Almost 800,000 people in the United States have a stroke every year (Benjamin et al., 2018), and it is currently the fifth leading cause of death in the United States (Herpich & Rincon, 2020). Some emergency therapies, like intravenous tissue plasminogen activator or thrombectomy, have contributed to survivorship; however, most stroke survivors are left with devastating disabilities and require inpatient rehabilitation before discharge.
Stroke survivors admitted to inpatient rehabilitation may have a right or left hemiparesis, cognitive impairment, dyspraxia, and visuospatial hemineglect and are more likely to experience a fall (Yang et al., 2021). There is also evidence to suggest that stroke survivors who fall are more likely to experience a physical injury (Wei et al., 2019). Injuries from falls in the stroke rehabilitation population range from 5% (Walsh et al., 2016) to 50% (Maeda et al., 2015). Even if a small percentage of falls result in physical injury, the consequences of that injury can be catastrophic in this population. Length of stay could be extended, and the patient may be unable to participate in much needed rehabilitation activities (Lee et al., 2018). There can also be significant psychological injury, leading to fear of falling and comorbid immobility, which negatively affects one's quality of life and longevity at discharge (Natan et al., 2016). Although there have been decades of fall prevention research, a precise, concise, and comprehensive method for reducing falls in the stroke rehabilitation population has yet to be developed (Quigley, 2016).
A phenomenon that has been theoretically linked to increasing the likelihood of falling in the stroke rehabilitation population is anosognosia for hemiplegia (AHP). Anosognosia for hemiplegia is defined as lack of knowledge of disease or disability in relation to a hemiplegic limb. It typically manifests as a delusional denial or recognition that a limb or side of the body is paralyzed (Antoniello & Gottesman, 2020). Believed to be a failure of the sensory-motor feedback loop, AHP causes an individual to receive inaccurate sensory information from the plegic limb that indicates that a movement has been completed as intended. This results in an insistence that the plegic hemibody is functioning normally (Langer & Bogousslavsky, 2020). It is not uncommon for those with AHP to obstruct early rehabilitative efforts and refuse to participate in treatments that could considerably improve prognosis (Cherney, 2006). It is also common for individuals with AHP to be unrealistic concerning their ability to return home without assistance (Orfei et al., 2007) and disregard appropriate safety measures intended to prevent falls (Kortte et al., 2015).
Though AHP is a relatively common occurrence following a stroke (Vidovic et al., 2019), there is no evidence in the literature that suggests AHP is being assessed or measured by nurses during a stroke rehabilitation admission or stay. This condition is different, though often occurs with unilateral neglect. With neglect, there is an impairment and inattention to stimuli in the hemispace contralateral to the lesion in the cerebral cortex. Anosognosia for hemiplegia, however, is defined as a motor impairment and is caused by a disconnection between the predicted sensory consequences of a movement and the actual sensory feedback that occurs. Individuals with AHP receive incorrect sensory feedback that they have completed a movement as desired when no movement has occurred at all (Kortte & Hillis, 2009). The most current AHP literature comes from the neuropsychological discipline (Antoniello & Gottesman, 2020; Barrett, 2021; Kirsch et al., 2021; Langer & Bogousslavsky, 2020). Investigating the potential link between AHP and fall events could lead to a more reliable measure by which to stratify stroke rehabilitation patients related to risk of falling. The purpose of this study was to explore the nature of the relationship between the presence and severity of AHP after stroke and fall events in an inpatient stroke rehabilitation population.
The quality health outcomes model was used to understand how AHP fits within the larger context of falls and fall risk in the poststroke population (Ayanian & Markel, 2016). In 1966, Avedis Donabedian proposed using a triad of structure, process, and outcomes to evaluate the quality of health care (Ayanian & Markel, 2016). Guided by this foundational work, Mitchell et al. (1998) introduced the classical framework of structure, process, and outcomes as a dynamic model that recognizes the feedback and relationships that occur between variables in the evaluation of health outcomes of patients. In this model, medical and nursing interventions affect and are also affected by system influences and patient characteristics, which lead to individual health outcomes. Fall events, which can be considered a health outcome and the dependent variable in this study, are a multifactorial problem that involves a number of systems, interventions, and patient factors. Understanding when AHP occurs and how its presence is related to falls may better help clinicians predict patients who may fall, as well as implement patient-specific fall prevention interventions.
Methods and Design
Because this study aimed to investigate the nature of the relationship between two variables of interest, a prospective correlational design was utilized. Logistic regression was used between a priori variables to determine the predictive nature of AHP and falls in the stroke rehabilitation population.
