity. In this setting, pacing is commonly framed as energy management that reduces the risk of post-exertional malaise and supports activity within current limits. The CDC clinical guidance (https://www.cdc.gov/me-cfs/hcp/clinical-care/treating-the-most-disruptive-symptoms-first-and-preventing-worsening-of-symptoms.html) recommends individual determination of physical and cognitive limits and integration of rest with activity. The NICE 2021 guideline (https://www.nice.org.uk/guidance/ng206/chapter/recommendations) explicitly defines a flexible, person-led strategy that includes physical, cognitive, emotional, and social demands.
The empirical evidence base for particular pacing formats remains heterogeneous. Sanal-Hayes and colleagues (2023) (https://doi.org/10.1186/s12967-023-04587-5) described 17 studies in a scoping review, including three randomized trials alongside observational designs. Their synthesis found variable directions of effects, and they cautioned that methodological quality and noncomparable measurements restricted inference. The review establishes the breadth of the pacing literature and its limitations; it is not a demonstration that one standardized regimen consistently improves symptoms. A defensible interpretation separates the safety rationale for avoiding foreseeable deterioration from the more demanding claim that a specific educational intervention produces superior long-term recovery.
Historical debates about exercise interventions in ME/CFS reveal the importance of diagnostic criteria, missing-data assumptions, prespecified outcomes, participant selection, and patient-reported harms. A label such as graded exercise has been applied to several programs. NICE defines the prohibited intervention in NG206 in terms of fixed incremental increases following an initial baseline and advises against this approach in ME/CFS. It allows carefully individualized physical activity support only under specialist oversight and with flexible adjustment to energy limits. The protocol definition, rather than the reassuring or alarming connotation of the label, must guide scientific comparison.
Long COVID and Phenotype-Specific Rehabilitation
Long COVID encompasses varied symptom patterns and mechanisms. Among them are fatigue, post-exertional symptom exacerbation, cognitive difficulties, breathlessness, orthostatic complaints, and possible cardiopulmonary abnormalities. A broad diagnosis does not determine exercise tolerance. The WHO's post-COVID guidance updated in June 2025 (https://www.who.int/teams/health-care-readiness/post-covid-19-condition) makes a strong recommendation to rule out and manage exertional desaturation and cardiac impairment before considering exercise training. It conditionally recommends education and energy-conservation strategies such as pacing for post-exertional symptom exacerbation and differentiates exercise approaches in patients without such exacerbation.
The 2026 European Respiratory Society long COVID guideline (https://publications.ersnet.org/index.php/content/erj/early/2026/03/05/1399300302611-2025) similarly distinguishes individuals with post-exertional malaise from those who may tolerate personalized rehabilitation training, while describing the low certainty and modest magnitude of much of the exercise evidence. This is a clinically significant conceptual distinction: exercise-based improvement in selected participants does not prove safety for a population excluded or underrepresented because of severe PEM. Likewise, evidence of harm from exceeding energy limits in a PEM phenotype does not preclude useful and safely monitored activity in other clinical contexts.
Research designs should measure PEM at baseline and stratify or prespecify subgroup analyses accordingly, together with autonomic, cardiopulmonary, neurological, and cognitive features. Randomization alone does not ensure a trial is informative for people excluded by its safety criteria. Reporting the denominator of screened patients and reasons for nonparticipation helps readers interpret what population the evidence describes.
Cross-Condition Comparisons and Limits of Transfer
Across multiple sclerosis, stroke, chronic pain, ME/CFS, and long COVID, fatigue shares a vocabulary but differs in chronology, possible mechanisms, risks, and response to intervention. The MS literature offers randomized evidence for educational and CBT-based fatigue management with variable durability and participation effects. Selected poststroke trials suggest potential benefits from cognitive treatment and, in suitable cases, graded activity. Operant pacing in chronic pain has heterogeneous trials and very low-confidence pooled narrative evidence. In ME/CFS, energy management is embedded in guidance centered on avoiding post-exertional deterioration, while the comparative effectiveness of individual pacing components remains inadequately established. Long COVID guidance requires stratification by PEM and cardiopulmonary risk.
