Abstract
Age at first stroke is decreasing, and most strokes are mild to moderate in severity. Executive function (EF) deficits are increasingly recognized in the stroke population, but occupational therapists have not altered their evaluation methods to fully accommodate changing patient needs. We present a hierarchical performance-based testing (PBT) pathway using data to illustrate how PBT could identify patients with mild stroke-related EF deficits in need of occupational therapy intervention. Data suggest that a substantial number of patients with EF deficits after mild stroke could benefit from occupational therapy services.
Improved stroke prevention efforts and increased use of tissue plasminogen activator have resulted in decreased stroke-related mortality and disability such that acute stroke impairment is most often mild (49.3%) or moderate (32.8%) in severity (Saver et al., 2013). The age at stroke onset is also declining, with approximately one-third of strokes occurring in people younger than age 65 (34%; Hall, Levant, & DeFrances, 2012). Mild stroke is defined by a National Institutes of Health Stroke Scale (NIHSS) score of <5 (Crespi et al., 2013; Khatri, Conaway, & Johnston, 2012; Poulin, Korner-Bitensky, & Dawson, 2013). Most patients scoring in the mild range on the NIHSS will have no activity of daily living (ADL) deficits (Edwards, Hahn, Baum, & Dromerick, 2006) and, especially when of younger age, will be expected to resume premorbid participation levels of work and community engagement. However, mild stroke-related neurological impairment is increasingly recognized as being consistent with compromised occupational performance and participation as a result of cognitive or executive function (EF) impairment (Morrison et al., 2013; Wolf, Baum, & Conner, 2009).
Wolf et al. (2009) advocated a change in the focus of occupational therapy for people with mild stroke, suggesting that the clinical practices of occupational therapists match the participation needs of the population needing services (Morrison et al., 2013; Wolf et al., 2009). This change suggests that clinicians assess the executive and cognitive abilities of people with mild stroke. With these recommendations, assessment of cognitive and EF ability moves beyond routinized ADL performance toward performance of complex instrumental activities of daily living (IADLs). Ultimately, the assessment of cognitive and EF abilities provides clinicians with important information that can be addressed through interventions designed to support people with mild stroke in returning to their premorbid levels of participation.
In line with this proposed shift, Wolf et al. (2009) recommended that a top-down assessment of participation be used to guide therapy interventions for people with acute stroke. Although client centered, participation measures rely on self-report methods to identify clients’ premorbid roles, activities, and routines. In the acute care setting, measures of participation are likely to be truly valid only when the patients are far enough poststroke to have had opportunities to experience stroke-related changes in their daily functioning. Moreover, clients with mild stroke rarely stay in acute care settings long enough for experiential learning to take place. Indeed, participation measures may have limited utility on inpatient floors, because patients with stroke have not yet returned to community living. Additionally, after acute stroke many people may experience limited insight into their cognitive and EF deficits. Reliance on self-report measures of participation in acute settings may, in fact, contribute to more disparity in access to rehabilitation services for patients with mild stroke.
It is well documented that people living with mild stroke-related cognitive and executive deficits often go undetected by health care providers. Edwards, Hahn, Baum, Perlmutter, et al. (2006) reported that among a sample of people with mild to moderate stroke, 71% were discharged with minimal or no services. Similarly, therapeutic interventions with a clear focus on work or community participation are rarely provided to people with mild to moderate stroke (Wolf et al., 2009).
Even as the population of people with stroke has changed so, too, has the practice of occupational therapy. During acute hospitalization for stroke, opportunities for naturalistic observation of ADLs or IADLs are limited. Occupational therapists may attempt to use ADL assessments as a screening tool, but ADL assessment alone is unlikely to be adequate in assessing people whose difficulties are limited to EF-based IADL impairments. Within the context of mild stroke, the role of impairments in EF is recognized as a major contributing factor to problems in community living (Morrison et al., 2013), but problems in EF are difficult to identify with tabletop cognitive or ADL evaluations (Edwards, Hahn, Baum, & Dromerick, 2006; Manchester, Priestley, & Jackson, 2004; Petruccelli & Delenick, 2013). Traditional evaluations of ADL status such as the FIM™ (Uniform Data System for Medical Rehabilitation, 1997) or Barthel Index (Mahoney & Barthel, 1965) are insensitive to EF deficits, nor do they predict ability to perform complex tasks (Petruccelli & Delenick, 2013).
