RESEARCH DOSSIER
Depression, Anxiety, and Emotional Recovery in Rehabilitation
Abstract. Depression, anxiety, and emotional recovery shape rehabilitation after neurological injury, physical trauma, and chronic illness. This fundamental research monograph investigates the constructs, historical foundations, psychological mechanisms, measurement constraints, longitudinal trajectories, intervention evidence, and social conditions governing emotional outcomes during rehabilitation. A critical narrative synthesis draws on original validation research, observational cohorts, randomized clinical trials, Cochrane reviews, and authoritative guidelines. The evidence identifies substantial emotional morbidity after stroke and traumatic brain injury, condition-specific problems of somatic symptom overlap, and complex relationships among mood, participation, pain, self-efficacy, and care transitions. Some treatments improve selected depression endpoints, while major trials demonstrate that emotional and functional recovery can diverge and adverse events matter. Evidence certainty varies by condition, measure, population, and intervention. The synthesis specifies methodological limits, clinically meaningful distinctions, and priorities for accessible measurement and rigorous longitudinal trials. This work reports no newly collected patient data, registered systematic review, new meta-analysis, or independent clinical trial.
Keywords: rehabilitation psychology; depression; anxiety; emotional recovery; post-stroke depression; traumatic brain injury; spinal cord injury; psychosocial adjustment; symptom assessment; somatic overlap; participation; rehabilitation outcomes; cognitive behavioral therapy; randomized clinical trials; biopsychosocial model; health inequality; Artificial Era
The Scientific Object of Emotional Recovery in Rehabilitation
Rehabilitation is a sustained process through which people with illness, injury, or disability pursue functioning, participation, and personally meaningful activities in changing environments. Depression and anxiety can affect this process at every stage: they may precede an injury, appear as reactions to new restrictions, accompany a neurological or medical condition, or persist after physical performance has improved. Emotional recovery therefore refers to change over time in mood, fear, perceived control, engagement, relationships, and the capacity to pursue valued life directions. It includes clinical improvement when a depressive or anxiety disorder is present, but its relevant outcomes also include ordinary distress, participation, and psychological well-being.
The central research object is the interaction between emotional symptoms and rehabilitation trajectories. The same individual may show improving mobility alongside worsening hopelessness, or may recover confidence and community participation while a permanent impairment remains. A coherent research account must measure emotional, functional, and social outcomes separately before examining their associations. The International Classification of Functioning, Disability and Health (WHO, 2001) (https://www.who.int/standards/classifications/international-classification-of-functioning-disability-and-health) provides an organizing language: bodily functions and structures, activities, participation, personal factors, and environmental conditions jointly describe functioning. Psychological distress can influence each domain and can be influenced by each domain.
This monograph asks how depression and anxiety are defined in rehabilitation populations; what is known about their distribution and course; how symptoms are distinguished from injury-related fatigue, sleep disruption, and neurological deficits; which models explain trajectories; which interventions show credible effects; and how emotional outcomes should be incorporated into interdisciplinary services. It integrates original measurement studies, cohort observations, randomized trials, evidence syntheses, and clinical guidance in a critical narrative review. It presents no new patient dataset, experimental trial, prospectively registered systematic review, or independent meta-analysis.
Historical Formation of the Research Field
Early medical rehabilitation concentrated heavily on impairment, survival, and restoration of physical capacity. Psychological questions became increasingly explicit as clinicians encountered differences between measured functional recovery and lived adjustment. The growth of rehabilitation psychology, psychiatric epidemiology, and behavioral medicine made it possible to investigate depression and anxiety as measurable dimensions of rehabilitation rather than treating emotional responses as incidental aspects of disability. This history was cumulative and interdisciplinary: psychiatry established diagnostic and treatment traditions, psychology studied coping and agency, occupational science examined participation, and disability scholarship brought social and environmental barriers into the analysis.
Engel's (1977) biopsychosocial proposal (https://doi.org/10.1126/science.847460) articulated a framework in which biological illness, subjective experience, and social conditions all contribute to clinical understanding. This framework has special relevance to rehabilitation because the same lesion may have different consequences depending on access to assistance, work demands, family relationships, income, or rehabilitation resources. Bandura's (1977) self-efficacy theory (https://doi.org/10.1037/0033-295X.84.2.191) added a testable psychological mechanism: expectations concerning one's capacity to carry out actions can shape effort, persistence, and avoidance. These theories identify dimensions worth studying; their historical influence should not be mistaken for direct proof that any particular intervention will improve depression.
The development of medical-setting measures was another methodological milestone. Zigmond and Snaith (1983) (https://doi.org/10.1111/j.1600-0447.1983.tb09716.x) designed the Hospital Anxiety and Depression Scale for medical outpatients, reducing dependence on bodily symptoms that might have nonpsychiatric causes. Later, Kroenke, Spitzer, and Williams (2001) (https://doi.org/10.1046/j.1525-1497.2001.016009606.x) and Spitzer and colleagues (2006) (https://doi.org/10.1001/archinte.166.10.1092) developed brief depression and anxiety measures widely used in health services. The chronology establishes a key change: emotional morbidity became accessible to structured detection, serial observation, and trial endpoints, although validation in one clinical setting never guarantees equivalent validity in all rehabilitation populations.
Definitions and Construct Boundaries
Depressive symptoms include low mood, diminished interest or pleasure, self-critical thinking, impaired concentration, disrupted sleep, reduced energy, and suicidal thoughts in different combinations. A depressive disorder is established through a structured clinical assessment applying the relevant diagnostic criteria, duration, impairment, exclusions, and differential diagnosis. In rehabilitation, persistent pain, fatigue, medication effects, sleep disorders, neurological injury, and physical dependence can contribute to overlapping symptoms. Therefore a questionnaire score is evidence of symptom burden and an invitation to evaluate, rather than a diagnosis by itself.
Anxiety encompasses apprehension about future threats, physiological arousal, worry, vigilance, and forms of avoidance. The content of anxiety matters: fear of another stroke, fear of falling, uncertainty about returning to work, and generalized worry are related yet distinct phenomena. Trauma-related intrusions and hyperarousal, delirium, cognitive disinhibition, and fear associated with actual medical danger require different clinical explanations. Post-stroke emotionalism, sometimes involving involuntary laughing or crying, further illustrates why visible emotional expression does not automatically indicate a depressive episode. Rehabilitation assessment must combine symptom description, context, functional consequences, temporal history, and specialist judgment.
Emotional recovery is best operationalized as a longitudinal, multidimensional construct. Symptom remission, clinical recovery, improved quality of life, self-efficacy, reduced avoidance, restoration of rewarding activities, and increased participation are potential outcomes with different meanings. A person can experience grief over lost opportunities while sustaining adaptive engagement, and can report symptom improvement while still encountering disabling barriers. Research must report which outcome changed and over what time scale. A single global adjustment score often conceals these distinctions.
A Biopsychosocial and Participation-Based Framework
The biopsychosocial framework directs attention to the interaction of neurological injury, inflammation, pain, sleep, treatment burden, psychological interpretation, family relationships, and environmental opportunity. An association between lesion burden and depression does not establish that depression is entirely neurologically determined; similarly, associations between isolation and depression do not establish that psychosocial influences alone explain the course. The appropriate explanatory level depends on the question, the available measurements, and the study design.