Scientific transfer is justified by mechanisms and populations, not by shared symptom names. Even within one diagnosis, symptoms may arise from multiple contributors. A defensible intervention study records how fatigue was defined, whether delayed exacerbation was assessed, what participants were able to tolerate, what the comparator received, and whether outcomes captured meaningful participation. The field will advance through trials that test active components and safety within clearly described phenotypes instead of using a generic diagnosis and one fatigue score as a substitute for mechanism.
Pacing as a Complex Rehabilitation Intervention
An activity pacing program is a complex intervention because multiple coordinated actions may produce an outcome: education, scheduling, rests, environmental modification, behavioral monitoring, communication with family members, adjustment of therapy intensity, and goal revision. These components may interact. A diary can improve recognition of delayed worsening; planned breaks can alter exposure; adaptive scheduling can maintain essential roles; and supportive clinical relationships can increase the person's ability to act on observations. A trial that randomizes a package estimates the overall effect of that package against a specific comparator. It does not demonstrate which component is necessary or sufficient.
Researchers should specify the intervention's logic model. For symptom-contingent pacing in an exertion-intolerant population, an intermediate outcome may be fewer delayed crashes or shorter recovery periods. For operant pacing in chronic pain, a relevant intermediate outcome may be reduced activity oscillation and more consistent engagement. For energy conservation in neurological rehabilitation, task efficiency and self-efficacy may mediate improved quality of life. The intervention is properly evaluated when the hypothesized pathway and the measured outcomes align. Labeling every package as pacing conceals these meaningful differences and makes cross-study synthesis unreliable.
### Baseline Assessment and Phenotyping
The starting point is a structured evaluation of fatigue presentation, course, associated symptoms, functional priorities, medical conditions, and environmental demands. In a clinical program, an appropriately qualified team establishes the diagnosis and investigates red flags; a research study records relevant inclusion and exclusion criteria and the clinical procedures used. The occurrence of delayed deterioration after exertion is especially important when evaluating activity progression. Studies should record the activities that commonly precipitate symptoms, including cognitive and social demands, and establish a sufficiently long monitoring interval to detect delayed patterns.
Baseline behavior should be interpreted in the context of the person's present life. A person may be performing only minimum self-care and may have little capacity for additional therapy. Another may sustain employment at the cost of severe evening exhaustion and the abandonment of other activities. These participants can present with similar average fatigue scores while requiring very different rehabilitation objectives. Baseline measurement should therefore include both fatigue and the opportunity structure of daily life, with particular attention to tasks that are necessary, valued, difficult to postpone, or unusually costly.
### Collaborative Goal Setting
Goals link the individual's priorities to observable outcomes. Examples include maintaining essential self-care, completing a course of study, participating in family life, attending therapy without a multiday symptom crash, or returning to a preferred leisure activity. The appropriate metric depends on the goal. Total physical activity may be an irrelevant or misleading proxy when the person values sustained cognitive work or predictable social engagement. A meaningful goal is specific enough to assess while allowing revision if the clinical condition changes.
Shared goal setting also addresses power in rehabilitation. Professionals contribute clinical evidence, risk assessment, and available options; the person contributes lived knowledge, preferences, and tolerable tradeoffs. Pacing plans can fail when providers schedule therapy around institutional convenience without considering the cumulative demands of travel, appointments, caregiving, or work. In such cases the intervention may require modifications in service delivery rather than additional efforts by the patient. This is a testable hypothesis about implementation and patient-reported outcomes, not an assertion that environmental modification will cure fatigue.
Designing Activity and Rest Schedules
A schedule can reduce unnecessary task switching, distribute demanding activities across days, and create predictable rest opportunities. Investigators should document whether breaks are scheduled before symptoms appear, triggered by symptoms, or incorporated into a fixed quota-based routine. These versions embody different theories. They may also produce different safety profiles when post-exertional deterioration is present. A schedule needs room for illness flares and unexpected obligations, and its evaluation should capture both symptom experience and valued participation.