Poulin et al. (2013) stressed the importance of performance-based testing (PBT) for EF deficits after stroke in their recent comprehensive review of stroke-specific EF assessments. Their review demonstrated that occupational therapists have the tools to identify significant changes in the functional capacity and participation of patients with mild stroke that are commonly missed by routine cognitive or ADL screening methods.
In this article, we advocate an approach to occupational performance assessment that supports an expanded role for occupational therapy to better meet the needs of the population with mild stroke. The challenge to the occupational therapist practicing in the early poststroke period is to find ways to systematically, efficiently, and cost effectively identify people with mild stroke who are in need of further assessment and occupational therapy treatment.
The American Occupational Therapy Association (AOTA; 2013) statement on cognition asserts that “cognitive functioning is always embedded in occupational performance and cannot be accurately understood in isolation” (p. S11). Consistent with this principle, we argue that PBT has significant, but as yet unrealized potential for poststroke occupational therapy practice (Petruccelli & Delenick, 2013; Poulin et al., 2013). Use of PBT allows for more accurate estimation of real-world function without reliance on (1) the patient’s awareness of deficits, (2) the patient’s experiential knowledge, and (3) performance on specific cognitive tests, which may be poorly predictive of actual functional competence (Manchester et al., 2004). Identifying and addressing performance problems support better long-term outcomes, particularly for the growing population of younger people with stroke (Edwards, Hahn, Baum, & Dromerick, 2006).
What Is Performance-Based Testing?
In PBT, a test is used to sample actual behavior, and the results are compared with a standard to assess competency. In psychology and neuropsychology, any standardized measure involving paper-and-pencil or tabletop activity may be considered a PBT (Wechsler, 1997). In contemporary occupational therapy practice, PBT is closely linked to the concept of ecological validity and performance in a real-world context (AOTA, 2014). Performance-based tests are often related to a specific area of functioning (i.e., are domain specific) and are intended to predict a particular area of competency (e.g., toileting ability would predict other aspects of ADL ability). Occupational therapists have developed a range of domain-specific performance-based tests that include the Test of Grocery Shopping Skills (Hamera, Rempfer, & Brown, 2005), handwriting measures (Duff & Goyen, 2010), and the Kettle Test (Hartman-Maeir, Harel, & Katz, 2009). ADL tests such as the Barthel Index (Mahoney & Barthel, 1965), and the FIM™ (Uniform Data System for Medical Rehabilitation, 1997) may also be considered domain-specific performance-based tests when used as assessments rather than rating scales.
The first notable attempt to use a standardized performance-based test to predict general as opposed to domain-specific functional abilities was the Allen Cognitive Level Screen (ACLS; Allen, 1985). The ACLS is an evaluation of functional cognition, including problem solving (a component of EF) and the ability to learn functional skills (Riska-Williams et al., 2007; Secrest, Wood, & Tapp, 2000). The ACLS, however, has not been demonstrated to be sensitive to cognitive and EF deficits after stroke. A range of newer performance-based tests more closely approximating the complex demands of community life (i.e., EF) have been developed or adapted by occupational therapists and include the Executive Function Performance Test (EFPT; Baum, Morrison, Hahn, & Edwards, 2003), the Assessment of Motor and Process Skills (Fisher & Bray Jones, 2010), and the Multiple Errands Test (MET; Morrison et al., 2013; Shallice & Burgess, 1991). These measures focus on ecologically valid assessment of EF and were developed to predict higher order thinking abilities that support participation in the home and community. Performance-based tests can therefore be categorized as (1) domain or function specific, (2) general with an EF component, and (3) EF specific (Poulin et al., 2013). Poulin et al. (2013) provided a review of stroke-specific EF measures.
