Understanding the Continuity Assessment Record and Evaluation (CARE) Tool

In the landscape of healthcare, particularly within acute and post-acute care settings, the ability to accurately and consistently measure patient functional status is paramount. This measurement is not merely about tracking progress; it’s fundamental for quality improvement, payment reform, and ensuring patients receive the most appropriate and effective care across different healthcare providers. Recognizing this critical need for standardized data, the Centers for Medicare & Medicaid Services (CMS) and the Assistant Secretary for Planning and Evaluation/Health Policy (ASPE/HP) have explored tools like the Continuity Assessment Record and Evaluation (CARE) item set, often discussed in the context of functional independence measurement (FIM), to establish robust quality metrics. This article delves into the purpose, analysis, and implications of using the CARE tool, drawing upon exploratory research aimed at providing recommendations for functional status quality metrics across various post-acute care (PAC) providers.

The drive to implement cross-cutting quality metrics stems from the evolving understanding of patient care pathways. Traditionally, Medicare’s setting-specific approach fostered silos, limiting collaboration and accountability among different types of care providers, including inpatient rehabilitation facilities (IRFs), long-term care hospitals (LTCHs), skilled nursing facilities (SNFs), and home health agencies (HHAs). Each of these provider types historically utilized unique patient assessment tools, such as the Minimum Data Set (MDS 3.0) for SNFs, the Inpatient Rehabilitation Facility – Patient Assessment Instrument (IRF-PAI) for IRFs, and the Outcome Assessment and Information Set (OASIS-C) for HHAs. While these instruments collected similar clinical and functional status data, inconsistencies in item definitions, measurement scales, and data collection procedures hindered direct data comparability and cross-provider quality assessments.

The CARE item set emerged as a potential solution to bridge this gap. Collected during the CMS PAC Payment Reform Demonstration (PAC-PRD), the CARE data set offers standardized patient assessment information, encompassing admission and discharge functional status data alongside crucial clinical data that influences functional outcomes. This uniformity allows for a more cohesive and comparable evaluation of patient progress and care quality across the diverse spectrum of PAC providers. The research project undertaken by RTI International leveraged this rich CARE data to explore the development of risk-adjusted, facility-level functional status outcome measures applicable across these different provider types. By examining key analytic issues and incorporating expert clinical perspectives, this work aimed to inform the creation of aligned, crosscutting quality metrics that could revolutionize how functional status is measured and utilized in post-acute care. This article will explore the methodologies, findings, and recommendations derived from this pivotal research, offering valuable insights for healthcare professionals, policymakers, and anyone interested in advancing the quality and efficiency of patient care in the PAC sector.

Summary of Environmental Scan: Functional Status and Quality Metrics

To effectively develop crosscutting functional status metrics using the Continuity Assessment Record and Evaluation (CARE) tool, a thorough understanding of the existing landscape of functional status measurement and quality metrics is essential. This involves examining established definitions of functional status and reviewing current quality metrics in use, particularly those endorsed by national bodies like the National Quality Forum (NQF). The environmental scan provides a crucial foundation for identifying gaps, opportunities, and best practices in the field, ensuring that the development of new metrics is informed by the most current knowledge and standards.

Defining Functional Status

The World Health Organization’s International Classification of Functioning, Disability and Health (ICF) provides a comprehensive framework for understanding “function.” According to the ICF, “function” is an umbrella term encompassing all body structures and functions, activities, and participation in daily life. This broad definition highlights that functional status is not limited to physical abilities but includes cognitive and social aspects of a person’s life. Within the ICF framework, examples of functional status span various components:

Body Structures and Functions: This includes physiological and psychological functions of body systems, such as swallowing, bladder and bowel continence, and sensory functions.
Activities: This component refers to the execution of tasks or actions by an individual, such as eating, bathing, dressing, walking, and communicating.
Participation: This encompasses involvement in life situations, reflecting a person’s engagement in society, work, education, and recreational activities.

For the purpose of developing quality metrics using the CARE tool, the focus is primarily on functional status within the area of activities. This is because activities of daily living (ADLs) are directly observable, measurable, and clinically relevant in assessing a patient’s functional independence and progress throughout the care continuum. The National Committee on Vital and Health Statistics, Subcommittee on Health emphasized the growing importance of functional status information, stating that it is “increasingly essential for fostering healthy people and a healthy population.” Understanding how health conditions impact an individual’s ability to perform basic activities and participate in life is crucial for achieving optimal health and well-being.

It’s important to recognize that measuring functional status in the area of activities is not as straightforward as measuring vital signs like height or weight. There is no universally standardized unit of measure or scale for functional status. Instead, it is typically assessed using subsets of specific daily activities, such as eating, bathing, walking, and stair climbing. A person’s performance in completing these activities is then evaluated using a rating or response scale, which provides a multiple-point option set to rank their performance level.

When considering functional assessment across different provider types, the relevance and feasibility of collecting data on certain functional status activities can vary. For instance, assessing dressing in street clothes might not be applicable in some LTCH settings where patients primarily wear hospital gowns due to medical acuity and equipment needs. Similarly, activities like walking 150 feet or climbing stairs might be challenging or unsafe for some patients upon admission to SNFs or IRFs, but they could be set as discharge goals. Therefore, missing data for specific functional status activities can occur due to a patient’s clinical condition, making it unsafe or impractical to perform the activity.

Motor Functional Status (Self-Care and Mobility) Quality Metrics

Research in post-acute care rehabilitation has heavily relied on instruments like the Functional Independence Measure (FIM®) to measure functional status. The FIM instrument, integrated into the IRF-PAI, has been instrumental in understanding the dimensions of functional status. Early analyses of FIM data revealed that functional status is composed of two primary constructs: motor function and cognitive function. Motor function, often referred to as physical function, encompasses mobility (the ability to move around), self-care (the ability to perform personal activities), and bowel and bladder management. Cognitive function, on the other hand, includes communication activities (comprehension and expression), as well as social interaction, problem-solving, and memory.

More recent research, particularly using self-report functional assessment scales like the Activity Measures for Post Acute Care (AM-PAC), suggests a further refinement within motor function. These studies propose that motor activities should be separated into two distinct constructs: mobility and self-care, especially when analyzing diverse patient populations with varying diagnoses. This distinction arises because patients with different primary diagnoses (e.g., stroke, hip replacement, spinal cord injury) often exhibit different patterns of motor activity limitations. Self-care limitations are more likely associated with conditions affecting upper body function, while mobility limitations are more related to balance and lower extremity function. Therefore, using subscales of mobility and self-care may offer more precise and clinically meaningful measures of patient functional status outcomes compared to a combined motor measure.