Task modification may include sitting for tasks previously performed standing, dividing complex cognitive work, changing environmental lighting and sound, using assistive devices, alternating types of demand, arranging transportation, or delegating routine work. The mechanism is the reduction or redistribution of task costs. The endpoint could be successful completion of a personally chosen activity with less subsequent symptom burden, even if total work or exercise output is unchanged. Researchers should assess whether intervention benefits persist when access to assistive technology or social support changes.
The idea of an energy budget is clinically intuitive but should be treated as a practical planning metaphor unless physiological capacity has been operationally established. People rarely possess a directly measurable daily quantity of energy convertible across physical, mental, and social tasks. Their tolerance may vary with the task, environment, comorbidities, and recovery interval. A symptom-responsive plan can still be effective as a behavioral strategy without implying that researchers have measured a literal energy reservoir. This distinction prevents explanatory metaphors from hardening into unsupported biological mechanisms.
Flare-Ups, Relapse, and Safety Monitoring
Safety monitoring needs a time horizon consistent with the symptom phenotype. A trial of physical training that records only immediate adverse events may overlook delayed post-exertional deterioration, while a symptom diary that captures several subsequent days can reveal clinically important patterns. Investigators should prespecify how severity, duration, and functional consequences of worsening will be recorded. They should distinguish expected transient discomfort from a clinically meaningful setback based on the relevant condition and specialist guidance.
In ME/CFS, NICE NG206 (https://www.nice.org.uk/guidance/ng206/chapter/recommendations) recommends reducing activity to within current limits during flares, reestablishing a sustainable baseline after stabilization, and avoiding fixed incremental programs. In post-COVID rehabilitation, WHO (2025) (https://www.who.int/teams/health-care-readiness/post-covid-19-condition) calls for screening and management of exertional desaturation and cardiac impairment before exercise training, together with adapted approaches in the presence of post-exertional symptom exacerbation. These recommendations concern clinical risk and should inform study design and qualified care. This monograph does not supply an individual exercise progression protocol.
A general symptom-monitoring framework also captures changes that suggest a new or worsening medical disorder. Severe unexplained breathlessness, chest pain, fainting, or other concerning new symptoms require clinical evaluation rather than reinterpretation as ordinary rehabilitation fatigue. Study investigators should establish procedures for adverse-event review and referral. A patient reported outcome about fatigue cannot substitute for medically necessary assessment of oxygenation, cardiovascular symptoms, neurological change, infection, or other potential causes.
Rehabilitation Across the Life Course
Age, developmental stage, and responsibilities shape both fatigue experience and available coping resources. A child or adolescent may face educational and peer-participation demands that cannot be inferred from adult employment measures. Family routines, school accommodations, and the capacity to recognize early warning signs influence feasibility. In older adults, multimorbidity, frailty, medication burden, caregiving roles, sensory limitations, and fall risk can complicate interpretation of fatigue and activity restriction. Clinical evidence for one age group should not be assumed to generalize to another.
Longitudinal studies should examine transitions, including discharge from inpatient care, return to work or school, changing family support, and movement between service providers. An intervention can initially function well in a controlled rehabilitation environment and lose effectiveness when transportation, household obligations, or the availability of rest changes. Assessment of long-term participation allows investigators to distinguish transient improvements under supervised conditions from sustainable everyday functioning. These are questions for prospective cohort studies, implementation evaluations, and pragmatic randomized trials.
Digital Rehabilitation and Artificial Systems
Digital diaries, scheduling applications, wearables, and remote rehabilitation platforms can assist longitudinal data collection and patient-selected activity management. Their scientific value depends on the match between measured variable and psychological construct. Step count cannot directly measure total exertion; heart rate varies with medication, orthostatic responses, stress, and illness; and automated summaries cannot establish the diagnosis underlying a symptom change. A system that recommends activity increases based solely on previous step counts may be unsafe in a person with delayed post-exertional worsening. Decision support must use clinically appropriate constraints and remain subject to qualified human oversight.