The WHO ICF (2001) (https://www.who.int/standards/classifications/international-classification-of-functioning-disability-and-health) distinguishes impairment from participation restrictions and environmental influences. For example, two people with similar walking limitations may differ greatly in their ability to visit friends because one has accessible transport and the other does not. Their levels of anxiety about public travel may also differ. These observations suggest specific causal pathways worth testing: environmental accessibility may increase meaningful activity, which may reduce avoidance and improve mood; or mood improvement may increase activity despite persistent barriers. Longitudinal and intervention designs are needed to evaluate the direction of those pathways.
A participation-based view also changes outcome selection. A trial can show a statistically significant reduction in depression questionnaire scores yet leave employment, family roles, and social participation unchanged. Conversely, an environmental change can increase participation without a large short-term shift in symptom scores. Neither result is inherently contradictory. They address different constructs in an interacting system and call for distinct measures rather than a single proxy for overall rehabilitation success.
Appraisal, Control, Self-Efficacy, and Behavioral Activation
Appraisal describes how a person interprets an event's significance, predictability, and implications for future goals. Rehabilitation often involves uncertainty: recovery may be nonlinear, functional gains may plateau, and individual prognosis may remain unsettled. Perceived control can therefore fluctuate with medical developments, communication quality, and earlier experiences. Catastrophic predictions about future incapacity or recurrent injury may promote avoidance, while realistic planning and supported experimentation may support engagement. These are plausible process models, not universal rules governing every person's emotional course.
Bandura's self-efficacy framework (https://doi.org/10.1037/0033-295X.84.2.191) predicts that successful performance and credible feedback can modify expectations of capability. In rehabilitation, mastery may involve managing transfers, using assistive technology, communicating needs, or returning to a social setting. The theoretically relevant construct is confidence in a specified action under specified conditions, rather than a generalized declaration that all outcomes are controllable. Interventions based on self-efficacy should test whether changes in confidence precede changes in activity and mood and whether those associations remain after accounting for impairment severity and access to rehabilitation.
Behavioral activation offers a complementary account. Low mood, pain, and fear can restrict participation in rewarding activity; reduced engagement can in turn limit opportunities for positive experiences and social reinforcement. Rehabilitation already contains behavioral tasks, but not every therapeutic exercise is intrinsically rewarding or personally meaningful. Emotional recovery may depend on connecting graded activity with valued roles and attainable choices. The intervention question is whether adjusting activity patterns improves depression beyond the improvement expected from natural medical recovery or standard rehabilitation.
Method of the Present Scientific Synthesis
This research is a critical narrative synthesis, organized by construct, research design, clinical population, and rehabilitation outcome. It prioritizes original validation studies, longitudinal observational evidence, randomized controlled trials, Cochrane reviews, and official multidisciplinary guidance. For each major claim it distinguishes the actual measurement, studied population, comparator, period of follow-up, and inferential strength. Where only abstracts or accessible summaries are available, conclusions remain confined to their documented findings.
A review of heterogeneous conditions must resist false numerical aggregation. Stroke, spinal cord injury, traumatic brain injury, chronic pain, and musculoskeletal trauma differ in their mechanisms and measurement difficulties. Even within a single condition, a questionnaire threshold, structured psychiatric interview, and treatment-seeking record generate different prevalence estimates. Consequently, this monograph does not produce a single global prevalence figure for depression or anxiety across rehabilitation. When study-specific estimates are discussed, their population and assessment method remain explicit.
The literature also contains historically influential concepts whose empirical support varies and current guidelines whose recommendations are broader than the trial evidence for an individual treatment. The synthesis separates theoretical explanations, observational associations, experimental effects, and policy recommendations. It identifies unanswered questions rather than translating absence of evidence into proof of ineffectiveness. These distinctions structure the subsequent analysis of clinical trajectories and interventions.
Depression Screening and Clinical Assessment
Screening is a structured method for detecting possible symptoms in people who might otherwise remain unidentified. It is particularly valuable in rehabilitation because attention can be consumed by physical impairment, medical procedures, or discharge planning. A symptom scale can standardize the opening question, but interpretation requires a clinical interview, consideration of chronology, and assessment of the contribution of other medical conditions. The PHQ-9 validation study (Kroenke et al., 2001) (https://doi.org/10.1046/j.1525-1497.2001.016009606.x) established useful properties in primary care and obstetrics-gynecology settings; those results cannot be imported uncritically to aphasia, serious neurological injury, or inpatient rehabilitation.
Several depression items overlap with common consequences of injury: poor sleep, low energy, altered appetite, impaired concentration, and reduced psychomotor activity. Overlap can produce false-positive symptom classification when injury-related features are interpreted as affective symptoms. An opposite error occurs when clinicians dismiss persistent sadness or anhedonia as an inevitable response to illness. The diagnostic task therefore requires evidence about mood and interest, onset, duration, cognitive content, previous episodes, medication exposure, and the functional significance of symptoms. The choice between inclusive and selective scoring of somatic items changes sensitivity and specificity; researchers should describe the rule they use.
A positive screening result should lead to an appropriate assessment pathway rather than an automatic treatment decision. This includes identifying urgency, evaluating suicidal thoughts in context, assessing access to communication, and documenting whether symptoms reflect a current depressive disorder, subthreshold depression, trauma reaction, delirium, grief, or a combination. NICE guidance on rehabilitation after traumatic injury (NG211) (https://www.nice.org.uk/guidance/ng211/chapter/recommendations) explicitly integrates psychological assessment and support into rehabilitation planning and recognizes that symptoms can emerge or recur during transitions.
Anxiety Screening and the Diversity of Fear
The GAD-7 (Spitzer et al., 2006) (https://doi.org/10.1001/archinte.166.10.1092) was developed as a brief measure of generalized anxiety symptoms. Its utility in rehabilitation is practical but not comprehensive: injury-related fear can be restricted to falling, pain exacerbation, medical procedures, future attacks, social exposure, or a particular task. A low score on a generalized-worry measure may coexist with disabling task-specific avoidance. Assessment should therefore identify feared consequences, avoidance behavior, safety behaviors, perceived probability of harm, and actual medical restrictions.
Fear can have protective functions. People are sometimes asked to avoid movements because tissue healing, instability, or other medical risks make them unsafe. A behavioral researcher who labels every cautious action as maladaptive avoidance may mistake medically appropriate protection for a psychological target. Conversely, persistent avoidance after clinicians judge the feared activity safe may reinforce loss of activity and confidence. The boundary is determined by a combination of medical evaluation, patient meaning, and observed patterns over time.
Anxiety and depression also coexist frequently, but co-occurrence has several interpretations. One disorder may precede the other, common stressors may raise both symptom dimensions, measurement items may overlap, or underlying pain, uncertainty, and disrupted sleep may jointly contribute. Their correlation is clinically important and does not itself prove a single shared mechanism. Research designs capable of separating specific symptom trajectories are essential for understanding whether changes in one emotional domain precede changes in the other.
The Hospital Anxiety and Depression Scale and Somatic Overlap
Zigmond and Snaith (1983) (https://doi.org/10.1111/j.1600-0447.1983.tb09716.x) proposed the Hospital Anxiety and Depression Scale for medically ill outpatients. Its two subscales were developed with the explicit problem of emotional assessment in a medical context in view. This makes the HADS a historically central instrument for rehabilitation research. Yet the scale remains a screening and severity measure, and its psychometric performance depends on the population, language, severity of impairment, and intended classification threshold.
The somatic-overlap problem can be expressed as a measurement model. Let an observed item response be influenced by depressive affect, disease-related symptoms, response style, and situational context. A symptom questionnaire sums responses, but the sum does not uniquely identify the latent causes that produced them. In stroke, motor slowing may affect reported activity; in spinal cord injury, sleep and energy can be changed by pain, autonomic disturbance, or medication; after brain injury, cognitive difficulties may affect the person's ability to interpret or recall question content. Different instruments place different weights on these sources of variance.