While numerous functional assessment instruments exist, only a limited number have been translated into functional status quality metrics for performance measurement at the facility level. Facility-level aggregation of functional assessment data, such as IRF-PAI/FIM motor function data, has been utilized for internal quality reporting and accreditation requirements like the Joint Commission’s ORYX initiative. However, the Joint Commission has recently moved towards standardized, federally mandated performance measures, suspending reporting requirements for ORYX non-core measures like IRF-PAI/FIM metrics in LTCHs and IRFs.

Currently, no NQF-endorsed functional status metrics are routinely employed in IRFs, SNFs, or LTCHs, despite the collection of functional status data in IRFs and SNFs using IRF-PAI and MDS respectively. A review of NQF-endorsed performance measures reveals a limited number of functional status quality metrics, with several being process measures that involve functional status data collection. The existing NQF-endorsed metrics primarily utilize the OASIS data set for HHAs, FOTO functional assessment items for outpatient settings, and AM-PAC measures, along with specific metrics for COPD patients and functional decline in long-term care residents.

Memory, Problem Solving, and Communication Metrics

Cognitive function is a broad domain encompassing various subdomains, including communication, problem-solving, and memory. Recent research suggests that cognitive abilities may be further differentiated into communication as a distinct construct and problem-solving and memory as another. Developing quality metrics in the area of communication presents unique challenges. Communication disorders are significant health issues that can lead to social isolation, depression, and loss of independence. These disorders can arise from impairments in voice, speech, language, hearing, cognition, or a combination thereof. Recovery rates and expected outcomes for patients with communication problems are highly variable, depending on the underlying etiology. For example, communication difficulties may stem from motor speech disorders, aphasia (language disorder), or cognitive impairments.

Current assessment instruments often include items focusing on communication ability, but they typically do not capture the underlying cause of the communication problem. However, understanding the etiology of a communication limitation is crucial when measuring care outcomes with a quality metric, as it can significantly influence the expected recovery trajectory and serve as an important risk adjuster.

Cognitive and communication status has been challenging to measure concisely for clinicians. Brief cognitive status instruments, such as the Brief Interview for Mental Status (BIMS) included in MDS 3.0 and the PAC-PRD CARE Item Set, are utilized clinically and in research. The BIMS assesses short-term memory and recall, but comprehensive cognitive or communication measures have not been extensively used for provider-level analyses. The FIM cognitive items have been used as an ORYX measure for IRFs.

The American Speech-Language-Hearing Association (ASHA) developed functional communication measures (FCMs), with seven receiving time-limited endorsement from the NQF in 2008. These NQF-endorsed measures focus on memory, speech production, writing, spoken language expression and comprehension, attention, and reading. However, these measures are under review and may be withdrawn from NQF consideration.

Developing cognition functional status metrics faces several challenges, including the need for detailed data on patient functional status at admission and discharge, as well as the underlying reasons for any limitations. Knowing the etiology of cognitive or communication limitations is crucial because it can significantly impact the expected recovery trajectory and serve as a critical factor in risk adjustment.

Analytic Approach: Utilizing CARE Data for Quality Metric Development

The core of developing effective functional status quality metrics lies in a robust analytic approach. For this project, RTI International employed a comprehensive methodology centered around the Continuity Assessment Record and Evaluation (CARE) Item Set data, collected during the CMS PAC Payment Reform Demonstration (PAC-PRD). This approach encompassed data description, descriptive analyses, and Rasch analysis, all designed to address key analytic issues in quality metric development.

Description of Data: The CARE Item Set

The primary data source for this research was the CARE Item Set data, gathered as part of the PAC-PRD from April 2008 to December 2010. This demonstration involved 206 providers across geographically diverse markets, chosen to represent both rural and urban populations with varying densities of PAC providers. Participating providers included short-stay acute hospitals and PAC facilities (LTCHs, IRFs, HHAs, and SNFs). These providers submitted patient assessment data on Medicare beneficiaries, capturing patient severity at both admission and discharge, particularly in PAC settings. The PAC-PRD aimed to predict variations in resource intensity and patient outcomes across different PAC settings. This project expands on the initial examination of functional change within the PAC-PRD to develop facility-level functional status quality metrics.

The CARE tool was specifically designed to measure differences in patient complexity, resource needs, and outcomes across acute and PAC settings. The key data elements collected by the CARE item set include:

Administrative Information: Provider, beneficiary, and payer data.
Admission Information: Prior healthcare utilization and premorbid functional status.
Current Medical Items: Diagnoses, procedures, major treatments, skin integrity, and physiological factors related to the current admission.
Cognitive Status and Interview-Based Items: Orientation, risk of delirium, depression, and pain.
Functional Status: Self-care, mobility, and instrumental activities of daily living (IADL) at admission and discharge.
Other Factors Affecting Outcomes: Frailty and other relevant conditions.
Discharge Items: Discharge destination and caregiver information.

To ensure data reliability, rigorous interrater reliability testing was conducted within a subsample of PAC-PRD providers. This involved collecting paired assessments and utilizing video-based patient vignettes to examine clinician agreement across different provider types. The results from these reliability studies indicated that the CARE items are reliable and performed consistently across provider types.

The functional status section of the CARE item set is particularly relevant for quality metric development. It includes a detailed list of activities, categorized into self-care and mobility domains. These activities are:

Self-Care: Eating, oral hygiene, toileting hygiene, dressing upper body, dressing lower body, wash upper body, shower/bathe self, put on/take off footwear.
Mobility: Roll left to right, sit to lying, lying to sitting on side of bed, sit to stand, bed to chair/wheelchair transfer, toilet transfer, walking or wheeling (in room, 50 feet, 100 feet, 150 feet), pick up object, one step curb, walk 50 feet with 2 turns, four steps, twelve steps, walk 10 feet on uneven surfaces, car transfers, wheel short ramp (if using wheelchair), wheel long ramp (if using wheelchair).

For each activity, clinicians assign a numeric code from 6 (complete independence) to 1 (complete dependence), reflecting the level of assistance needed by the patient. Letter codes are used when an activity is not completed: “M” (medical issues), “S” (safety issues), “N” (not applicable), “P” (patient refused), and “A” (attempted but not completed).