In the Artificial Era, interactive artificial systems may help organize prompts, summarize self-reports, recognize patterns across time, and prepare discussions with clinicians. These are functions that can be evaluated through usability studies, validation against clinician-reviewed records, randomized implementation research, and safety audits. Empirical claims require data on accuracy, false reassurance, missed adverse events, privacy, accessibility, and the effect on participation. The theoretical possibility of artificial cognitive support is distinct from evidence of a deployed system's clinical benefit. This dossier treats Artificial as a broader historical context for changes in the organization of care while keeping claims about rehabilitation efficacy grounded in observable research.
Digital participation is also an equity issue. People who cannot use smartphone interfaces, who have limited connectivity, or whose severe fatigue prevents frequent logging can be excluded by designs that appear efficient to researchers. Accessible alternatives include low-burden prompts, offline recording, assistive input, and optional caregiver assistance under the person's consent. The outcome should include treatment burden: a tool that improves the quality of a graph but consumes a substantial portion of an individual's daily capacity can undermine the goals of rehabilitation.
Evidence Quality and Methodological Challenges
The literature includes well-defined randomized trials, small pilots, uncontrolled feasibility work, observational cohorts, psychometric validation, scoping reviews, and official guidelines. Each source answers a different question. A randomized comparison can estimate the effect of assignment to an intervention package under trial conditions; a pilot primarily informs feasibility and helps determine what a definitive trial would require; a diary study estimates temporal association under modeling assumptions; and an evidence review evaluates the strength and consistency of studies within its scope. Guideline recommendations additionally weigh values, harms, resources, and clinical uncertainty.
Selective reviews can provide conceptual depth but cannot support claims of exhaustive retrieval or formal evidence certainty unless they follow a prespecified protocol. The present monograph uses a critical narrative approach with a traceable set of original and synthesis sources. It describes numerical results only when supported by the linked primary record or official guidance. It does not calculate an overall effect of pacing because the construct, population, comparator, and outcomes are too heterogeneous for an unqualified pooled estimate, and no independent meta-analytic procedure was undertaken here.
### Bias and Confounding
Nonblinded educational programs are especially susceptible to differences in expectations and self-report. Trial arms can differ in contact time, group support, instructor enthusiasm, or access to usual care. Attrition may correlate with fatigue severity, leading complete-case estimates to favor the people best able to tolerate treatment. Small samples can produce imprecise effects, and the choice of multiple secondary endpoints increases the risk of chance-positive results. Long-term follow-up is necessary to determine whether initial improvements remain and whether participation or quality of life changes.
Confounding also complicates observational data. Individuals who adopt pacing may be more severely affected at baseline; those with greater resources may gain more from the same intervention; and symptom improvement can permit activity rather than follow from it. Temporal ordering, controls for baseline disease severity, appropriately justified models, and sensitivity analyses are necessary. When reporting a causal pathway, researchers should establish that the relevant mediator was measured and that serious alternative explanations have been considered.
### Outcome Selection and Interpretation
A clinically meaningful endpoint should correspond to the intervention's stated purpose. The FACETS trial demonstrated changes in fatigue self-efficacy and selected fatigue measures without clear quality-of-life advantages, while selected stroke studies found advantages on fatigue or endurance measures. These results cannot be combined under a vague conclusion that all pacing improves rehabilitation. Differences may reflect intervention mechanisms, outcomes, populations, follow-up windows, and comparator effects. Researchers should disclose prespecified primary outcomes, absolute and standardized effects where available, uncertainty intervals, adverse events, and the extent of missing data.
Participation can be measured through the fulfillment of valued roles, perceived autonomy, quality of engagement, and environmental access. Improvement may involve greater predictability and a reduction in disabling crashes rather than a larger count of activities. Research systems that reward activity volume alone risk overlooking sustainability. A strong outcome set captures symptom severity, duration of worsening, chosen activity, functioning, acceptability, equity, and safety while avoiding excessive respondent burden.