The venlafaxine spinal cord injury trial (Fann, Bombardier, Richards, et al., 2015) (https://doi.org/10.1001/jamapsychiatry.2014.2482) illustrates why outcome selection is consequential. The trial evaluated the 17-item Hamilton Depression Rating Scale and a core-symptom Maier subscale excluding somatic items; significant between-group improvement occurred on the latter but not the full Hamilton scale. The methodological lesson extends beyond this drug: selecting a scale capable of distinguishing relevant change from stable impairment-related symptoms may determine which effects are observable, and it changes what the result means.
Accessible Communication and Assessment Validity
Many rehabilitation populations cannot use standard self-report measures without adaptation. Aphasia may impair language comprehension and expression after stroke; traumatic brain injury can disrupt attention, memory, insight, and processing speed; sensory impairment may block ordinary forms and interviews. A missing questionnaire is therefore an assessment-access problem, not evidence that the person lacks emotional symptoms. Excluding such participants from research produces a selected population and can systematically distort estimates of depression, anxiety, and response to treatment.
Accessible assessment may involve supported communication, simplified wording, visual presentation, additional processing time, interpreter support, alternative response modalities, or carefully selected observer information. Each adaptation presents its own validity question. Caregiver reports can provide useful behavioral observations but may diverge from the person's internal experience. A clinician may observe withdrawal while the person reports frustration rather than sadness. A well-designed assessment records both perspectives and keeps the sources distinguishable rather than pooling them into an apparent consensus.
Researchers should report the proportion of eligible individuals excluded because of communication or cognitive difficulties; whether adapted measures were validated; the handling of missing data; and whether assessment mode changed over repeated visits. An apparent decrease in depression could emerge because participants with persistent impairment were less likely to complete follow-up. Sensitivity analysis, participant retention, and measurement invariance are therefore substantive components of evidence quality, especially in neurological rehabilitation.
Longitudinal Measurement and Recovery Trajectories
Emotional recovery unfolds through transitions from acute care to inpatient rehabilitation, discharge, community reentry, and longer-term life adjustment. A cross-sectional questionnaire at one point estimates current symptom burden but reveals little about individual trajectories. Repeated assessments can identify early-onset symptoms, delayed onset, persistence, recurrence, and partial remission. They also allow researchers to test whether emotional changes precede activity changes or whether functional gains precede improved mood.
A meaningful longitudinal design requires consistent measurement intervals, baseline mental-health history, and sufficiently long follow-up to observe later transitions. It should distinguish changes within individuals from differences between individuals. A population can show a stable mean score while some participants recover and others deteriorate. Group means alone conceal this heterogeneity; latent class and growth-curve approaches may help describe patterns, provided sample size, model identification, and external validation justify them.
Retention is a major issue. People with severe impairments, socioeconomic instability, or high emotional burden may be less able to attend follow-up appointments. If dropout depends on the outcome, complete-case analysis can bias estimates. Studies should specify missing-data assumptions, describe whether missingness plausibly depends on unobserved emotional status, and offer sensitivity analyses. Reliable measurement across time further requires attention to response shift: a person's understanding of well-being and participation can change as values, expectations, and adaptive strategies develop.
Measuring Outcomes Beyond Symptom Scores
Depression and anxiety scores are indispensable for certain clinical and research questions, but they represent only part of rehabilitation success. Functional independence, participation in home life, work and education, health-related quality of life, satisfaction with personal goals, social connection, perceived agency, and use of health services each represent different outcomes. A useful trial design identifies which is primary and which are secondary before analysis, limits multiplicity-related false positives, and includes outcome measures relevant to the lived rehabilitation context.
Measurement should also distinguish performance under standardized conditions from performance in everyday settings. A person may demonstrate the capacity to walk a short distance in a clinical facility while remaining reluctant to travel alone. Environmental barriers, confidence, assistive devices, access to transport, and social support may intervene between capacity and actual participation. Emotional outcomes can plausibly mediate or moderate these pathways, but this requires an explicit analysis plan and repeated measurement, not retrospective storytelling around correlated changes.
The best outcome battery is thus multidomain and parsimonious. It includes clearly defined depression and anxiety endpoints, medically relevant functional indicators, a participation measure, and a measure of the hypothesized psychological mechanism. It accounts for assessment burden and recognizes that persons with fatigue or cognitive difficulties may find lengthy instruments distressing or inaccessible. A smaller valid set of measures is scientifically more useful than a long collection of poorly interpretable questionnaires.
Depression and Anxiety After Stroke
Stroke provides one of the most extensively studied populations in rehabilitation psychiatry. Neurological injury, altered functional capacity, social disruption, communication difficulty, prior mental-health history, and vascular or other medical conditions may all contribute to depressive and anxiety symptoms. Hackett and Pickles (2014) (https://doi.org/10.1111/ijs.12357) pooled 61 observational studies involving 25,488 people and reported an estimated frequency of depression of 31% (95% confidence interval 28%–35%). This estimate is anchored to their search ending in May 2013, to the included instruments and recruitment strategies, and to the observational populations; it is not a timeless or universal prevalence for every survivor of stroke.
The meaning of the estimate depends on how depression was defined. Some studies used structured diagnostic interviews and others used symptom scales or thresholds. Timing varied from acute stroke to several years afterward, and participation of individuals with severe aphasia or cognitive problems could be restricted. These design differences produce heterogeneity that a pooled point estimate cannot remove. The contribution of this literature is a strong demonstration that depressive morbidity is common enough to deserve systematic attention across rehabilitation settings, together with evidence that its reported frequency is sensitive to sampling and ascertainment.
Anxiety after stroke forms a related but separately measured field. Campbell Burton and colleagues (2013) (https://doi.org/10.1111/j.1747-4949.2012.00906.x) systematically reviewed observational studies of anxiety frequency after stroke. Such evidence establishes that anxiety is a recurring clinical and research concern, while its magnitude depends on assessment method and time since stroke. Fear of another event, falls, dependence, and social exposure may all accompany generalized anxiety. Measures and interventions must therefore be connected to the particular concern under investigation rather than assuming that a global anxiety score captures all functionally important fear.
Emotional Distress After Traumatic Brain Injury
Traumatic brain injury offers a different research configuration. Emotional change can reflect psychological appraisal of injury consequences, disruption of neural systems involved in affective and executive regulation, premorbid vulnerabilities, social circumstances, and treatment-related effects. Disturbances in initiation, attention, irritability, sleep, and fatigue can complicate clinical interpretation. An assessment focused only on outward behavior can mistake cognitive or neurological difficulties for either deliberate disengagement or a psychiatric symptom. Diagnosis requires temporal context and structured assessment, including collateral information when appropriate.
In a prospective cohort study, Bombardier and colleagues (2010) (https://doi.org/10.1001/jama.2010.599) followed adults after traumatic brain injury and documented substantial incidence of major depressive disorder in the first year. Because the investigation used longitudinal follow-up and psychiatric assessment, it contributes evidence about the emergence of diagnosable depression rather than merely cross-sectional distress. Its population and care setting determine the generalizability of its findings. People with milder injuries, severe communication impairments, or different health care access may have different rates and trajectories.