Setting
The setting for this study was an inpatient stroke and general rehabilitation unit affiliated with a major medical university in the Southern United States. The stroke and general rehabilitation unit receives most of its patients from the affiliated major medical university, which is designated as a certified stroke center by the American Heart Association, American Stroke Association, and the Joint Commission. Admission to the inpatient rehabilitation unit is determined by an admissions team led by a physician.
Inclusion Criteria
Inclusion criteria included any patient admitted to the stroke rehabilitation unit with the primary diagnosis of ischemic stroke. Patients excluded from the study included those diagnosed with hemorrhagic stroke, preexisting diagnosis of a movement disorder, dementia, and previous strokes, as all these conditions predispose one to falls independent of the presence of AHP or stroke. A consecutive sampling method was used, and all individuals meeting inclusion criteria were asked to participate. The institutional review board approved a partial waiver of authorization for recruitment so that medical records could be reviewed and subject appropriateness established before the patient was approached for consent by the primary investigator.
Cognitive changes after a stroke present many potential ethical issues for those interested in poststroke research (Cherney, 2006). As part of the routine admission to the unit, each patient undergoes an intake cognitive assessment conducted by a neuropsychologist. The result of the cognitive assessment is documented in the patient's electronic medical record. Prior to the informed consent process, the patient's medical record was examined, and if there was a cognitive deficit noted by the neuropsychologist on admission to the unit, then consent per the legally authorized representative was initiated.
Instrumentation and Variables of Interest
Key variables of interest included the patient's age, gender, ethnicity, side of stroke, number of falls while admitted to the stroke and general rehabilitation unit, and the presence and severity of AHP as measured by the Visual-Analogue Test for Anosognosia for motor impairment (VATA-m; Della Sala et al., 2009). The purpose of the VATA-m, created specifically for the poststroke patient, is to diagnose the presence of AHP and allow for the comparison of the patient's score with a clinician score. This instrument is also suitable for patients with language deficits (Della Sala et al., 2009) and was created specifically for the poststroke population. Because the VATA-m uses clinician scores, demographic information from the clinicians who participated in the study was collected as well.
Another instrument used was the Morse Fall Risk Scale. The Morse Fall Risk Scale is one of three traditional scales that are more frequently used in the stroke rehabilitation setting, though this scale has not been validated for use in this population (Campbell & Mathews, 2010). All patients admitted to the general and stroke rehabilitation unit are assessed every 12 hours using the Morse Fall Risk Scale. The scale addresses issues thought to be associated with falls, including a history of falling, secondary diagnoses, the use of ambulatory aids, issues with gait and transferring, and a final question concerning mental status. Risk is assigned based on the score the patient receives, and there are tiered fall prevention interventions that are associated with each score. For example, a moderate fall risk is a score of 24-44, and the interventions tiered to that range include communication between clinicians via safety huddles, purposeful hourly rounding, and keeping walkways clear of medical equipment (Morse et al., 1989).
In the final question of the Morse Fall Risk Scale, the clinician is asked to address if the patient is oriented to their ability, or if they overestimate and forget their limitations. This question, though it cannot distinguish between AHP and a cognitive impairment, was intended to be used by the principal investigator as a preconsent screening tool to identify which patients to approach for inclusion in the study. If a clinician noted that their patient was not aware or overestimated their abilities, then that patient would be approached by the investigator for possible consent and inclusion into the study. However, during the triweekly interviews, it was noted that every eligible stroke patient was deemed aware of their limitations by the clinician. Therefore, every eligible patient was approached for possible inclusion in the study.
Datal Collection Procedure
To administer the VATA-m, two paper copies of the instrument were used (Della Sala et al., 2009). The 12 experimental questions each has a picture of a motor task, a description of a motor task, and a visual analog scale that ranges from 0 to 3. Zero on the scale indicates there is no difficulty in completing the task, whereas a three on the scale indicates major difficulties in completing the task. For this study, the participant was shown the picture, and the principal investigator read the description of the motor task. The participant was then asked to verbally indicate their ability to complete the task on a scale from 0 to 3. If the patient was aphasic, they were asked to point to a number on the visual analog scale that corresponded with their ability to perform the task. To obtain normative and comparative data, the patient's clinician (nurse or patient care technician) was also asked to complete the VATA-m and rate how well they believed the patient could perform the task listed. This assessment occurred during the same data collection period as the patient but took place in another room. Each instrument was scored independently, with a maximum score of 36 and a minimum of 0. From here, the patient's self-evaluation score was subtracted from that of the clinician, and the discrepancy score or difference was used to determine the presence and degree of AHP. To use the VATA-m, written correspondence with the original authors, Drs. Sergio Della Sala and Gianna Cocchini, was initiated, and permission to use the instrument was received.