Toward a Better Experimental Program
Future randomized trials should stratify patients by clinically meaningful fatigue phenotype, especially the presence of delayed post-exertional worsening; specify whether pacing is symptom-contingent, energy conservation, or operant; identify the intervention components and comparator; and prespecify both benefit and harm outcomes. Trials should include adequate follow-up to examine sustained participation and should report differences by baseline severity, comorbidity, environment, and access to support. Factorial and component studies could test which elements of a multicomponent program drive outcomes, while implementation studies could establish whether apparently effective protocols remain usable in ordinary services.
Repeated-measures designs should connect symptoms to actual daily activities and their delayed effects. Within-person approaches offer the opportunity to characterize fluctuating patterns and perhaps improve individualized planning, but findings must undergo independent replication. Patient involvement in selecting outcomes is essential because a statistically meaningful average fatigue change may have little value if it does not preserve chosen aspects of everyday life. Future science must also determine when symptom stabilization creates conditions for carefully monitored activity expansion and when increased activity remains unsafe or undesirable.
The strongest research program will couple clinical precision with ecological validity. Its central questions concern which person, with which condition and temporal symptom profile, benefits from which pacing mechanism, under what conditions, with what tradeoffs, and for how long. The quality of that answer depends on measuring activities that people genuinely value, the environments that enable those activities, and both the immediate and delayed consequences of intervention.
Critical Scientific Synthesis
The evidence supports five conclusions about fatigue, symptom variability, and activity pacing in rehabilitation. First, fatigue is a distinct subjective and functional construct that requires direct measurement. Its relationship with physical performance, cognition, sleep, pain, mood, and participation changes with clinical condition and context. Second, temporal pattern is constitutive of the research problem. Mean fatigue scores cannot reveal whether symptoms are stable, fluctuate with daily demands, or worsen after a delayed response to exertion. Third, pacing is a family of interventions defined by purpose and mechanism: energy conservation, symptom-contingent management within tolerance, and operant planning of behavior are related methods with different assumptions and treatment goals.
Fourth, experimental studies show benefits for some specific programs and endpoints, especially in multiple sclerosis and selected poststroke populations. The presence of a randomized treatment effect establishes evidence for that intervention, comparator, participant group, measurement, and follow-up period. It does not establish a universal pacing treatment effect. Several trials show improvements in fatigue ratings or self-efficacy without equivalent improvements in participation or broader quality of life; some effects attenuate during longer follow-up. The 2023 ME/CFS pacing scoping review and the 2026 operant-pacing chronic pain systematic review both document considerable heterogeneity and important limits in inference. The field requires more component-specific randomized studies and stronger monitoring of delayed outcomes.
Fifth, safety constraints have explanatory and clinical importance. The presence of post-exertional symptom exacerbation changes the meaning of activity exposure and the acceptable risk profile of rehabilitation. Clinical guidance for ME/CFS and post-COVID condition emphasizes flexible energy management and screening for clinically relevant risks rather than automatic escalation of activity. The effect of exercise training demonstrated in other diagnoses cannot override a clearly defined condition-specific risk. The scientifically appropriate decision is based on a verified clinical phenotype, explicit intervention mechanism, collaborative objectives, and appropriate monitoring.
### Distinguishing Established Findings, Inferences, and Open Questions
Established evidence includes the existence of validated fatigue measures for selected populations, documented fatigue-related interference with rehabilitation, randomized trials showing condition- and outcome-specific benefits from several management programs, and the clinical importance of post-exertional symptom exacerbation in ME/CFS and subsets of long COVID. Strong claims about a single underlying mechanism shared across all fatigue syndromes remain beyond the evidence assessed here. Proposed within-person feedback models, activity-cost frameworks, and artificial support systems are research propositions whose predictive and intervention value must be tested.
Important unresolved questions concern whether activity pattern changes mediate clinical benefits; which baseline features predict improvement or harm; how cognitive, physical, emotional, and sensory exertion interact; whether sensor-derived data reliably capture meaningful exertion; how long activity modification benefits persist; and which outcomes best represent autonomous participation. Comparative trials need better characterization of intervention components and consistent reporting of adverse events and delayed functional deterioration. Severe and underrepresented populations demand special attention because results from high-functioning trial volunteers may have narrow applicability.