Clinical intervention evidence also demands careful reading. In a randomized study involving 100 adults with major depression after complicated mild to severe traumatic brain injury, Fann and colleagues (2015) (https://doi.org/10.1089/neu.2014.3423) compared telephone or in-person cognitive behavioral therapy with usual care. The pooled CBT comparison did not demonstrate a statistically significant advantage on the clinician-rated Hamilton depression outcome. Some patient-reported analyses favored telephone CBT, and participants reported greater improvement and satisfaction. These findings support feasibility and possible benefits for selected outcomes, while falling short of proving a robust effect on the trial's principal clinician-rated endpoint.
Spinal Cord Injury and Mood Outcomes
Spinal cord injury often involves enduring changes in mobility, sensation, autonomic function, pain, bowel or bladder management, sexuality, and environmental accessibility. Emotional trajectories are influenced by the interaction of these changes with autonomy, social support, participation, and economic resources. A generic model in which depression follows inevitable stages of disability adjustment cannot adequately represent the observed variability. Some individuals experience sustained distress, others adapt with changing priorities, and many move between periods of improvement and renewed difficulty as personal circumstances change.
Williams and Murray (2015) (https://doi.org/10.1016/j.apmr.2014.08.016) conducted a meta-analysis of the prevalence of depressive disorders after spinal cord injury. The synthesis established the importance of diagnostic definitions and methodological choices for prevalence estimation. Its results should be interpreted against the included study designs, rather than used to predict the mental-health course of a newly injured individual. An individual's prior mental-health history and current circumstances remain essential to clinical assessment.
The randomized Fann and colleagues (2015) venlafaxine XR trial (https://doi.org/10.1001/jamapsychiatry.2014.2482) enrolled 133 adults with spinal cord injury and depression in a 12-week comparison with placebo. The estimated improvement was not statistically different on the full 17-item Hamilton Depression Rating Scale (between-group treatment effect 1.0, 95% CI −1.4 to 3.4, P = .42), whereas the Maier core-symptom subscale favored treatment (effect 1.6, 95% CI 0.3–2.9, P = .02). The trial reported improvement on a measure of disability and one more frequent adverse effect, blurred vision. This is an instructive example of outcome sensitivity and adverse-event monitoring, not a basis for universal medication choice or individual prescribing.
Musculoskeletal Trauma, Pain, and Uncertain Recovery
Serious musculoskeletal trauma may change the person's sense of bodily reliability, independence, appearance, social role, and capacity for employment. Repeated operations, prolonged healing, and uncertainty about chronic pain can create emotionally difficult periods even when the initial injury is medically stabilized. Depression and anxiety may coexist with fear of reinjury, financial strain, and loss of meaningful activity. Their relationship to functional recovery is likely bidirectional and requires designs that can distinguish symptom effects from impairment severity and socioeconomic conditions.
Pain-related fear may promote guarded movement and limited activity, but some restrictions reflect legitimate medical advice. Research should therefore identify the relevant phase of tissue healing, the movement's actual medical risk, the person's appraisal, and the frequency of avoided behavior. An intervention that promotes indiscriminate movement could be inappropriate; psychologically informed graded activity is meaningful only within a safe rehabilitation plan. This logic connects psychological mechanisms to medically grounded constraints and patient-specific goals.
NICE guideline NG211 (https://www.nice.org.uk/guidance/ng211/chapter/recommendations) recommends attention to psychological support, recognition of depression and anxiety, and multidisciplinary coordination following traumatic injury. It explicitly notes that emotional problems can recur or deteriorate during discharge or changes in setting. This recommendation has practical importance: an apparently successful inpatient episode does not establish that emotional recovery will continue when professional support decreases, daily responsibilities return, or accessible transport and home assistance become uncertain.
Chronic Conditions and Heterogeneous Rehabilitation Populations
Rehabilitation extends beyond acute events. People with chronic respiratory disease, cardiac disease, persistent pain, neurological conditions, cancer-related impairments, and progressive illnesses may receive rehabilitation while confronting long-term symptom variability and uncertain prognosis. The psychological object is therefore broader than adaptation to a single injury. It includes recurrent activity limitations, shifting treatment requirements, role renegotiation, and the challenge of maintaining participation under changing conditions.
General depression guidelines for adults living with chronic physical conditions, including NICE CG91 (https://www.nice.org.uk/guidance/cg91/chapter/Recommendations), help organize care but do not provide identical evidence for every diagnosis. It is methodologically unsound to generalize an effect from a post-stroke trial to a person with a progressive neuromuscular disease merely because both attend rehabilitation. The severity and causes of fatigue, cognitive effects of medication, social context, prognosis, and medically achievable activity goals can all alter intervention feasibility and the meaning of observed change.
The transdiagnostic approach has a precise role here. It identifies repeated psychological processes such as uncertainty, avoidance, low reward exposure, reduced agency, and disrupted social roles across different medical diagnoses. It then asks whether their effects and treatment responses are sufficiently consistent to support shared intervention principles. A credible transdiagnostic model retains condition-specific moderators rather than replacing clinical variation with an overly uniform account of rehabilitation distress.
Depression, Anxiety, and Suicidality as Separate Assessment Problems
Depressive symptoms are associated with suicide risk in many clinical populations, but neither a depression scale score nor a physical impairment by itself establishes an individual's immediate risk. Assessment requires direct attention to suicidal ideation, intent, planning, previous behavior, protective factors, changing circumstances, and capacity to access timely care. Rehabilitation professionals must be able to activate a clear, documented response pathway involving qualified clinical personnel when concerning information emerges.
The value of this principle is particularly evident at transitions. Discharge from a protected setting can bring greater independence and new stressors at the same time. A patient who anticipates returning to an inaccessible environment or to an unsupported caregiving arrangement may encounter distress that was not visible during inpatient care. NICE NG211 (https://www.nice.org.uk/guidance/ng211/chapter/recommendations) recommends asking about self-harm and suicidal thoughts regularly during psychological assessment, especially at discharge and changes in setting. This is a service-level safety recommendation, not an assertion that all people undergoing rehabilitation are at comparable risk.
Research on suicidality must also avoid conflating different outcomes. Passive thoughts about death, active suicidal ideation, medically documented attempts, and mortality records have distinct ascertainment methods and base rates. Studies require valid measurement, appropriate ethical protections, and careful reporting of event counts and follow-up. A model aimed at explaining depression severity cannot automatically be treated as a validated suicide prediction tool.
Psychological Therapies in the Rehabilitation Setting
Psychological interventions in rehabilitation commonly adapt techniques from cognitive behavioral therapy, behavioral activation, problem-solving approaches, supportive counseling, and other evidence-informed forms of psychotherapy. Their clinical purposes may include reducing depressive symptoms, modifying avoidance, improving coping with uncertainty, supporting valued activities, and addressing the social consequences of illness. Treatment feasibility depends on cognitive capacity, communication access, fatigue, physical discomfort, session scheduling, transportation, and the extent of multidisciplinary coordination.
Evidence from one population cannot establish universal efficacy. The Fann et al. (2015) (https://doi.org/10.1089/neu.2014.3423) traumatic brain injury trial shows why the outcome and comparison must be specified: combined CBT did not significantly outperform usual care on the principal clinician-rated depression scale, although certain patient-reported and telephone-treatment analyses showed more encouraging findings. A clinical program can therefore be feasible and acceptable without demonstrating a clear effect on every endpoint. The distinction is especially important where small samples, heterogeneous injuries, varying attendance, and pragmatic treatment settings influence statistical precision.