In addition to CARE data, the analysis also utilized IRF-PAI/FIM data from 2008-2010, particularly when diagnostic specificity was required for certain analyses.

Data Analysis: Descriptive and Rasch Approaches

The data analysis strategy involved two primary approaches: descriptive analyses and Rasch analysis.

Descriptive Analyses: Initial analyses focused on understanding the basic characteristics of the CARE functional status data. This involved examining:

Distributions of Numeric and Letter Codes: Calculating the frequency and percentage of each numeric assistance level (1-6) and letter code (M, S, N, P, A) for each functional status item. These distributions were analyzed separately for each provider type (LTCH, IRF, SNF, HHA) and at both admission and discharge. This provided a detailed overview of patient functional abilities and limitations across different settings and time points.
Scatter Plots: Creating scatter plots to visualize the relationship between functional assessment change scores (discharge score minus admission score) and functional status discharge scores. This helped explore the correlation and patterns between improvement and final functional status levels.
Regression Analyses: Conducting regression analyses, particularly for case-mix adjustment work. This included testing for statistical interactions between various patient characteristics and functional outcomes to inform risk adjustment strategies.

Rasch Analysis: To gain a deeper understanding of the functional status items and their measurement properties, Rasch analysis was employed. Rasch analysis is particularly valuable when measuring latent traits—concepts that are not directly observed but are inferred from multiple indicators (in this case, functional activities). The one-parameter Rasch model was used to:

Examine Item Hierarchy: Determine the order of functional status items (activities) from easiest to hardest, reflecting the underlying construct of functional ability. This helps understand the relative difficulty of different activities and how they contribute to overall functional status.
Create Interval-Level Scales: Transform ordinal-level data (assistance levels 1-6) into interval-level data. This is a key advantage of Rasch analysis, as it ensures that the distances between points on the scale are equal, allowing for more accurate and meaningful comparisons and statistical analyses.
Address Missing Data: Rasch analysis can estimate person measures even with some missing data, making it robust for handling real-world clinical data where not all activities may be assessed for every patient.
Inform Item Selection: Identify key items that effectively span the entire range of patient functioning, ensuring that the functional status scales are sensitive to both high and low functioning patients. It also helps identify redundant or overlapping items.

Rasch analysis uses scored responses to create a functional status scale or “ruler.” This ruler represents a continuum of functional ability, with each item and each level of assistance placed along this continuum based on its difficulty or the level of ability it represents. By applying Rasch analysis, the research aimed to create more robust and psychometrically sound functional status measures for quality metric development. The Winsteps 3.75 software program was used to perform the Rasch rating scale analysis.

Results: Key Findings from CARE Data Analysis

The comprehensive analysis of the CARE data, employing both descriptive and Rasch methodologies, yielded significant results that informed recommendations for functional status quality metrics. These results address several key analytic issues critical to the development of effective and meaningful metrics.

Descriptive Analyses: Understanding Data Distributions

Initial descriptive analyses provided a foundational understanding of the CARE data. Examining the distribution of numeric scores, letter codes, and missing data for each functional status item, stratified by provider type and assessment time (admission and discharge), revealed several important patterns. These distributions, visualized in bar graphs and detailed in appendix tables, highlighted the varying levels of patient function and data completeness across different PAC settings.

Key observations from the descriptive analyses included:

LTCH Patient Acuity: Patients in LTCHs consistently showed the most frequent use of letter codes, indicating that functional status activities were often not completed due to medical or safety reasons. LTCH patients also exhibited a higher percentage of patients who were dependent in both self-care and mobility activities at admission. This aligns with the expected higher acuity and medical complexity of patients in LTCH settings.
HHA Patient Independence: HHA patients, as anticipated, generally showed the least frequent use of dependence codes, indicating a higher level of baseline functional independence compared to patients in other PAC settings. However, they still presented with functional limitations requiring home health services.
“Setup Assistance” Relevance: For self-care items, the level 5 code (setup or cleanup assistance) was frequently used at both admission and discharge. This level was less common for mobility activities. This suggests that setup assistance is a particularly relevant category for self-care activities, reflecting the need for preparatory or follow-up help with equipment or items, even when patients can perform the activity themselves.
Letter Codes for Challenging Items: Letter codes, indicating that an activity did not occur, were most frequently used at admission and for more challenging mobility items (e.g., steps, uneven surfaces, car transfers). This is understandable, as patients upon admission might not be medically stable or physically able to attempt these more demanding activities.
Missing Data for “Pick up Object”: Interestingly, the item “picking up object from floor” had a high rate of missing data or was not assessed at discharge for all provider types except SNFs. This suggests potential issues with the practicality or perceived relevance of this item at discharge assessment across different PAC settings.
Overall Score Distribution: The overall distribution of scores for the CARE function items was consistent with the expected complexity and patient populations in each PAC setting. This provides initial validation of the CARE tool’s ability to capture meaningful differences in functional status across the care continuum.

These descriptive analyses provided valuable context for interpreting subsequent, more complex analyses and informed the practical considerations for implementing functional status quality metrics across diverse PAC settings.

Use of Single Items vs. Multiple Items in Quality Metrics

One of the fundamental questions addressed in this research was whether functional status quality metrics should be based on single items or multiple items combined into a scale. This issue balances measurement precision with the practicality and feasibility of data collection in busy clinical environments.

Analysis and Results: To investigate this, Rasch analysis was applied to the CARE functional status data. The analysis focused on the relative relationships between items and rating scale levels to construct a stable functional status scale across time and settings. Letter codes were recoded to level 1 (dependent) for the Rasch analysis. The Rasch scale was set from 0-100, with higher values indicating greater independence.

Figure 4-1 from the original document visually represents the Rasch analysis of motor items.

Figure 4-1. Rasch Analysis of Motor Items (Self-Care and Mobility) in the Continuity and Assessment Record and Evaluation Tool

This figure illustrates the “motor ruler” created by the Rasch analysis, showing the placement of each item’s rating scale levels along the functional status continuum. The analysis revealed that the “distance,” or amount of ability, needed to move between rating scale steps was not uniform across the scale or across different items. For example, the range on the ruler for level 3 (partial/moderate assistance) for “eating” is smaller than the range for level 5 (setup assistance), indicating that the functional difference between levels is not consistent across the rating scale. This finding underscores the limitations of treating ordinal-level data as interval-level data, as is often done with summed raw scores.