### Implications for Rehabilitation Research
Future programs should use a layered outcome architecture. The first level documents symptom severity, interference, variability, and delayed exacerbation. The second describes measurable task demands, activity patterns, rest, performance changes, and contextual exposures. The third evaluates outcomes that people regard as meaningful: community access, chosen roles, cognitive work, emotional life, relationship participation, and the sustainability of rehabilitation routines. The fourth measures equity, treatment burden, feasibility, and adverse events. Together these levels prevent the research from equating greater movement or a lower symptom score with complete rehabilitation success.
Methodological pluralism is appropriate where each method answers a defined question. Randomized studies identify intervention effects under specified conditions; qualitative research clarifies the meaning of chosen activity and the burdens of treatment; time-intensive repeated measures examine temporal organization; psychometric work establishes whether the constructs are captured adequately; and implementation studies evaluate use in real services. Their integration is a scientific synthesis grounded in convergent, distinguishable forms of evidence.
Conclusion
Fatigue, symptom variability, and activity pacing form a coherent research object because the impact of fatigue on rehabilitation is determined through time: the changing relationship between capacity, demand, symptom response, recovery, and meaningful participation. A complete scientific account asks how exhaustion is measured, which mechanisms contribute in a specific clinical condition, when activity leads to worsening or improvement, and how interventions redistribute demands without sacrificing safety or autonomy.
The research reviewed here establishes that fatigue-related rehabilitation benefits are population-specific and outcome-specific. Energy conservation and cognitive-behavioral programs have demonstrated improvements on selected endpoints in multiple sclerosis. Selected poststroke trials support additional cognitive and graded activity approaches in suitable populations. Evidence for operant pacing in chronic pain and for distinct pacing formats in ME/CFS remains heterogeneous. Current guidelines concerning post-exertional malaise, exertional desaturation, and cardiac impairment establish essential constraints for clinical application. They require careful symptom phenotyping and monitoring that reaches beyond the immediate treatment session.
The central scientific conclusion is that sustainable participation provides the integrative horizon for rehabilitation research. Symptom scores, performance measures, delayed recovery, individual goals, environmental accessibility, and the distribution of daily demands must be examined together. This understanding supports a research agenda in which fatigue management is evaluated by identifiable mechanisms, clinically appropriate comparisons, transparent evidence, and the person's durable capacity to pursue meaningful life activities.
References
- Antcliff, D., Burns, L., English, E., MacKenzie, R., Geraghty, A. W. A., Marriott-Smith, J., Stansfield, R., Storms, K., Jinks, C., & Holden, M. A. (2026). Operant approach activity pacing interventions for the management of chronic pain: A systematic review. Clinical Journal of Pain, 42(9), e1393. https://doi.org/10.1097/AJP.0000000000001393](https://doi.org/10.1097/AJP.0000000000001393)
- Cao, B., Soriano, J. B., Wang, Q., Al-Aly, Z., Gu, X., Wang, G., Leo, Y.-S., Jin, Y., Shi, Q., Ohmagari, N., Liu, E., Paredes, R., Lu, H., Kim, E. S., Zhang, D., Estill, J., Li, L., Spruyt, K., Yang, W., ... Chen, Y. (2026). Clinical practice guideline for long COVID prevention and treatment. European Respiratory Journal, 68(1), Article 2502611. https://doi.org/10.1183/13993003.02611-2025](https://doi.org/10.1183/13993003.02611-2025)