A rehabilitation-specific psychotherapy model must also define the role of the therapist in relation to the wider team. Psychological sessions may explore hopelessness and avoidance while occupational and physical therapy create opportunities for graded participation. Consistent goals can make treatment more coherent, whereas contradictory messages about risk or capability can increase uncertainty. Any asserted benefit of integrated care, however, needs direct evaluation using well-defined comparators, rather than assumption that greater multidisciplinary contact necessarily produces superior outcomes.
Treatment Evidence for Depression After Stroke
The most substantial aggregated intervention evidence in this field comes from stroke research. Allida and colleagues (2023) (https://doi.org/10.1002/14651858.CD003437.pub5) updated a Cochrane review of pharmacological, psychological, non-invasive brain stimulation, and combination approaches for depression after stroke. The review included 65 trials, 72 comparisons, and 5,831 participants. It found evidence suggesting fewer participants met depression criteria at treatment completion after pharmacological therapy, psychological interventions, or some combinations, with substantial uncertainty across comparisons. The review's authors judged the certainty of evidence low to very low and highlighted small samples, methodological limitations, and heterogeneous populations.
That review also identified harms. Participants receiving pharmacological interventions reported central nervous system and gastrointestinal adverse effects more frequently than controls in the relevant trials. Estimates for non-invasive brain stimulation did not consistently show improved depression outcomes. Consequently, conclusions should describe both the possibility of symptom improvement and the limitations of the evidence used to infer it. This is different from claiming that any intervention reliably improves neurological function or long-term community participation.
Research reports often summarize a successful trial by its direction of effect, omitting distinctions between risk ratios, mean changes, response, remission, and functional outcomes. A rigorous synthesis reports the endpoint actually studied, the comparator and follow-up interval, and whether confidence intervals exclude clinically unimportant effects. It also records how many participants withdrew or were excluded. The scientific question is not simply whether depression can be treated after stroke; it is which approach benefits which subgroup, on which outcomes, with which harms, and under what rehabilitation conditions.
Medication Trials and the Meaning of Clinical Outcomes
Medication can be appropriate within qualified clinical management of an established depressive or anxiety disorder, but rehabilitation introduces special considerations. Medical comorbidity, polypharmacy, fall risk, cognitive impairment, pain treatment, and changing physical status may influence tolerability, monitoring, and treatment decisions. This monograph examines published evidence rather than providing dosing instructions or individualized selection of an antidepressant. Regulatory indications and safety information must be checked against current official product labeling when a clinical recommendation is made.
In the spinal cord injury venlafaxine XR trial (https://doi.org/10.1001/jamapsychiatry.2014.2482), the difference between the full Hamilton depression measure and the Maier subscale should prevent overgeneralization. The primary full-scale comparison was not statistically significant, while the core-symptom subscale and a disability measure favored the active-treatment group. Outcomes are not interchangeable, and their divergence can reflect differences in construct coverage, baseline somatic burden, measurement precision, or chance. The proper interpretation preserves each result and the population from which it arose.
Preventive use of medication after neurological injury requires a separate evidentiary judgment. A treatment that reduces subsequent depressive symptoms may still fail to improve physical recovery, while adverse effects can outweigh a hypothesized functional benefit. Trials of fluoxetine after stroke, including the large FOCUS, EFFECTS, and AFFINITY programs, illustrate why mood outcomes and global functional recovery should not be collapsed into a single claim. The broader lesson is that prevention of depression and improvement of physical independence are different causal questions requiring separate randomized evidence.
Behavioral Activation, Graded Participation, and Fear
Behavioral activation provides an experimentally testable pathway between depression and rehabilitation outcomes. Reduced activity can lower opportunities for mastery, enjoyment, and contact; low mood can make further activity more difficult. Rehabilitation services can translate this model into collaborative identification of valued tasks, graduated participation, monitoring of barriers, and attention to the subjective meaning of success. The hypothesis is that increased engagement in rewarding or meaningful activities contributes to emotional change, possibly in combination with other therapies.
Fear-based avoidance requires more careful differentiation. A person may avoid walking because of an objectively high risk of injury, because of a past fall and an exaggerated prediction of future catastrophe, or because transportation and the physical environment are inaccessible. These pathways require different interventions and combinations of medical, psychological, and environmental change. The research should distinguish fear reduction, perceived control, actual activity, and adverse events. Evidence of reduced fear without evidence of safe participation remains a limited outcome.
The mechanism can be evaluated by mediator analysis in randomized studies: if an intervention changes meaningful activity, does that change statistically account for later improvement in depressive symptoms? Such models are strongest when the mediator is measured before the outcome and relevant confounding is addressed. Simple correlations between activity and mood cannot establish mediation. Likewise, self-efficacy may function as a moderator rather than mediator: confidence at baseline could identify participants who respond differently to a particular rehabilitation approach.
Interdisciplinary Coordination and Psychological Safety
Rehabilitation involves interdependent professional decisions. Physical therapists may assess movement and balance, occupational therapists may evaluate daily activities, physicians may address pain and medical safety, speech-language professionals may support communication, and psychologists may assess mood, cognition, adjustment, and behavior. Emotional distress can interfere with attendance and participation, but the appropriate response is an accessible assessment of its causes rather than a moral judgment that the person is insufficiently motivated.
NICE NG211 (https://www.nice.org.uk/guidance/ng211/chapter/recommendations) describes a coordinated approach to psychological and physical assessment after traumatic injury. It recommends emotional support from practitioners and referral to suitably experienced psychological services when distress significantly affects rehabilitation. Its guidance also emphasizes assessment during changes of care setting. These principles can inform service organization, while the effectiveness of a specific local model remains an empirical question.
Coordination should preserve patient decision-making and privacy. People may value different goals: independence in self-care, relief from pain, return to family roles, or development of a viable routine with assistance. Goals imposed by a clinician can appear technically successful while remaining personally irrelevant. Research on integrated interventions should therefore include the person's preferences and meaningful outcomes, describe who made decisions about therapy, and examine whether participation increased beyond clinical performance measures.
Digital and Remote Delivery of Psychological Care
Remote psychological support can reduce travel burdens for people with mobility restrictions, inaccessible transport, rural residence, or complex schedules. Its potential advantages are observable: sessions may be easier to attend, specialist expertise may reach underserved areas, and continuity may improve after discharge. Its limitations include privacy, internet access, hearing or visual impairment, cognitive load, lack of a safe setting, and difficulty establishing a therapeutic relationship through available technology. Access and effectiveness are distinct research outcomes.
The Fann et al. (2015) (https://doi.org/10.1089/neu.2014.3423) trial directly investigated telephone-administered CBT after traumatic brain injury. The trial supported feasibility and some patient-reported benefits, but the overall primary clinician-rated effect was nonsignificant. This evidence is a basis for further careful testing of remote approaches, not for asserting that telephone and in-person modalities are equivalent across conditions or that automation can replace specialist assessment. Remote treatment should be evaluated in relation to treatment fidelity, emergency protocols, communication accessibility, and retention.
The Artificial Era also introduces digital cognitive aids, adaptive interfaces, and conversational support systems into rehabilitation. Their psychological significance can be studied through adherence, perceived autonomy, trust, anxiety, disclosure, and changes in therapeutic engagement. Claims that artificial systems provide effective psychotherapy for medically complex rehabilitation populations require population-specific trials and safety assessment. A philosophy of artificial agency is conceptually relevant to the emerging technological setting; it is not itself clinical evidence of treatment effectiveness.