To monitor falls on a weekly basis, the principal investigator accessed an intranet organization-wide system used to monitor patient safety events. It is the institution's policy that all safety events, including falls, are entered into a database maintained by the Office of Risk Management and Insurance. The nurse at the bedside, who witnessed or discovered the fall, enters the appropriate information into the database. This information includes the patient's name and medical record, additional demographic information, injuries sustained during the fall, and a narrative report concerning the circumstances surrounding the fall. A physician note is also entered into the electronic health record.
Data Collection
Data collection began January 1, 2020, with the original goal to have a total of 34 stroke rehabilitation patients enrolled in the study per a power analysis using G*Power. However, in mid-March 2020, COVID-19 became a national concern, and data collection was stopped. Preliminary data analysis revealed that 15 of the 16 enrolled patients had AHP. Because of these results, the decision was made to move forward with the completion of the study, noting that the sample size would be a limitation to the statistical results.
Data Analysis
Data analysis for this study was performed using IBM SPSS Statistics Version 26 software. Analysis of the variables began with examining descriptive statistics, including mean, median, mode, and standard deviation for continuous variables (AHP score and number of falls). Frequency and proportions were calculated and reported for the categorical variables of gender, ethnicity, side of stroke, AHP (as a category), falls, and fall injury. To explore the nature of the relationship between the presence and severity of AHP and fall incidents, Fisher's exact test was used. Logistic regression was then computed to determine what percentage of variance in patient falls (outcome variable) was explained by the presence of AHP (predictor variable).
Results
A total of 101 stroke and general rehabilitation patients were screened for eligibility. Thirty-eight patients (38%) of those screened had a diagnosis of stroke, which prompted a more detailed medical record review. Of the 38 patients with a stroke diagnosis, 18 (47%) had experienced a hemorrhagic stroke and were excluded from participation. Other exclusions included one patient who refused, one with a preexisting movement disorder, one with a previous stroke, and one patient with diagnosed dementia. This left 16 patient-and-clinician dyads who were enrolled and participated in the data collection process (see Table 1).
Sample Description
The average age of the participants enrolled in the study was 57 years (SD = 15.1), with 10 identifying as male and six identifying as female. More than half of the sample was Caucasian (nine participants), five were African American, one was Latinx, and the remaining participant declined to answer.
The majority of the participants (n = 13) experienced a right-sided stroke, and the remaining three participants were diagnosed with a left-sided stroke. Major anatomical locations of the ischemic events included the middle cerebral artery (eight participants), the corona radiata (three participants), and the basal ganglia (two participants). The remaining three participants experienced a stroke in their medulla, posterior inferior cerebellar artery, and thalamus.
Presence and Severity of AHP
Of the 16 participants enrolled in this study, 15 showed evidence of AHP per the VATA-m. The prevalence of AHP in this sample is considerably higher than what is noted in the literature. It is estimated that between 10% (Baier & Karnath, 2005) and 77% (Vidovic et al., 2019) of stroke survivors show evidence of AHP. Although the prevalence was higher in the current study (see Table 2), it must be noted that the VATA-m is still only a subjective measurement of awareness of ability (Della Sala et al., 2009).
Per the definitions used by the original instrument authors, the severity of AHP ranged from none to severe. Seven participants (44%) had scores that indicated mild AHP. Another seven participants (44%) had scores that indicated the presence of moderate AHP. The final participants had a score that indicated severe AHP.
Presence and Severity of Fall Events
Of the 16 participants enrolled, four experienced a fall during the data collection period. One fall involved the participant with severe anosognosia. The other three falls occurred with individuals categorized as having mild AHP. All four falls occurred in the participant's private room. One fall was witnessed, whereas the other three were unwitnessed. One participant slipped and fell while in the shower alone and pulled the emergency cord to notify the unit staff that help was needed. Two participants were found by unit staff, and the circumstances surrounding the event were unclear as neither individual was able to recall the circumstances of the fall. The one witnessed fall occurred when a family member was attempting to help a patient transfer from the wheelchair to the bed.