- Cella, D., Riley, W., Stone, A., Rothrock, N., Reeve, B., Yount, S., Amtmann, D., Bode, R., Buysse, D., Choi, S., Cook, K., DeVellis, R., DeWalt, D., Fries, J. F., Gershon, R., Hahn, E. A., Lai, J.-S., Pilkonis, P., Revicki, D., ... Hays, R. (2010). The Patient-Reported Outcomes Measurement Information System (PROMIS) developed and tested its first wave of adult self-reported health outcome item banks: 2005–2008. Journal of Clinical Epidemiology, 63(11), 1179–1194. https://doi.org/10.1016/j.jclinepi.2010.04.011](https://doi.org/10.1016/j.jclinepi.2010.04.011)
- Centers for Disease Control and Prevention. (2024). Strategies to prevent worsening of symptoms (ME/CFS clinical care). https://www.cdc.gov/me-cfs/hcp/clinical-care/treating-the-most-disruptive-symptoms-first-and-preventing-worsening-of-symptoms.html](https://www.cdc.gov/me-cfs/hcp/clinical-care/treating-the-most-disruptive-symptoms-first-and-preventing-worsening-of-symptoms.html)
- Centers for Disease Control and Prevention. (n.d.). Managing post-exertional malaise (PEM) in ME/CFS [Clinical toolkit]. https://www.cdc.gov/me-cfs/pdfs/toolkit/managing-pem_508.pdf](https://www.cdc.gov/me-cfs/pdfs/toolkit/managing-pem_508.pdf)
- Chalder, T., Berelowitz, G., Pawlikowska, T., Watts, L., Wessely, S., Wright, D., & Wallace, E. P. (1993). Development of a fatigue scale. Journal of Psychosomatic Research, 37(2), 147–153. https://doi.org/10.1016/0022-3999(93)90081-P](https://doi.org/10.1016/0022-3999(93)90081-P)
- Cotler, J., Holtzman, C., Dudun, C., & Jason, L. A. (2018). A brief questionnaire to assess post-exertional malaise. Diagnostics, 8(3), Article 66. https://doi.org/10.3390/diagnostics8030066](https://doi.org/10.3390/diagnostics8030066)
- de Gier, M., Beckerman, H., Twisk, J., Knoop, H., & de Groot, V. (2023). Blended versus face-to-face cognitive behavioural therapy for severe fatigue in patients with multiple sclerosis: A non-inferiority RCT. Multiple Sclerosis Journal, 29(10), 1316–1326. https://doi.org/10.1177/13524585231185462](https://doi.org/10.1177/13524585231185462)
- Gay, M. C., Cassedanne, F., Barbot, F., Vaugier, I., Thomas, S., Manchon, E., Bensmail, D., Blanchere, M., & Heinzlef, O. (2024). Long-term effectiveness of a cognitive behavioural therapy (CBT) in the management of fatigue in patients with relapsing remitting multiple sclerosis (RRMS): A multicentre, randomised, open-label, controlled trial versus standard care. Journal of Neurology, Neurosurgery & Psychiatry, 95(2), 158–166. https://doi.org/10.1136/jnnp-2023-331537](https://doi.org/10.1136/jnnp-2023-331537)
- Intercollegiate Stroke Working Party. (2023). National clinical guideline for stroke (2023 ed.): Rehabilitation and recovery—motor recovery and physical effects of stroke. https://www.strokeguideline.org/chapter/motor-recovery-and-physical-effects-of-stroke/](https://www.strokeguideline.org/chapter/motor-recovery-and-physical-effects-of-stroke/)
- Krupp, L. B., LaRocca, N. G., Muir-Nash, J., & Steinberg, A. D. (1989). The fatigue severity scale: Application to patients with multiple sclerosis and systemic lupus erythematosus. Archives of Neurology, 46(10), 1121–1123. https://doi.org/10.1001/archneur.1989.00520460115022](https://doi.org/10.1001/archneur.1989.00520460115022)
- Lai, J.-S., Cella, D., Choi, S., Junghaenel, D. U., Christodoulou, C., Gershon, R., & Stone, A. (2011). How item banks and their application can influence measurement practice in rehabilitation medicine: A PROMIS fatigue item bank example. Archives of Physical Medicine and Rehabilitation, 92(10 Suppl), S20–S27. https://doi.org/10.1016/j.apmr.2010.08.033](https://doi.org/10.1016/j.apmr.2010.08.033)