Identity, Loss, Narrative, and Meaning
Major illness and injury can alter the everyday continuity of a person's roles. Someone accustomed to being a wage earner, parent, athlete, or independent traveler may have to renegotiate those activities. Emotional responses may include grief, shame, anger, fear, relief, hope, and a revised sense of possibility. A clinical study should not assume that adaptation follows a fixed sequence or that accepting disability means abandoning previous ambitions. Identity change can involve maintaining important commitments through new means, discovering activities, or resisting social expectations that underestimate capabilities.
Narrative interviews can reveal aspects of emotional recovery hidden by symptom inventories. Participants may describe the significance of being treated as capable, receiving accurate information, retaining choice, or rebuilding relationships. These narratives are evidence about lived experience and perceived mechanisms. They do not, without an appropriate comparison, establish that a particular psychological intervention causes improved depression outcomes. Combining qualitative and quantitative methods can help define meaningful endpoints before conducting trials.
The distinction between disorder and understandable distress matters here. Sadness after loss may be consistent with the circumstances and still deserve support. Persistent major depression may require clinical evaluation and treatment even when the social causes of distress are intelligible. A person-centered response examines intensity, duration, functional impairment, risk, and the person's own priorities. This permits clinical precision without treating the person's emotional response as a defect of motivation or as an unavoidable consequence of impairment.
Health Inequality, Structural Barriers, and Access
Financial resources, insurance coverage, employment protection, accessible transportation, housing, language, disability discrimination, and geographic proximity to specialist services can all shape rehabilitation opportunities. They may influence emotional burden directly and indirectly through the capacity to maintain relationships, obtain care, and exercise choices. Because these factors are correlated with illness severity and other characteristics, observational studies should adjust for plausible confounders and report how social conditions were measured. Broad claims that one variable alone causes depression are rarely supported by such designs.
Selection bias is a serious concern in clinical cohorts. People who enter specialist rehabilitation centers may differ from those who receive home-based services or no formal treatment. Trials can further exclude participants with communication impairment, severe comorbidities, or limited capacity to attend appointments. An intervention found effective in a highly selected urban sample may not have the same reach in remote communities or under-resourced settings. External validity depends on explicit description of who was included, who was not, and which resources the intervention required.
Implementation research should examine referral completion, missed appointments, waiting times, treatment acceptability, accessible assessment, clinician workload, and continuity after discharge. Emotional recovery may remain poorly supported even when screening instruments are available if there is no pathway from detection to timely assessment. This is a measurable health-system problem. Studies that report screening uptake alone cannot establish that patients ultimately received effective or wanted care.
Age, Development, and Cultural Context
Age and developmental stage alter the social meaning and practical tasks of rehabilitation. Adolescents may confront school, peers, identity formation, and dependence on caregivers; working-age adults may focus on employment and family responsibilities; older adults may face multiple chronic conditions and transitions in living arrangements. These are population tendencies that require individual assessment rather than fixed assumptions about goals. A scale validated in middle-aged general medical outpatients is not automatically appropriate for a child or a person with advanced cognitive impairment.
Cultural context shapes the interpretation and reporting of emotional symptoms, beliefs about illness, expectations of family support, the language of distress, and preferences for treatment. Translation of a questionnaire is only a first step toward valid cross-cultural measurement. Researchers need evidence concerning comprehension, factor structure, threshold interpretation, response style, and clinical utility in the target language and population. When instruments work differently across groups, apparent prevalence differences may partly reflect measurement inequivalence.
A life-course model can accommodate these differences without dividing people into rigid categories. It considers prior mental-health history, cumulative socioeconomic experience, changing responsibilities, injury-related limitations, and the availability of new opportunities. Emotional recovery may be influenced by past adversity, present resources, and anticipated future roles. Carefully measured moderation effects can test whether age, social position, cultural context, or condition type alters intervention response.
Care Transitions and the Dynamics of Relapse
Rehabilitation pathways include discontinuities. Transfer from an acute ward to a rehabilitation unit, discharge home, changes in employment, and the end of funded therapies can each transform the person's environment. A measure obtained at discharge often captures the state of the individual within a service system that is about to change. Later anxiety may reflect uncertainty about managing alone; later low mood may reflect reduced contact, a painful setback, or unmet expectations. These possibilities argue for follow-up models that include transition events.
NICE NG211 (https://www.nice.org.uk/guidance/ng211/chapter/recommendations) specifically warns that psychological difficulties can recur or deteriorate during discharge and relocation and recommends regular attention to anxiety and depression during rehabilitation reviews. This guidance supports continuity of assessment. A study of an intervention delivered entirely within inpatient care should therefore avoid projecting its short-term effects into long-term community adjustment without evidence.
Rehabilitation planners can treat transitions as measurable exposure points. Research can record dates of discharge, service interruption, recurrence of symptoms, change in social support, and functional progress. It can test whether a structured follow-up pathway improves symptom detection or care continuity. Any claim of reduced depression or hospitalization must still be substantiated by a suitably designed comparison. The mere presence of a follow-up program does not establish benefit.
The Artificial Era and Emerging Rehabilitation Environments
Digital systems can organize rehabilitation schedules, store symptom histories, support remote communication, and help clinicians interpret patient-reported outcomes. These tools create new research questions about whether increased information access improves perceived control, whether automated reminders support engagement, and how technology changes reliance on clinicians or caregivers. For people with communication or motor impairments, the design of interfaces may determine whether an emotional assessment is possible at all. Accessibility, interpretability, and privacy are therefore part of the psychological research object.
Human–artificial interaction also introduces issues of agency and trust. A conversational system may offer information or perceived companionship, but the meaning of such interaction should be evaluated through measurable outcomes such as understanding, adherence, disclosure, autonomy, social participation, and adverse psychological consequences. A system that gives reassuring responses is not thereby shown to provide effective treatment for depression after stroke or traumatic brain injury. Clinical benefit requires independent evidence in the population for whom the intervention is intended.
Within the broader Artificial Era, rehabilitation psychology can contribute an empirical account of how nonbiological systems enter human care environments. It can identify which roles are suitable for aids, which decisions require specialist human responsibility, and which forms of cognitive partnership patients find helpful. These are hypotheses and design questions situated within a philosophical framework, not established clinical effects. Research in this area should use predefined safety criteria, robust comparators, accessibility testing, and transparent documentation of system limitations.
Critical Evaluation of Observational Evidence
Observational studies provide the principal evidence about the frequency, distribution, and course of emotional symptoms during rehabilitation. They are indispensable for establishing that a problem is common, describing temporal patterns, and identifying subgroups that need closer examination. Their limits arise from selection, measurement, unmeasured confounding, and the inability of association alone to establish a causal pathway. A finding that depression predicts poorer participation may reflect the effect of depression on activity, the effect of severe disability on mood, a shared cause such as pain, or a changing mixture of these influences.
The Hackett and Pickles (2014) (https://doi.org/10.1111/ijs.12357) stroke meta-analysis demonstrates both the strength and constraints of pooled observational evidence. The 31% estimate draws from many cohorts and shows substantial burden, but the studies differ in time from stroke, recruitment, instruments, and thresholds. Pooling can improve precision around an average while leaving unexplained variation between populations. The scientifically appropriate conclusion describes what was estimated and the limits of extending that estimate to individuals or other diagnoses.
Prospective cohorts improve temporal understanding but still face attrition, historical treatment changes, and incomplete baseline assessment. In the Bombardier et al. (2010) (https://doi.org/10.1001/jama.2010.599) traumatic brain injury cohort, the repeated evaluation of major depression is more informative about emergence over time than a single questionnaire. Yet differences in clinical access and inclusion criteria require consideration before applying findings to all persons with brain injury. For causal interpretation, study reports should specify plausible confounders, the temporal ordering of variables, and the handling of missing assessments.