The severity of falls was based on preexisting injury categories defined by the National Database of Nursing Quality Indicators (NDNQI), the largest repository of nursing quality metrics in the United States (Montalvo, 2007). According to the NDNQI, there are five categories of fall injury. A "minor" fall indicates that there is injury requiring the application of a dressing, ice, cleaning of a wound, or the presence of an abrasion. The additional categories speak to more serious injuries, including suturing, surgery, or patient death in relation to injuries sustained from the fall. Three of the four falls that occurred were classified as "none" or having no evidence of physical injury from the incident. The additional fall resulted in a "minor" injury, described as a bruise above the participant's eye. None of the falls that occurred during the data collection period required any medical intervention.
Relationships Between AHP and Falls
To understand the relationship between the presence and severity of AHP and the presence and severity of fall events, Fisher's exact test was used. Primary outcome results indicated a nonsignificant relationship between the presence of AHP and patients who fell during the study (p = .99, two-tailed Fisher's exact test). The same statistical technique was used to understand the relationship between the severity of AHP and the severity of falls, and the outcome again indicated a nonsignificant relationship (p = .138, two-tailed Fisher's exact test; see Table 3).
Logistic regression was used to determine if the presence or severity of AHP was predictive of falls in this sample. In the initial model, the predictor was the presence of AHP, and the dependent variable was the incidence of a fall during admission. According to the model, AHP was not predictive of fall events, F(1, 14) = 0.318, p = .582, with an R2 of .22. An additional regression model was calculated to predict the effect of severity of AHP on the incidence of falls, and this model was also not predictive of fall events, F(1, 14) = 0, p = 1, with an R2 of -.071. In this sample, neither the presence nor severity of AHP was predictive of fall events.
Discussion and Clinical Significance
This study aimed to characterize the presence and severity of AHP and its relationship to patient falls in the stroke rehabilitation population. The results of this analysis concluded that neither the presence nor the severity of AHP contributed to the variability of falls in this sample. This does not mean that a relationship does not exist, as the sample size for this project was small because of COVID-19 restrictions. A similar study aimed at understanding the impact of AHP on global rehabilitation outcomes, with a much large sample size, may yield different results.
The proportion of participants who demonstrated the presence of AHP in this sample is considerably different from the results of other studies in which the estimates of AHP range from 10% (Baier & Karnath, 2005) to 77% (Vidovic et al., 2019). Fifteen of the 16 patients enrolled had a discrepancy score suggestive of AHP. Variability in rates can be attributed to many variables, but one of note is the use of the VATA-m in this population (Della Sala et al., 2009). The VATA-m has the unique feature of including a visual scale that allows individuals with expressive aphasia to participate when, in other studies, those with expressive aphasia are excluded (Hartman-Maeir et al., 2001). As a result, three left-sided stroke patients with diagnosed aphasia were able to communicate and participate in the assessment. This is relevant in that previously published studies focused on right-sided lesions, potentially excluding the aphasic population, which may have minimized the impact of AHP on stroke survivors (Antoniello & Gottesman, 2020; Barrett, 2021; D'Imperio et al., 2017; Fowler et al., 2018; Kirsch et al., 2021).
Falls that occurred during the data collection period did not differ from other reported fall rates in the same population. Although the sample size was limited, the number of falls experienced by those enrolled was proportional to what other studies reported, about six falls per 1,000 patient days (Quigley, 2016). Only one patient sustained a minor injury based on preexisting NDNQI classifications of the participants who fell. This finding is also congruent with other literature that suggests that though the incidence of falls in the stroke rehabilitation population is high, the risk for serious injury is low (Teasell et al., 2002). However, there could be emotional or psychological consequences of the fall that the clinician cannot discern or perceive with the traditional postfall assessment. For example, some patients who fall may be afraid of falling again, leading to chronic immobility and a prolonged rehabilitation period (Natan et al., 2016).
In the rehabilitation setting, most falls occur in the patient room, are unwitnessed, and happen at night or during the evening shift (Quigley, 2016; Wong et al., 2011). The falls that occurred in this sample were no different; all the falls occurred in the patient's private hospital room, three of the four falls were unwitnessed, and three of the four occurred at night or during the evening shift when patients were presumed to be in bed or asleep. Though no study has found correlations between variables to be significant enough to suggest a connection (Quigley, 2016; Wong et al., 2011), all four of the patients who fell had AHP per the VATA-m (Della Sala et al., 2009). Although there was not a statistically significant relationship, clinically significant patterns should be considered. Maintaining adequate clinical staffing in the overnight hours, specifically on floors and units that care for patients with AHP, should be a priority for nursing leadership to positively impact patient outcomes.