- Mathiowetz, V. G., Finlayson, M. L., Matuska, K. M., Chen, H. Y., & Luo, P. (2005). Randomized controlled trial of an energy conservation course for persons with multiple sclerosis. Multiple Sclerosis, 11(5), 592–601. https://doi.org/10.1191/1352458505ms1198oa](https://doi.org/10.1191/1352458505ms1198oa)
- National Institute for Health and Care Excellence. (2021). Myalgic encephalomyelitis (or encephalopathy)/chronic fatigue syndrome: Diagnosis and management (NICE Guideline NG206). https://www.nice.org.uk/guidance/ng206/chapter/recommendations](https://www.nice.org.uk/guidance/ng206/chapter/recommendations)
- National Institute for Health and Care Excellence. (2023). Evidence reviews for the optimal tool for the assessment of fatigue: Stroke rehabilitation in adults (update), evidence review B. https://www.ncbi.nlm.nih.gov/books/NBK600101/](https://www.ncbi.nlm.nih.gov/books/NBK600101/)
- National Institute for Health and Care Excellence. (2023). Stroke rehabilitation in adults (NICE Guideline NG236). https://www.nice.org.uk/guidance/ng236/chapter/Recommendations](https://www.nice.org.uk/guidance/ng236/chapter/Recommendations)
- Nguyen, S., Wong, D., McKay, A., Rajaratnam, S. M. W., Spitz, G., Williams, G., Mansfield, D., & Ponsford, J. L. (2019). Cognitive behavioural therapy for post-stroke fatigue and sleep disturbance: A pilot randomised controlled trial with blind assessment. Neuropsychological Rehabilitation, 29(5), 723–738. https://doi.org/10.1080/09602011.2017.1326945](https://doi.org/10.1080/09602011.2017.1326945)
- Sanal-Hayes, N. E. M., McLaughlin, M., Hayes, L. D., Mair, J. L., Ormerod, J., Carless, D., Hilliard, N., Meach, R., Ingram, J., & Sculthorpe, N. F. (2023). A scoping review of pacing for management of myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS): Lessons learned for the long COVID pandemic. Journal of Translational Medicine, 21, Article 720. https://doi.org/10.1186/s12967-023-04587-5](https://doi.org/10.1186/s12967-023-04587-5)
- Schwartz, J. E., Jandorf, L., & Krupp, L. B. (1993). The measurement of fatigue: A new instrument. Journal of Psychosomatic Research, 37(7), 753–762. https://doi.org/10.1016/0022-3999(93)90104-N](https://doi.org/10.1016/0022-3999(93)90104-N)
- Thomas, S., Thomas, P. W., Kersten, P., Jones, R., Green, C., Nock, A., Slingsby, V., Smith, A. D., Baker, R., Galvin, K. T., & Hillier, C. (2013). A pragmatic parallel arm multi-centre randomised controlled trial to assess the effectiveness and cost-effectiveness of a group-based fatigue management programme (FACETS) for people with multiple sclerosis. Journal of Neurology, Neurosurgery & Psychiatry, 84(10), 1092–1099. https://doi.org/10.1136/jnnp-2012-303816](https://doi.org/10.1136/jnnp-2012-303816)
- van den Akker, L. E., Beckerman, H., Collette, E. H., Twisk, J. W. R., Bleijenberg, G., Dekker, J., Knoop, H., & de Groot, V. (2017). Cognitive behavioral therapy positively affects fatigue in patients with multiple sclerosis: Results of a randomized controlled trial. Multiple Sclerosis, 23(11), 1542–1553. https://doi.org/10.1177/1352458517709361](https://doi.org/10.1177/1352458517709361)
- World Health Organization. (2025). Clinical management of COVID-19: Living guideline, June 2025. https://iris.who.int/bitstream/handle/10665/381920/B09467-eng.pdf?sequence=1](https://iris.who.int/bitstream/handle/10665/381920/B09467-eng.pdf?sequence=1)
- Zedlitz, A. M. E. E., Rietveld, T. C. M., Geurts, A. C., & Fasotti, L. (2012). Cognitive and graded activity training can alleviate persistent fatigue after stroke: A randomized, controlled trial. Stroke, 43(4), 1046–1051. https://doi.org/10.1161/STROKEAHA.111.632117](https://doi.org/10.1161/STROKEAHA.111.632117)