Furthermore, the Rasch analysis highlighted potential floor and ceiling effects when using single items. For instance, relying solely on the “eating” item might not capture improvement in patients who are already independent in eating but improve in more challenging motor activities. A patient scoring “independent” (level 6) on eating would be estimated to have a measure of ~47 or higher on the motor ruler, limiting the ability to detect further improvement beyond this point if only this single item is used.

Technical Expert Panel (TEP) Discussion and Conclusions: The TEP unanimously favored multiple-item scales for measuring functional status. Experts emphasized that multiple items provide a more global and comprehensive perspective on patient function, enhancing the face validity of the measure. They also noted that multiple items improve reliability, reduce floor and ceiling effects, and offer more precise discrimination along the spectrum of functional ability. While acknowledging potential concerns about increased burden and expense associated with collecting multiple items, TEP members generally agreed that this information should be routinely collected for comprehensive patient care assessment.

Conclusion: Based on the Rasch analysis and expert consensus, the research concluded that multiple-item scales are preferred over single items for developing functional status quality metrics. Multiple items offer greater precision, reliability, and a more comprehensive assessment of patient functional abilities, which are crucial for robust quality measurement.

Combined Motor Scale vs. Separate Self-Care and Mobility Metrics

Another critical analytic issue was whether to use a combined motor scale or to separate motor function into two distinct quality metrics: self-care and mobility. This question addresses the dimensionality of motor function and whether it’s best represented as a single construct or as separable components.

Analysis and Results: To investigate this, IRF-PAI/FIM motor functional status data were analyzed, focusing on patients with different clinical diagnoses: stroke, central cord syndrome/spinal cord injury, and unilateral hip replacement. These diagnoses were chosen to represent varying patterns of motor and self-care impairments. The analysis aimed to determine if a single motor functional status quality metric would be appropriate for diverse patient populations or if separate metrics for self-care and mobility would provide more accurate and clinically meaningful assessments.

Figure 4-2 from the original document shows the mean admission FIM scores for motor items by diagnosis.

Figure 4-2. Mean Admission FIM® Scores for Items by Diagnosis

This figure reveals that the patterns of motor function scores vary across the three diagnosis groups. The lines representing mean scores for each diagnosis are not parallel, indicating different patterns of activity limitations. For example, patients with hip replacements show relatively lower scores on bathing, lower body dressing, stairs, and walking/wheelchair compared to patients with stroke or central cord syndrome.

Further Rasch analyses were conducted separately for each diagnosis group to examine the hierarchy of motor items. Figures 4-3, 4-4, and 4-5 from the original document illustrate these findings.

Figure 4-3. Map of Motor FIM® Motor Items for Patients With a Stroke

Figure 4-4. Map of Motor FIM® Motor Items for Patients With Central Cord Syndrome

Figure 4-5. Map of FIM® Motor Items for Patients With a Hip Replacement

These figures demonstrate that the order of motor items, from easiest to hardest, differs for patients with different diagnoses. For instance, “toileting” is relatively more challenging for stroke patients compared to patients with central cord syndrome or hip replacement. This varying item hierarchy across diagnoses suggests that the operational definition of motor function is not consistent across these groups, undermining the validity of a single, combined motor scale for diverse populations.

Scatter plots further reinforced this finding. Figure 4-6 and 4-7 from the original document compare self-care and mobility measures against the combined motor score.

Figure 4-6. Scatter Plot of Self-Care and Motor Measures for Patients With a Stroke, Central Cord Syndrome, or Hip Replacement

Figure 4-7. Scatter Plot of Mobility and Motor Measures for Patients With a Stroke, Central Cord Syndrome, or Hip Replacement

These scatter plots show that self-care and mobility measures do not consistently align with the combined motor score across diagnoses. For example, stroke patients tend to have relatively lower self-care scores compared to their overall motor scores, while patients with central cord syndrome tend to have relatively lower mobility scores compared to their overall motor scores. This further supports the idea that self-care and mobility are distinct constructs that should be measured separately, especially when comparing outcomes across diverse patient populations.

Technical Expert Panel (TEP) Discussion and Conclusions: TEP members agreed that separate quality metrics for self-care and mobility would be more appropriate than a single combined motor metric, particularly for measures intended to apply to patients with various diagnoses. They emphasized that separate metrics would provide more precise and accurate measurements, offering better feedback for clinicians and enhancing their ability to improve patient outcomes. Separating motor function into self-care and mobility scales was also seen as improving the face validity and clinical relevance of the quality metrics.

Conclusion: Based on the analysis of IRF-PAI/FIM data and expert consensus, the research concluded that separate quality metrics for self-care and mobility are recommended over a combined motor scale. This separation accounts for the distinct patterns of functional limitations observed in different patient populations and provides more precise and clinically meaningful measures of functional status outcomes.

Change Value vs. Discharge Status for Performance Scores

A significant methodological question in quality metric development is whether performance scores should be calculated as a change value (discharge score minus admission score) or as a discharge status, both adjusted for admission functional status. Both approaches have been used in existing quality metrics, and each has its own set of advantages and limitations.

Analysis and Results: To explore this issue, the relationship between change scores and discharge scores was examined using IRF-PAI/FIM data for a clinically homogeneous group of stroke patients (case-mix group 0110). Scatter plots were created to visualize this relationship for both self-care and mobility scores. Figures 4-8 and 4-9 from the original document depict these relationships.

Figure 4-8. Scatter Plot of Change and Discharge Scores: Self-Care

Figure 4-9. Scatter Plot of Change and Discharge Scores: Mobility

These scatter plots reveal a strong positive correlation between change scores and discharge scores for both self-care (correlation = 0.876) and mobility (correlation = 0.976) within this homogenous patient group. As change scores increase, discharge scores also tend to increase, indicating a close relationship between improvement and final functional status. This suggests that, at least for this patient subgroup, using either change or discharge scores for functional status quality metrics might yield similar results.

However, it’s crucial to acknowledge the methodological limitations associated with change scores, including:

Regression to the Mean: Change scores can be influenced by regression to the mean, particularly for extreme baseline scores.
Reliability Issues: Change scores can be less reliable than individual baseline and follow-up scores due to the accumulation of measurement error.
Floor and Ceiling Effects: Instruments may not be sensitive to the full range of change, especially for patients starting at the extremes of the functional spectrum.
Interpretability: The clinical meaning of a change score can be unclear without established clinically meaningful thresholds.