Randomized Trials, Statistical Precision, and Harms
Random assignment reduces systematic differences between comparison groups when implemented and analyzed appropriately. The clinical meaning of trial results nevertheless depends on the intervention, comparator, outcome definition, follow-up, adherence, sample size, and missingness. A statistically significant change on a secondary scale can provide useful evidence about that scale while leaving a primary outcome uncertain. Multiple testing and selective reporting can exaggerate apparent efficacy if a single favorable contrast receives disproportionate attention.
The spinal cord injury venlafaxine XR trial (Fann et al., 2015) (https://doi.org/10.1001/jamapsychiatry.2014.2482) and the traumatic brain injury CBT trial (Fann et al., 2015) (https://doi.org/10.1089/neu.2014.3423) each illustrate the distinction between primary and secondary results. The venlafaxine study did not find a significant difference on the full Hamilton scale but did on a core-symptom subscale. The pooled CBT comparison after brain injury did not achieve a significant clinician-rated depression advantage, while selected patient-reported comparisons were encouraging. Neither pattern justifies simplifying the result into unconditional effectiveness or ineffectiveness.
Larger preventive stroke trials reinforce the need to separate domains. The FOCUS Trial Collaboration (2019) (https://doi.org/10.1016/S0140-6736(18)32823-X), EFFECTS Trial Collaboration (2020) (https://doi.org/10.1016/S1474-4422(20)30219-2), and AFFINITY Trial Collaboration (2020) (https://doi.org/10.1016/S1474-4422(20)30207-6) investigated fluoxetine after acute stroke. They did not establish improvement in global functional recovery. EFFECTS reported less newly diagnosed depression but more fractures and hyponatremia; AFFINITY reported more falls, fractures, and seizures. These trials demonstrate that improvement in one emotional endpoint may coexist with unchanged functional recovery and meaningful harms. They do not support routine use of fluoxetine solely to improve physical recovery after stroke.
Evidence Synthesis, Replication, and Selective Publication
Systematic reviews can improve understanding by locating eligible studies, applying prespecified inclusion rules, evaluating bias, and quantitatively or narratively integrating results. However, a pooled estimate is only as reliable as its contributing trials and the assumptions used to combine them. A review of depression interventions after stroke may group psychologically, pharmacologically, and procedurally different treatments. If estimates from highly heterogeneous designs are summarized without clear distinctions, readers may infer a degree of commonality that the data do not support.
The Allida et al. (2023) (https://doi.org/10.1002/14651858.CD003437.pub5) Cochrane review included 65 trials with 5,831 participants and rated key treatment comparisons as low- or very-low-certainty evidence. This judgment is essential to interpreting apparent benefits: small trials, inconsistent methods, and incomplete reporting make broad clinical generalizations premature. Even when a pooled effect has a favorable direction, its uncertainty can remain important for practice. An evidence review should display harms and attrition alongside efficacy, rather than treating adverse events as peripheral information.
Publication bias may arise when small positive trials are more likely to appear in the literature than null findings. Selective outcome reporting can occur if researchers highlight results after observing the data. Statistical tests for funnel plot asymmetry have limited interpretability with few heterogeneous trials; trial registration, published protocols, and access to original outcome definitions provide more direct safeguards. Independent replication across settings and populations is especially valuable because a treatment may work differently when delivered with fewer resources, to people with greater impairment, or by staff with different expertise.
Construct Validity and Causal Models
A central scientific problem is whether the same questionnaire score reflects the same underlying emotional construct across conditions. Fatigue following chemotherapy, disrupted sleep after spinal cord injury, and slowed performance after traumatic brain injury can elevate symptom totals for reasons unrelated to a primary depressive disorder. Comparing average scores across these groups without testing construct equivalence may create misleading differences. Measurement invariance and clinical validation are therefore relevant to every cross-condition synthesis.
A plausible causal model includes injury severity, treatment burden, pain, participation barriers, emotional symptoms, and prior vulnerabilities. Some paths may operate in both directions. Mood can affect adherence and perceived capability; impaired participation can change mood through loss of reinforcement or social contact. Mediating variables such as self-efficacy, activity avoidance, or available support need repeated measurement and clearly specified timing. Merely adjusting for every observed variable can introduce statistical bias if a variable is itself a consequence of the exposure; causal diagrams can help clarify the target comparison.
Clinical implementation further complicates causal inference. Services that provide more psychological support may treat people with more severe symptoms, creating confounding by indication in observational comparisons. Conversely, those who complete an intensive program may be healthier, wealthier, or more motivated than those who discontinue. Analyses restricted to program completers can therefore overestimate benefit. Randomized designs, careful baseline measurement, pragmatic implementation trials, and transparent sensitivity analyses offer complementary ways to address these problems.
Unresolved Questions and Research Priorities
Future studies should characterize trajectories of anxiety and depression from the earliest phase of rehabilitation through community participation, with repeated measures that distinguish symptom dimensions, diagnostic status, and functional outcomes. Research should include participants with severe communication or cognitive impairments through validated accessible procedures whenever ethically and methodologically possible. Exclusion and missing-data reporting are critical because the least accessible populations may also have the greatest unmet needs.
Intervention studies need larger, well-characterized samples, credible comparators, transparent prespecified outcomes, and sufficient follow-up to test durable effects. Studies should determine whether an emotional intervention also improves participation, quality of life, or care continuity and should identify adverse effects and opportunity costs. Heterogeneity is not merely a nuisance to average away: condition, injury severity, prior depression, social support, treatment access, and rehabilitation stage may genuinely moderate response. General principles should be built from replicated patterns rather than assumed in advance.
A further priority is integration of qualitative accounts with quantitative measurement. What people consider meaningful emotional recovery may differ from a researcher's chosen symptom threshold. Participatory design can improve relevance, though its causal benefits must also be evaluated. Research on remote and artificial support tools should examine privacy, dependency, accessibility, safety escalation, and impact on human professional relationships in addition to convenience or satisfaction. These questions extend rehabilitation psychology into the Artificial Era while keeping evidence requirements explicit.
Conclusion
Depression, anxiety, and emotional recovery are central dimensions of rehabilitation because they interact with bodily health, psychological processes, access to opportunities, and the reconstruction of daily life. The evidence demonstrates substantial emotional morbidity after conditions such as stroke and traumatic brain injury and establishes the need to assess psychological functioning across care transitions. At the same time, emotional symptoms, psychiatric diagnoses, and participation outcomes require distinct operational definitions. A change in one domain is not proof of change in another.
The strongest theoretical account is multidimensional. Injury characteristics, pain and fatigue, perceived control, learning, avoidance, social relationships, environmental barriers, and institutional arrangements can each influence trajectories. Their relative contribution varies by person and condition and must be evaluated empirically. Observational studies describe frequency and longitudinal association; randomized trials examine causal treatment effects; qualitative research describes meaning and experience; and guidelines organize responsible practice. Scientific synthesis depends on retaining those differences.
Existing treatment research supports the possibility of benefit from psychological and pharmacological approaches in selected contexts, while the certainty of much post-stroke intervention evidence remains low to very low. Large trials of fluoxetine for global functional recovery after stroke show why mood-related outcomes and physical rehabilitation outcomes cannot be conflated, and why adverse events must be considered. The next stage of research requires accessible assessment, explicit causal models, rigorous trials, person-defined outcomes, and long-term observation beyond discharge. Emotional recovery in rehabilitation is therefore best understood as the evolving relationship between clinical symptoms, meaningful participation, and the conditions that enable people to live and act within their changing circumstances.