Problems With and Opportunities for Performance-Based Testing in Occupational Therapy Practice
Hospital environments are unlike the world that the patient will occupy postdischarge. Hospital environments are purposefully kept simple and inadvertently result in patients being the passive recipients of care. Staff at hospitals commonly provide simple one-step cues, and activities are often serial (patients do not need to multitask). Because hospital routines are developed with the goal of reducing task demands, most spontaneously available and naturally observable behaviors are likely to be poor predictors of real-world functional abilities.
Patients with mild stroke are typically hospitalized for 1–3 days and are not usually provided with any rehabilitative services. Acute and subacute occupational therapy evaluations typically take place in a patient’s hospital room. Evaluations focus on providing ADL care and are scheduled into the daily care plan organized by nursing staff. Patients are typically assessed for the level of support they need to perform ADL tasks. It is not common practice to ask patients to multitask or perform dynamic and novel tasks with graded levels of distractions to increase task challenge.
Evaluations of ADL functioning may inadvertently provide excessive supports and reduced task demands during personal care routines. As a result, occupational therapists may fail to identify mildly to moderately impaired patients who have EF deficits, who could benefit from intervention, and who are at risk for hospital readmission or reduced levels of community participation after discharge from the hospital. Edwards, Hahn, Baum, and Dromerick (2006) found that therapists identified significantly more deficits among patients recovering from stroke when they used a comprehensive screening battery rather than the therapists’ routine ADL clinical evaluation method. They suggested that patients with stroke who appear neurologically healthy may nonetheless have cognitive and perceptual deficits that compromise their ability to resume roles and participate in their work, community, and social life (Edwards, Hahn, Baum, Perlmutter, et al., 2006; Gresham, Duncan, & Stason, 1995).
Assessment Path for People With Stroke That Includes Executive Function Assessment
A performance-based assessment protocol for people with mild to moderate stroke might include, in hierarchical order, an ADL measure (e.g., the FIM), an IADL measure (e.g., the EFPT), and a measure of higher order EF (e.g., the MET). Patients who have ADL deficits will most likely be referred for rehabilitative services, during the course of which they are likely to be assessed for cognitive function and EF. ADL measures such as the FIM, however, may not be sensitive to subtle EF impairments; therefore, further assessment is needed to predict the potential for independent community skills.
If a patient is free of basic ADL deficits, the EFPT could be used to assess for potential IADL impairment. Bedside-administration EFPT subtests, in particular the Medication Management Task and the Bill Pay Task, are practical for acute and subacute settings because they are sensitive to mild deficits, are portable, and can be administered within 10 min. The MET is not an off-the-shelf measure and has to be adapted for use in each location in which it is to be administered, a consideration that requires a commitment to its use. The MET, with an administration time of 30 min or more, may not be considered practical for use in acute care settings. The American Heart Association (AHA) Stroke Guidelines (Go et al., 2013), however, recommended a follow-up visit with a neurologist specializing in stroke 4–6 wk postdischarge. If at that time functional challenges are noted, we would suggest referral to an occupational therapist for further PBT using the medication management and bill pay items on the EFPT along with the MET and Stroke Impact Scale (SIS; Lai, Studenski, Duncan, & Perera, 2002) to identify whether referral to outpatient occupational therapy is indicated.
Preliminary Test of a Proposed Assessment Path and Treatment Guide
As an illustration of the potential effectiveness of this proposed assessment path, we analyzed a prospectively collected dataset of 420 patients 6 mo after mild stroke to identify the percentage of patients who reported participation limitations and whose cognitive or EF deficits were not identified by conventional methods. Participation deficits were operationalized as scores of ≤80 on the SIS (a measure of participation after cerebrovascular accident that is presumed to indicate normal participation at its maximum [100] score).
Of the entire sample with mild stroke, 94% were independent on the FIM and 58% reported no or minimal limitations in participation. Of the patients who described limitations in participation, 87% scored as modified independent or independent (scores > 108, hereinafter termed independent) on the FIM. The FIM captured only 13% of people reporting decreased participation on the SIS, indicating that it is not an appropriate screening tool for the identification of problems leading to reduced participation among the population with mild stroke. Among participants scoring as independent on the FIM, 39% reported scores of ≤80 on the SIS (i.e., that they had impaired community participation). Of people identified as independent on the FIM and who reported decreased participation on the SIS, 46% scored as impaired (≥3) on the EFPT.