Despite these limitations, change scores are often intuitively appealing as they directly reflect improvement during the care episode. Discharge status, on the other hand, represents the patient’s functional level at the end of care, which is also a critical outcome measure, particularly in terms of returning patients to the community.

Technical Expert Panel (TEP) Discussion and Conclusions: TEP members suggested further examination of both change and discharge score options. Some members preferred discharge scores, considering them suitable for both stay-level and episode-level quality metrics, and emphasized the importance of discharge status as a predictor of community reintegration. Concerns were raised about the methodological limitations of change scores, but their face validity in representing improvement was also acknowledged. The panel noted that the choice between change and discharge scores might depend on the specific patient subgroup and the goals of care. Importantly, it was highlighted that change and discharge scores could be comparable if appropriately case-mix adjusted, including adjustment for admission status. Several TEP members recommended developing both change and discharge metrics to provide a more comprehensive assessment of functional status outcomes.

Conclusion: The analysis and expert discussion suggest that both change value and discharge status are viable options for functional status quality metrics. Given the strong correlation observed between them and the different strengths and weaknesses of each approach, further research and consideration are needed to determine the optimal choice or combination of these measures, potentially developing both types of metrics to offer a more complete picture of functional outcomes.

Mean, Median, or Percentage Meeting a Benchmark for Reporting Scores

Another key decision in quality metric development is how to aggregate and report functional status scores at the facility level. The options considered were reporting mean (average) values, median values, or the percentage of patients meeting or exceeding a pre-defined benchmark. Each of these approaches presents different ways of summarizing facility performance and has implications for interpretability and sensitivity to variations in patient populations.

Analysis and Results: The research did not directly analyze the comparative performance of these reporting methods using CARE data in this phase. However, the issue was thoroughly discussed with the Technical Expert Panel (TEP), drawing upon existing practices and considerations in quality measurement.

Currently, the Joint Commission ORYX measures for IRF-PAI/FIM instruments utilize mean change scores and mean discharge scores for IRFs. In contrast, some quality metrics, like the “Change in basic mobility as measured by the AM-PAC,” report the proportion of patients meeting a target level of improvement. Benchmarks could be defined based on national averages, clinically important values, or minimal detectable change (MDC) thresholds.

A key concern with using mean or median values is that they might obscure important information about individual patient outcomes, particularly given the heterogeneity of patient populations in PAC settings. Mean values can be influenced by outliers and may not accurately represent the typical patient experience within a facility. Median values are less sensitive to outliers but still provide an average-like summary.

Percentage-based measures, using benchmarks, offer the potential to focus on clinically meaningful targets and may be more easily interpretable by clinicians and consumers. Benchmarks can be set to represent clinically significant improvement or the achievement of specific functional goals. However, defining appropriate and universally applicable benchmarks can be challenging.

Technical Expert Panel (TEP) Discussion and Conclusions: Some TEP members favored using percentage values, such as the percentage of patients reaching a benchmark score, as they can be more clinically meaningful and easier to understand. The idea of a staging approach, categorizing patients into functional stages and tracking movement between stages, was also discussed as a potentially valuable direction, although it required further research with CARE data.

Concerns were raised about using mean or median values alone, as they might suggest that 50% of patients are below average, which could be misinterpreted or demotivating. The panel emphasized that patient populations are heterogeneous, and important individual patient information might be lost in a mean-only measure. Some members supported a national benchmark approach, focusing on patients achieving clinically meaningful goals. The importance of reporting confidence intervals alongside mean or median values was also mentioned to reflect the uncertainty inherent in performance estimates.

Conclusion: The TEP discussion highlighted that no single reporting method is universally optimal. Further analyses are needed to compare the performance and interpretability of mean, median, and percentage-based reporting methods using CARE data. A combination of approaches, potentially including percentage of patients meeting benchmarks alongside mean or median values with confidence intervals, might provide a more nuanced and informative picture of facility performance on functional status outcomes.

Summed Raw Scores vs. Rasch Measures for Scaling

When using multiple items to construct functional status scales, a critical choice arises: should quality metrics use summed raw scores (unweighted or weighted) or Rasch-based measures? This decision impacts the psychometric properties of the metric and the interpretability of scores.

Analysis and Results: Figure 4-10 from the original document illustrates the relationship between raw, unweighted, summed self-care scores and Rasch self-care measures using CARE item set data.

Figure 4-10. Scatter Plot of Raw Self-Care Scores and Rasch Self-Care Measures

This figure shows that in the middle range of the scale, the relationship between raw summed scores and Rasch measures is approximately linear. However, at the extremes of the scale (very low and very high functional status), the relationship becomes non-linear. For instance, at the high end of the scale, a large difference in Rasch measures (e.g., 80 to 100) corresponds to a relatively smaller difference in raw summed scores (e.g., 38 to 42).

This non-linearity arises because raw scores are ordinal-level data, where the intervals between scale points are not necessarily equal. Rasch analysis, in contrast, transforms ordinal data into interval-level measures, ensuring equal intervals across the scale. This transformation can be particularly important at the extremes of the functional spectrum, where raw scores may not accurately reflect the magnitude of functional differences.

For patients with functional status levels in the middle range, the choice between raw summed scores and Rasch measures might not drastically alter conclusions. However, for patients at very low or very high functional levels, using Rasch measures may provide a more accurate and sensitive assessment of functional status and change, as they account for the non-linear nature of ordinal scales. Furthermore, Rasch measures offer the advantage of being estimable even with some missing data, which is a practical benefit in clinical settings. However, the calculation of Rasch-derived measures is less transparent than simply summing raw scores, which could be a concern for some stakeholders.

Technical Expert Panel (TEP) Discussion and Conclusions: TEP members generally supported continued exploration of both summed raw scores and Rasch measures. They emphasized the need for the analysis to be meaningful and understandable to the target audience. Some members felt comfortable with Rasch measures scaled from 0 to 100, considering this range understandable to the public. The panel recognized the psychometric advantages of Rasch measures but also acknowledged the simplicity and familiarity of raw scores.

Conclusion: The research suggests that both summed raw scores and Rasch measures have potential for use in functional status quality metrics. Rasch measures offer psychometric advantages, particularly in terms of interval-level scaling and handling non-linearity and missing data. However, raw scores are simpler and more transparent. The optimal choice might depend on the specific goals of the quality metric, the target audience, and the need for psychometric rigor versus interpretational simplicity. Further comparative analyses and stakeholder input are needed to guide this decision.