References
- AFFINITY Trial Collaboration. (2020). Safety and efficacy of fluoxetine on functional outcome after acute stroke (AFFINITY): A randomised, double-blind, placebo-controlled trial. The Lancet Neurology, 19(8), 651–660. https://doi.org/10.1016/S1474-4422(20)30207-6](https://doi.org/10.1016/S1474-4422(20)30207-6)
- Allida, S. M., Hsieh, C.-F., Cox, K. L., Patel, K., Rouncefield-Swales, A., Lightbody, C. E., House, A., & Hackett, M. L. (2023). Pharmacological, non-invasive brain stimulation and psychological interventions, and their combination, for treating depression after stroke. Cochrane Database of Systematic Reviews, 2023(7), CD003437. https://doi.org/10.1002/14651858.CD003437.pub5](https://doi.org/10.1002/14651858.CD003437.pub5)
- Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191–215. https://doi.org/10.1037/0033-295X.84.2.191](https://doi.org/10.1037/0033-295X.84.2.191)
- Bombardier, C. H., Fann, J. R., Temkin, N. R., Esselman, P. C., Barber, J., & Dikmen, S. S. (2010). Rates of major depressive disorder and clinical outcomes following traumatic brain injury. JAMA, 303(19), 1938–1945. https://doi.org/10.1001/jama.2010.599](https://doi.org/10.1001/jama.2010.599)
- Campbell Burton, C. A., Murray, J., Holmes, J., Astin, F., Greenwood, D., & Knapp, P. (2013). Frequency of anxiety after stroke: A systematic review and meta-analysis of observational studies. International Journal of Stroke, 8(7), 545–559. https://doi.org/10.1111/j.1747-4949.2012.00906.x](https://doi.org/10.1111/j.1747-4949.2012.00906.x)
- EFFECTS Trial Collaboration. (2020). Safety and efficacy of fluoxetine on functional recovery after acute stroke (EFFECTS): A randomised, double-blind, placebo-controlled trial. The Lancet Neurology, 19(8), 661–669. https://doi.org/10.1016/S1474-4422(20)30219-2](https://doi.org/10.1016/S1474-4422(20)30219-2)
- Engel, G. L. (1977). The need for a new medical model: A challenge for biomedicine. Science, 196(4286), 129–136. https://doi.org/10.1126/science.847460](https://doi.org/10.1126/science.847460)
- Fann, J. R., Bombardier, C. H., Richards, J. S., Wilson, C. S., Heinemann, A. W., Warren, A. M., Brooks, L., McCullumsmith, C. B., Temkin, N. R., Warms, C., & Tate, D. G. (2015). Venlafaxine extended-release for depression following spinal cord injury: A randomized clinical trial. JAMA Psychiatry, 72(3), 247–258. https://doi.org/10.1001/jamapsychiatry.2014.2482](https://doi.org/10.1001/jamapsychiatry.2014.2482)
- Fann, J. R., Bombardier, C. H., Vannoy, S., Dyer, J., Ludman, E., Dikmen, S., Marshall, K., Barber, J., & Temkin, N. (2015). Telephone and in-person cognitive behavioral therapy for major depression after traumatic brain injury: A randomized controlled trial. Journal of Neurotrauma, 32(1), 45–57. https://doi.org/10.1089/neu.2014.3423](https://doi.org/10.1089/neu.2014.3423)
- FOCUS Trial Collaboration. (2019). Effects of fluoxetine on functional outcomes after acute stroke (FOCUS): A pragmatic, double-blind, randomised, controlled trial. The Lancet, 393(10168), 265–274. https://doi.org/10.1016/S0140-6736(18)32823-X](https://doi.org/10.1016/S0140-6736(18)32823-X)
- Hackett, M. L., & Pickles, K. (2014). Part I: Frequency of depression after stroke: An updated systematic review and meta-analysis of observational studies. International Journal of Stroke, 9(8), 1017–1025. https://doi.org/10.1111/ijs.12357](https://doi.org/10.1111/ijs.12357)
- Kroenke, K., Spitzer, R. L., & Williams, J. B. W. (2001). The PHQ-9: Validity of a brief depression severity measure. Journal of General Internal Medicine, 16(9), 606–613. https://doi.org/10.1046/j.1525-1497.2001.016009606.x](https://doi.org/10.1046/j.1525-1497.2001.016009606.x)
- National Institute for Health and Care Excellence. (2009). Depression in adults with a chronic physical health problem: Recognition and management (Clinical guideline CG91). https://www.nice.org.uk/guidance/cg91](https://www.nice.org.uk/guidance/cg91)
- National Institute for Health and Care Excellence. (2022). Rehabilitation after traumatic injury (NICE guideline NG211). https://www.nice.org.uk/guidance/ng211](https://www.nice.org.uk/guidance/ng211)
- Spitzer, R. L., Kroenke, K., Williams, J. B. W., & Löwe, B. (2006). A brief measure for assessing generalized anxiety disorder: The GAD-7. Archives of Internal Medicine, 166(10), 1092–1097. https://doi.org/10.1001/archinte.166.10.1092](https://doi.org/10.1001/archinte.166.10.1092)
- Williams, R., & Murray, A. (2015). Prevalence of depression after spinal cord injury: A meta-analysis. Archives of Physical Medicine and Rehabilitation, 96(1), 133–140. https://doi.org/10.1016/j.apmr.2014.08.016](https://doi.org/10.1016/j.apmr.2014.08.016)
- World Health Organization. (2001). International Classification of Functioning, Disability and Health (ICF). https://www.who.int/standards/classifications/international-classification-of-functioning-disability-and-health](https://www.who.int/standards/classifications/international-classification-of-functioning-disability-and-health)
- Zigmond, A. S., & Snaith, R. P. (1983). The hospital anxiety and depression scale. Acta Psychiatrica Scandinavica, 67(6), 361–370. https://doi.org/10.1111/j.1600-0447.1983.tb09716.x](https://doi.org/10.1111/j.1600-0447.1983.tb09716.x)
Social Relationships, Caregiving, and Emotional Adjustment
Rehabilitation occurs within networks of relationships. Family members, partners, friends, peers, professional caregivers, and community institutions may facilitate or constrain recovery. Assistance with transportation, communication, or daily tasks can create opportunities for participation; the same arrangements can create conflict when preferences and responsibilities are unspoken. Emotional recovery therefore has an interpersonal dimension in addition to individual affect. Research should specify whether perceived social support, actual received support, relationship quality, or social participation is being measured, because these constructs are related without being identical.
Caregivers may experience distress of their own, especially when rehabilitation involves uncertain prognosis, changes in employment, financial burden, or sustained care demands. The person's emotional symptoms and caregiver strain can be associated in both directions. Longitudinal dyadic research can investigate reciprocal patterns, whereas single-informant cross-sectional data cannot identify the direction of influence. A psychologically informed intervention should recognize the roles of consent, privacy, decision authority, and accessible family communication rather than assuming that all additional family involvement is beneficial.
Social support also interacts with institutional conditions. A patient may have an attentive family yet lack accessible housing, community transportation, suitable employment, or affordable follow-up services. Psychosocial adjustment cannot be inferred solely from the quality of close relationships. The WHO ICF (https://www.who.int/standards/classifications/international-classification-of-functioning-disability-and-health) provides a basis for connecting interpersonal resources with structural facilitators and barriers while leaving empirical questions about effect magnitude open.