Though the purpose of this study was to examine the relationship between the presence and severity of AHP after stroke and fall events in the stroke rehabilitation setting, it must be noted that falls in any setting are the consequence of multiple variables involving the system, medical interventions, and patient-specific characteristics, which are most easily understood referencing the quality health outcomes model (Mitchell et al., 1998). The most common system issues contributing to the outcome of patient falls include inadequate assessment, lack of leadership, failures of communication, deficiencies and safety hazards, and inadequate staffing levels (Besharati et al., 2015; Chen et al., 2015). Understanding that the phenomenon of AHP exists and could be present in stroke rehabilitation patients adds to the patient-specific knowledge of treating these patients.
Another clinically significant finding that was not anticipated is clinician unawareness of patients' AHP. Morse Fall Risk scores were monitored and collected for this sample of patients. The Morse Fall Risk Scale question that was of most interest asked the clinician to answer if they thought that the patient was aware of their disability (Morse et al., 1989). In all 16 patients in this sample, 100% of the clinicians documented that the patient was aware of their disability. These data are not consistent with the finding that 15 of the 16 patients had AHP and were unaware of their limitations. If the patients in this sample were aware of their limitations, as scored by the clinician, they would not have had a significant AHP score. This finding demonstrates an overestimation of awareness being made by clinicians at the bedside, where fall prevention interventions are put into place. There is no quick bedside method currently available to assess a patient's awareness of their physical limitations. The Morse Fall Risk Scale does not discriminate between awareness of AHP and awareness of a disability in general. The clinical staff may be erroneously using alertness, cognitive status, or verbal ability as a proxy for understanding one's disability versus assessing if the patient is aware that a disability is present. Thus, incorporating the VATA-m, or a component of the VATA-m, into standard patient assessment may provide important information to help improve patient care.
Study Strengths
This study has demonstrated that AHP exists in stroke rehabilitation patients and that previous estimations of the phenomenon may be low. Although the sample size was small, using a visual analog scale, patients with aphasia were included into the sample, instead of being excluded because they could not speak. The principal investigator was also responsible for screening patients, data collection, input, and analysis, contributing to the reliability of the study.
Study Limitations
There are acknowledged limitations of this study. The SARS-CoV-2 pandemic significantly limited the sample size and affected the data collection process. Based on a power analysis conducted before the study began, a sample size of 34 patient-and-clinician dyads was needed to achieve a large effect size between the variables. Unfortunately, data collection was in progress when the virus emerged in the United States, and because of the unknown nature of the pathogen, only essential workers were allowed in the hospital for many months. It was not clear during this time when nonessential healthcare workers, including scientists, would be allowed back in the patient care areas. Because of this, the decision was made to stop data collection and move forward with analyzing the data collected. Though statistical techniques were used, no conclusive determinations can be ascertained because of the sample size, and further study is needed to understand the impact of AHP on stroke survivors. Likewise, data collected at one site limits generalizability or external validity of the results to other settings.
Conclusion
Findings from this study support a more extensive exploration of the effects of AHP on patients' stroke rehabilitation including missed rehabilitation time, falls, discharge disposition, return to work, return to driving, and quality of life after stroke. This study provides evidence that the phenomenon exists; however, having a larger sample and expanding outcome variables will provide further insight into the clinical impact of AHP. The incidental finding of the clinician overestimating patient awareness raises an essential question of patient assessment at the bedside. What is not clear from the data collected is why this diffence exists. It appears that clinical staff use proxy variables to estimate patient awareness of disability, and an opportunity for future work is bridging the gap between patient awareness of disability and clinician understanding of disability. A qualitative study may be helpful in understanding this phenomenon from both a patient view and a clinician view. Understanding that AHP can exist as a disorder of motor function rather than cognition adds a layer of complexity to understanding levels of consciousness and supports the idea of an individual assessment of awareness beyond assumption based on alertness or cognition. An objective assessment technique that can be quickly administered at the bedside, like the VATA-m, should be considered for use. Ongoing work and research in these areas should continue to be a priority in the stroke rehabilitation population to understand the true impact of AHP on patient outcomes.
Key Practice Points
* AHP is a unique phenomenon that causes a disconnect in motor pathways of patients with stroke.
* AHP may be a reason that individuals in stroke rehabilitation continue to fall at higher rates than other nursing units.
* Clinicians may overestimate their patient's awareness of disability.
* Clinicians who work at the bedside with stroke rehabilitation patients should incorporate a portion or all of the VATA-m to assess awareness.
Conflict of Interest
The authors declare no conflict of interest.
Funding
This work was partially supported by the Department of Veterans Affairs, Veterans Health Administration, Office of Academic Affiliations VA Quality Scholars Advanced Fellowship Program. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the U.S. government.
Acknowledgments
None.
REFERENCES