Defining the Target Population for Quality Metrics

Defining the target population (denominator) for functional status quality metrics is crucial for ensuring that metrics are relevant, fair, and accurately reflect the quality of care provided by different PAC provider types. This involves specifying which patients should be included in the metric calculations for IRFs, SNFs, LTCHs, and HHAs, considering the distinct patient populations and care goals within each setting.

Analysis and Results: The research did not involve specific data analyses to define target populations. Instead, this issue was primarily addressed through discussions with the Technical Expert Panel (TEP), drawing upon their clinical expertise and understanding of PAC settings.

The core question is whether functional status quality metrics should apply to all patients within each provider type or only to specific subgroups. Functional status is relevant for virtually all patients, but the primary goals of care and the intensity of rehabilitation services vary across PAC settings:

IRFs: Primarily focus on intensive rehabilitation with restorative goals. Functional improvement is a central aim for most patients.
SNFs: Offer a range of services, including rehabilitation and skilled nursing care. Restorative goals are common, but intensity may be lower than in IRFs.
HHAs: Provide care in patients’ homes, often with functional goals to maintain or improve independence in daily living. Services may also include medical care, such as wound care.
LTCHs: Treat medically complex patients, often with limited mobility. Medical management is the primary focus, with functional goals often secondary and more modest.

Despite these differences, functional status is a relevant aspect of care in all PAC settings. Even for patients with primarily medical goals, maintaining or preventing decline in function is an important aspect of quality care. For example, preventing functional decline in medically complex LTCH patients or HHA patients receiving wound care is a valid quality goal.

Technical Expert Panel (TEP) Discussion and Conclusions: The consensus among TEP members was to include all patients within each provider type in the target population. They argued that functional status is an important outcome for all patients in PAC settings, regardless of their primary diagnosis or care goals. While acknowledging the need to account for different patient treatment goals through case-mix adjustment, the panel felt that excluding certain patient groups would limit the scope and impact of functional status quality metrics. They emphasized that even for patients with primarily medical needs, preventing functional decline is a relevant quality indicator.

Conclusion: Based on expert consensus, the research recommends that the target population for functional status quality metrics should encompass all patients within each PAC provider type (IRF, SNF, LTCH, HHA). This broad inclusion ensures that the metrics are comprehensive and reflect the importance of functional status across the entire spectrum of post-acute care. Case-mix adjustment is essential to account for variations in patient populations and care goals across settings and diagnoses.

Exclusion Criteria for Functional Status Quality Metrics

Defining appropriate exclusion criteria is crucial for ensuring the fairness and validity of functional status quality metrics. Exclusion criteria aim to remove patients from the metric calculations for whom functional status outcomes are not meaningfully attributable to the quality of PAC services or are unduly influenced by factors outside of provider control.

Analysis and Results: The research did not involve data analysis to determine exclusion criteria. This issue was primarily addressed through discussions with the Technical Expert Panel (TEP), leveraging their clinical judgment and experience.

Potential exclusion criteria considered included:

Incomplete Stays: Patients who do not complete their intended course of treatment due to death, unexpected discharge to acute care, or discharge against medical advice. Including these patients could distort facility-level performance as functional status at discharge may not be accurately assessed or reflect the full impact of PAC services.
Lack of Baseline Functional Limitations: Patients who have no functional limitations in a specific domain (e.g., self-care, mobility, communication) at admission. Measuring change or discharge status is not meaningful for these patients in that specific domain, as there is limited room for improvement. However, these patients might still be included in other functional status metrics if they have limitations in other domains.
Unpredictable Functional Outcomes: Patients with conditions that inherently lead to unpredictable or fluctuating functional status, such as persistent vegetative state or rapidly progressive neurological conditions like multiple sclerosis. Outcomes for these patients may be less sensitive to PAC interventions and more driven by the underlying disease trajectory.
Expected Functional Decline: Patients with conditions where functional decline is expected despite optimal care, such as amyotrophic lateral sclerosis (ALS). Including these patients could unfairly penalize providers caring for these complex populations.

Technical Expert Panel (TEP) Discussion and Conclusions: TEP members generally agreed on the appropriateness of several exclusion criteria:

Exclude patients with incomplete stays: Including discharges to acute care, discharges against medical advice, and patient deaths.
Exclude patients with unpredictable functional outcomes: Such as persistent vegetative state and multiple sclerosis.
Exclude patients with expected functional decline: Such as ALS.

The panel felt that further discussion was needed regarding the exclusion of patients who do not need or receive therapy treatment. While functional status is relevant for all patients, the direct impact of rehabilitation services might be less pronounced for patients primarily receiving medical management. However, excluding patients based on therapy receipt could introduce complexity and potential bias in metric application.

Conclusion: Based on expert consensus, the research recommends excluding patients with incomplete stays, unpredictable functional outcomes, and expected functional decline from functional status quality metric calculations. Further consideration is needed regarding the exclusion of patients who do not receive therapy treatment. Careful and precise definitions of exclusion criteria are essential to ensure the fairness and validity of functional status quality metrics across PAC providers.

Case-Mix Adjustment Process and Variables

Case-mix adjustment, also known as risk adjustment, is a critical component of functional status quality metrics. It aims to account for differences in patient populations across facilities, allowing for fair comparisons of provider performance. Patients’ functional status outcomes are influenced not only by the quality of care but also by various patient characteristics, such as age, clinical complexity, diagnosis, and admission functional status. Case-mix adjustment statistically controls for these non-treatment-related factors, isolating the effect of provider care quality on patient outcomes.

Analysis and Results: To inform case-mix adjustment strategies, regression analyses were conducted using IRF-PAI/FIM data. These analyses examined the relationship between discharge self-care scores (dependent variable) and admission self-care scores and age (independent variables), separately for patients with stroke, hip fracture, lower extremity joint replacement, and all patients combined.

Figure 4-11 from the original document graphically represents the regression model predicting discharge self-care scores by diagnosis.