Analysis of a separate dataset found that 22% of people identified as unimpaired on the EFPT (<2) also reported decreased participation on the SIS and scored as impaired on the MET. When taken together, data from these two datasets suggest that, using the EFPT and the MET, a hit rate of ≤68% may be possible in identification of EF deficits that may be related to patient-identified problems in participation after mild stroke and that are unidentified with routine administration of the FIM. These data demonstrate that among people living in the community after mild stroke and who report problems in participation, cognitive and EF deficits can be identified using PBT.
Evaluation methods implemented in a logical sequence can limit the time required of therapists and the stress on patients and yield more data at each step, with the FIM being the least sensitive; the EFPT, of intermediate sensitivity; and the MET, most sensitive. Clearly, this post hoc data analysis provides only preliminary evidence and does not control for other factors that can lead to reduced SIS participation ratings beyond the cognitive or EF deficits identified by the EFPT or the MET (e.g., depression, perceptual impairments, environmental factors). Despite these limitations, these data illustrate the potential magnitude of the unmet need for occupational therapy services and suggest that a large group of people with mild stroke with EF deficits currently go unrecognized and, thus, untreated. On the basis of our analysis and extrapolating from the incidence of mild stroke reported by the AHA (Saver et al., 2013), we estimate that approximately 84,000 people with mild stroke would be identified and potentially provided with occupational therapy service each year if this hierarchical assessment path was implemented by occupational therapists across the United States.
Discussion
Previous studies have noted the reluctance of clinicians to use standardized testing protocols (Baum, Perlmutter, & Edwards, 2001). The literature suggests that many therapists believe that standardized assessments are warranted only if a patient shows clear signs of impairment (Dunn, 2001; Edwards, Hahn, Baum, Perlmutter, et al., 2006). Historically, longer lengths of stay allowed acute care clinicians to spend more time with patients and families, allowing for naturalistic observation and increased detection of impairments during acute hospitalization. Clinicians with limited time and who are under pressure to evaluate a high volume of patients to provide discharge recommendations may be reluctant to engage in routine PBT in the absence of obvious impairments, but EF deficits are unlikely to be obvious during routinized task performances in controlled hospital environments.
Pressure to discharge patients quickly is highest for ambulatory patients without gross motor, cognitive, or language impairment (Wolf et al., 2009). The majority of people with mild stroke are discharged directly to home from the acute setting without the benefit of inpatient rehabilitation. We suggest that people with mild stroke commonly experience chronic cognitive and EF impairments and decreased occupational performance and participation. This observation is consistent with the report of Khatri et al. (2012), who found that 29% of the patients with mild stroke had Modified Rankin Scale scores of 2–6 at 90 days, which they describe as a poor outcome. The Modified Rankin Scale is a global measure of outcome, so although the data reported here are consistent with Khatri et al.’s results, we include more refined measures of functional performance and indicate the importance of EF as an additional cause of disability. Most third-party payers rely on frontline clinicians to determine access to rehabilitation, and failure to document impairments may permanently deprive patients of the opportunity for therapy they might otherwise receive (Wolf et al., 2009).
Real life is complex, and performance-based tests provide ecologically valid task challenges. Additionally, real-world PBT tasks inherently make sense to patients, thereby increasing their motivation to participate in the clinical evaluation. The nature and extent of the evaluation are driven by the expected occupational performance demands on patients and patients’ expectations of themselves.
Much remains to be done in establishing the appropriate application of PBT. For example, the degree of challenge presented through current standardized PBT varies significantly, and therapeutic knowledge is required to adequately match a patient’s abilities to the appropriate PBT. Occupational therapists need to establish a true hierarchy of PBT and the appropriate decision and entry points for administering each measure. The factors that contribute to the ability of PBT to accurately assess EF are not yet well delineated. In this article, we have focused on the acute and subacute occupational therapy evaluation of patients with mild stroke, but there are known EF performance-based deficits associated with many medical conditions, including mood disorders, multiple sclerosis, cardiac health disorders, cancer, and postchemotherapy states, that might also benefit from this approach.