Figure 4-11. Graphic of Regression Model Predicting Discharge Self-Care Scores by Diagnosis

This figure illustrates the regression lines for each diagnosis group, showing the predicted discharge self-care score based on admission self-care score. The slopes of these lines (regression coefficients) differ significantly across diagnoses, particularly between patients with lower extremity joint replacement and patients with stroke. The regression coefficient for admission functional status was 0.88 for all patients, 0.93 for stroke patients, 0.81 for hip fracture patients, and 0.46 for lower extremity joint replacement patients. The substantial difference in slope for lower extremity joint replacement patients suggests that the effect of admission functional status on discharge functional status varies by diagnosis. This finding indicates that stratifying by diagnosis might be necessary for effective case-mix adjustment.

The research considered three main approaches to risk adjustment:

Stratification: Dividing patients into risk strata (e.g., high-risk, low-risk groups) based on key risk factors and reporting outcomes separately within each stratum.
Regression Modeling: Using regression models with demographic and clinical covariates to predict outcomes and adjust for patient-level risk factors. Facility-specific predicted and expected values can be calculated for performance comparison.
Combined Approach (Stratification + Regression): Combining stratification by key risk factors with regression modeling within each stratum. This approach is particularly relevant when the effect of covariates on outcomes varies across strata.

The regression analysis results, particularly the diagnosis-specific regression slopes, suggest that a combined approach of stratification and regression might be most appropriate for functional status quality metrics. Stratification by primary diagnosis could account for the varying relationships between admission status and discharge status across different diagnostic groups. Regression modeling within each stratum could then further adjust for other relevant risk factors.

Technical Expert Panel (TEP) Discussion and Conclusions: TEP members agreed that further analyses are needed to explore different case-mix adjustment approaches. They supported the idea of defining strata based on primary diagnosis. Additional factors suggested by TEP members for potential inclusion as risk adjusters included:

Case-mix index
Surgical vs. nonsurgical treatment
Baseline functional status (admission scores)
Cognitive function at admission
Depression levels at admission
Geographic region
Urbanicity
Length of stay at the acute care hospital prior to PAC
Prior level of function
Age
Gender

The panel emphasized the importance of selecting risk adjusters that are clinically relevant, reliably measured, and not directly influenced by provider interventions.

Conclusion: Based on the regression analyses and expert consensus, the research recommends a combined approach of strata and regression modeling for case-mix adjustment of functional status quality metrics. Stratification by primary diagnosis is suggested, with regression modeling within each stratum to further adjust for other relevant patient-level risk factors. Further research is needed to refine the specific risk adjustment model, including the selection of optimal risk adjusters and the statistical methodology (e.g., hierarchical models vs. fixed-effects models).

Conclusions and Recommendations for Functional Status Quality Metrics

The research project, encompassing an environmental scan, data analyses using the Continuity Assessment Record and Evaluation (CARE) tool and other functional status data, and input from a Technical Expert Panel (TEP), provides a strong foundation for advancing the development of functional status quality metrics in post-acute care. Based on the findings, several key conclusions and recommendations emerge:

1. Focus on Self-Care and Mobility Metrics: The research recommends further development of two distinct motor functional status quality metrics: self-care and mobility. Separating these constructs provides more accurate and clinically meaningful assessments compared to a combined motor scale, particularly for diverse patient populations across PAC settings.

2. Utilize Multiple Items for Scale Construction: Multiple-item scales are preferred over single-item measures for functional status quality metrics. Multiple items enhance reliability, reduce floor and ceiling effects, and provide a more comprehensive and precise assessment of patient function.

3. Further Explore Change and Discharge Score Metrics: The research recommends additional analyses comparing metrics based on discharge scores and metrics based on change scores. Both approaches have merits and limitations, and the optimal choice or combination may depend on the specific context and goals of the quality metric. Developing both types of metrics could offer a more complete picture of functional status outcomes.

4. Investigate Percentage-Based and Mean/Median Reporting: Further analyses are needed to compare metrics that report the percentage of patients meeting a benchmark with metrics that use mean or median values. Percentage-based measures may be more clinically interpretable, while mean/median values offer different summary perspectives. A combination of reporting approaches might be optimal.

5. Employ Rasch Measures and Consider Raw Scores: Both Rasch measures and summed raw scores warrant further consideration for scaling functional status items. Rasch measures offer psychometric advantages, particularly for interval-level scaling and handling non-linearity. However, raw scores are simpler and more transparent. The choice should be guided by the desired balance between psychometric rigor and interpretational ease.

6. Target Population Should Include All Patients: The target population for functional status quality metrics should include all patients within each PAC provider type (IRF, SNF, LTCH, HHA). Functional status is a relevant outcome for all patients in these settings, and broad inclusion ensures comprehensive quality assessment.

7. Implement Exclusion Criteria for Specific Situations: Exclusion criteria should be implemented to remove patients with incomplete stays, unpredictable functional outcomes, and expected functional decline from metric calculations. Precisely defined exclusion criteria are essential for fair and valid quality measurement.

8. Utilize a Combined Case-Mix Adjustment Approach: A combined approach of strata and regression modeling is recommended for case-mix adjustment. Stratification by primary diagnosis, combined with regression modeling within each stratum, can effectively account for patient-level risk factors and ensure fair comparisons of provider performance.

9. Future Directions for Cognition and Swallowing Metrics: While this research primarily focused on motor function, the report notes the need for further development of quality metrics in the area of cognition. Challenges include the complexity of measuring cognitive function and communication, as well as accounting for varying recovery trajectories based on etiology. Swallowing disorders are also identified as a potential future area for functional status quality metrics development, given the availability of relevant items in the CARE data set and MDS 3.0.

In conclusion, this research provides valuable insights and recommendations for the development of robust and clinically meaningful functional status quality metrics in post-acute care, leveraging the Continuity Assessment Record and Evaluation (CARE) tool. By addressing key analytic issues and incorporating expert clinical input, this work paves the way for improved quality measurement, payment reform, and ultimately, enhanced patient care across the PAC continuum.

References

Ash, A. S., Fienberg, S. E., Louis, T. A., Normand, S.-L. T., Stukel, T. A., & Utts, J. (2012). Statistical issues in assessing hospital performance. Retrieved from http://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/HospitalQualityInits/Downloads/Statistical-Issues-in-Assessing-Hospital-Performance.pdf

Collins, L. M., & Johnston, M. V. (1995). Analysis of stage-sequential change in rehabilitation research. American Journal of Physical Medicine & Rehabilitation, 74(2), 163-170.

DeSalvo, K. B., Fan, V. S., McDonell, M. B., & Fihn, S. D. (2005). Predicting mortality and healthcare utilization with a single question. Health Serv Res, 40(4), 1234-1246. doi:10.1111/j.1475-6773.2005.00404.x

Donabedian, A. (2005). Evaluating the quality of medical care. 1966. Milbank Q, 83(4), 691-729. doi:10.1111/j.1468-0009.2005.00397.x

Food and Drug Administration. (2009). Guidance for industry: Patient-reported outcome measures: Use in medical product development to support labeling claims. Retrieved from http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM193282.pdf

Fortinsky, R. H., Garcia, R. I., Joseph Sheehan, T., Madigan, E. A., & Tullai-McGuinness, S. (2003). Measuring disability in Medicare home care patients: Application of Rasch modeling to the outcome and assessment information set. Med Care, 41(5), 601-615. doi:10.1097/01.MLR.0000062553.63745.7A

Granger, C. V., Hamilton, B. B., Linacre, J. M., Heinemann, A. W., & Wright, B. D. (1993). Performance profiles of the functional independence measure. Am J Phys Med Rehabil, 72(2), 84-89.

Grimby, G., Tennant, A., & Tesio, L. (2012). The use of raw scores from ordinal scales: Time to end malpractice? J Rehabil Med, 44(2), 97-98. doi:10.2340/16501977-0938

Haley, S. M., Andres, P. L., Coster, W. J., Kosinski, M., Ni, P., & Jette, A. M. (2004). Short-form activity measure for post-acute care. Arch Phys Med Rehabil, 85(4), 649-660.

Haley, S. M., Ni, P., Lai, J. S., Tian, F., Coster, W. J., Jette, A. M., Straub, D., & Cella, D. (2011). Linking the activity measure for post acute care and the quality of life outcomes in neurological disorders. Arch Phys Med Rehabil, 92(10 Suppl), S37-43. doi:10.1016/j.apmr.2011.01.026

Heinemann, A. W., Linacre, J. M., Wright, B. D., Hamilton, B. B., & Granger, C. (1993). Relationships between impairment and physical disability as measured by the functional independence measure. Arch Phys Med Rehabil, 74(6), 566-573.

Hula, W. D., Doyle, P. J., & Austermann Hula, S. N. (2010). Patient-reported cognitive and communicative functioning: 1 construct or 2? Arch Phys Med Rehabil, 91(3), 400-406. doi:10.1016/j.apmr.2009.11.013

Iezonni, L. (2003). Risk adjustment for measuring healthcare outcomes (3rd ed.). Chicago, IL: Health Administration Press.

Jette, A. M., Haley, S. M., & Ni, P. (2003). Comparison of functional status tools used in post-acute care. Health Care Financ Rev, 24(3), 13-24.

Jette, A. M., Haley, S. M., Tao, W., Ni, P., Moed, R., Meyers, D., & Zurek, M. (2007). Prospective evaluation of the AM-PAC-CAT in outpatient rehabilitation settings. Phys Ther, 87(4), 385-398. doi:10.2522/ptj.20060121

Kane, R. L., & Radosevich, D. M. (2011). Conducting health outcomes research. Sudbury, MA: Jones & Bartlett.

Linacre, J. M. (2012). Winsteps® Rasch Measurement computer program (Version 3.75). Beaverton, OR: Winsteps.com.

Linacre, J. M., Heinemann, A. W., Wright, B. D., Granger, C. V., & Hamilton, B. B. (1994). The structure and stability of the Functional Independence Measure. Archives of Physical Medicine & Rehabilitation, 75(2), 127-132.

McGlynn, E. A., & Asch, S. M. (1998). Developing a clinical performance measure. Am J Prev Med, 14(3 Suppl), 14-21.

National Committee on Vital and Health Statistics Subcommittee on Health. (2001). Classifying and Reporting Functional Status (Vol. 2012, pp. 1).

National Quality Forum. (2012). Quality Positioning System Retrieved Sept. 21, 2012, from http://www.qualityforum.org

Stineman, M. G., Henry-Sanchez, J. T., Kurichi, J. E., Pan, Q., Xie, D., Saliba, D., Zhang, Z., & Streim, J. E. (2012). Staging activity limitation and participation restriction in elderly community-dwelling persons according to difficulties in self-care and domestic life functioning. American Journal of Physical Medicine & Rehabilitation, 91(2), 126-140.

Stineman, M. G., Maislin, G., Fiedler, R. C., & Granger, C. V. (1997). A prediction model for functional recovery in stroke. Stroke, 28(3), 550-556.

Stineman, M. G., Ross, R. N., Fiedler, R., Granger, C. V., & Maislin, G. (2003a). Functional independence staging: conceptual foundation, face validity, and empirical derivation. Archives of Physical Medicine & Rehabilitation, 84(1), 29-37.

Stineman, M. G., Ross, R. N., Fiedler, R., Granger, C. V., & Maislin, G. (2003b). Staging functional independence validity and applications. Archives of Physical Medicine & Rehabilitation, 84(1), 38-45.

Stineman, M. G., Xie, D., Pan, Q., Kurichi, J. E., Saliba, D., & Streim, J. (2011). Activity of daily living staging, chronic health conditions, and perceived lack of home accessibility features for elderly people living in the community. Journal of the American Geriatrics Society, 59(3), 454-462.

Streiner, D. L., & Norman, G. R. (2008). Health measurement scales: A practical guide to their development and use (4th ed.). New York: Oxford University Press.

The Joint Commission. (2012). Performance measurement. JointCommisionOnline, (August 29, 2012). Retrieved from

Velozo, C. A., Byers, K. L., Wang, Y. C., & Joseph, B. R. (2007). Translating measures across the continuum of care: Using Rasch analysis to create a crosswalk between the Functional Independence Measure and the Minimum Data Set. J Rehabil Res Dev, 44(3), 467-478.

Wang, Y. C., Byers, K. L., & Velozo, C. A. (2008). Rasch analysis of Minimum Data Set mandated in skilled nursing facilities. J Rehabil Res Dev, 45(9), 1385-1399.

World Health Organization. (2001). International Classification of Functioning, Disability and Health: Retrieved from http://apps.who.int/classifications/icfbrowser/

Wright, B. D., Linacre, J. M., Smith, R. M., Heinemann, A. W., & Granger, C. V. (1997). FIM measurement properties and Rasch model details. Scand J Rehabil Med, 29(4), 267-272.