The Clinical Meaning of Rankin ‘Handicap’ Grades After Stroke
Background and Purpose The Rankin Scale is a frequently used handicap index in stroke outcome research. However, relatively little is known about its validity. The purpose of this study was to investigate the clinical meaning of Rankin grades by identifying the functional health aspects that contribute to Rankin scores.
Methods We studied 438 patients 6 months after stroke. Data were collected on the following functional health indicators: alertness, communication, independence, disability in activities of daily living, mobility, instrumental disability, social interaction, and recreation. Disability in activities of daily living was assessed with the Barthel Index, whereas the other indicators were measured with subscales of the Sickness Impact Profile. The association between functional health and Rankin Scale was expressed in terms of relative frequencies and Somers’ D statistic. Linear regression analysis (after ordinal transformation) was used to identify the significant health factors that explain Rankin scores.
Results Mobility, disability in daily and instrumental activities, and living arrangements showed a stronger association with Rankin scores (Somers’ D range, 0.60 to 0.74) than cognitive and social functioning (Somers’ D range, 0.34 to 0.47). Disability in activities of daily living turned out to be the most important explanatory factor of Rankin scores (R2=67%).
Conclusions The Rankin Scale is not a pure handicap measure but should be viewed as a global functional health index with a strong accent on physical disability. The index is useful as a simple and time-efficient outcome measure in large-scale multicenter trials. It is argued that at present there is no clear need to assess handicap as the primary outcome in medically oriented stroke intervention studies.
In 1980, the World Health Organization published the International Classification of Impairments, Disabilities, and Handicaps (ICIDH). Impairments are defined as organic dysfunctions, disabilities as a patient’s difficulty with tasks, and “handicap” is defined as the social “disadvantage for a given individual, resulting from an impairment or a disability, that limits or prevents the fulfillment of a role that is normal (depending on age, sex, social, and cultural factors) for that individual.”1
In the past years, many health indexes have been constructed within this conceptual framework. Most of these instruments focus on impairments (eg, dysarthria, ataxia, and hemianopsia), disabilities in activities of daily living (ADL) (eg, bathing, dressing, and toileting), and instrumental disabilities (eg, meal preparation, shopping, and outdoor mobility).
Of the few available handicap instruments, the Rankin Scale is without doubt the most well-known measure in stroke outcome research.2 The original index was a 5-point rating scale that graded patients on their overall level of independence with reference to previous activities. Although the reliability of the Rankin Scale in terms of interobserver agreement and reproducibility has been evaluated with satisfactory results,3 4 relatively little is known about the validity of the index. One of the main objections to the Rankin Scale is that it does not rate handicap but rather disability.5 In the original scale, for example, the word “disability” was used, whereas the ability to walk was an explicit grading criterion. To overcome this conceptual problem, the original scale has been modified by replacing the word “disability” by “handicap,” by introducing the term “lifestyle” into the scoring categories, and by deleting ambulation as a scoring indicator.6
The present study was undertaken to investigate the clinical meaning of Rankin grades by identifying which aspects of cognitive, physical, and social health domains are associated with Rankin scores.
Subjects and Methods
The study group consisted of 438 patients who had had a stroke 6 months earlier. They were the survivors of an original cohort of 760 consecutively admitted stroke patients who participated in a multicenter quality-of-care study in The Netherlands. Two hundred fifty-eight patients died after the stroke, and 64 patients were not included in the study: 17 patients declined to participate in the study, for 3 patients there was no Rankin score available, and for 44 patients no data on other functional health aspects were collected during the follow-up interview. The reason for this was that the patient burden associated with the length of the interview was unacceptably high, so the research assistant had to stop the interview prematurely.
Six months after the stroke, patients were interviewed by trained research assistants using a semistructured questionnaire. One hundred twelve patients (26%) were not able to communicate because of severe speech, language, or cognitive disorders. These latter patients were rated by a proxy respondent (most typically the partner). During the interviewer training, it was stressed that any limitations in the patient’s social role functioning should be assessed.
Interview data were collected on the following indicators of functional health: cognitive alertness behavior and ability to communicate; level of disability in ADL, mobility, and extended daily activities (instrumental disability); patients’ present living arrangements; and limitations in social interactions and recreation. These cognitive, physical, and social indicators approximately reflect the formal ICIDH handicap dimensions: orientation handicap, physical independence handicap, mobility handicap, and social integration handicap.1 Since only a minority of the patients (20%) were employed at stroke onset, we did not focus on indicators that reflect the occupation and economic self-sufficiency handicap dimensions.
ADL disability was assessed with the Barthel Index,7 whereas cognitive alertness behavior, ability to communicate, mobility, instrumental disability, social interaction, and recreation were measured with subscales of the Sickness Impact Profile (SIP).8 On the basis of this information, the interviewer scored the patient’s level of handicap with the modified Rankin Scale.6
To evaluate the meaning of Rankin grades, the associations between the functional health indicators and the Rankin Scale were examined. To enhance the comparability between the various (sub)scales, we divided the indicator scores into three categories. SIP scores were categorized as mildly (score ≤median), moderately (score >median and ≤75th percentile), and severely (score >75th percentile) limited. Barthel Index scores were classified as no disability (score=20), mild disability (score ≥10 and <20), and severe disability (score <10). Patients’ living arrangements were divided into independent, rehabilitation center, and nursing home categories. The association between indicator scores and Rankin scores was expressed in terms of relative frequencies and Somers’ D statistic. Somers’ D is an asymmetrical index of relation between one ordered independent variable and another ordered dependent variable, and it ranges in values between −1 and +1.9 The extremes reflect a perfect association, whereas the value 0 indicates an absence of association. We assumed that, for the Rankin Scale to be a valid handicap index, not only the physical but also the cognitive and social indicators had to associate substantially with the Rankin Scale.
Because it was expected that various aspects of functional health would be interrelated (eg, ADL disability, mobility, living arrangements, and social integration), the association of functional health indicators with Rankin scores was also analyzed with linear regression analysis. Because of the ordinal characteristic of the Rankin Scale, an ordinal transformation of the Rankin scores was performed (using the SAS procedure TRANSREG). The indicator scores were treated as continuous variables. In this analysis all variables were included into the model.
The study group consisted of 438 patients (54% men; median age, 73 years). Characteristics of the 64 excluded patients of stroke type, lesion location, age, and sex were comparable to those of the study group.
The majority of the patients (62%) scored in the middle range of the Rankin Scale (minor or moderate handicap), whereas only a few patients (7%) were grouped in the extreme score categories. The median Rankin score was 3 (Table 1⇓).
The association between functional health indicators and Rankin Scale scores is presented in Table 2⇓. Patients who were mildly disabled in cognitive, physical, and social functions, as well as those who were living independently, had higher relative frequencies of Rankin categories 0, 1, and 2. An opposite pattern was seen in the patients with serious cognitive impairments or social limitations. These patients had higher relative frequencies of Rankin grades 3, 4, or 5. Patients with severe disability in mobility and (extended) daily activities showed a somewhat different picture. These physically disabled patients were clustered more frequently in the fourth and fifth Rankin scoring category. Most of the severely affected patients were living in nursing homes. In general, the moderately disabled patients were clustered in Rankin category 3. Mobility, ADL and instrumental activities, and living arrangement were more highly associated with Rankin grades (Somers’ D range, 0.60 to 0.74) than the level of cognitive and social dysfunction (Somers’ D range, 0.34 to 0.47).
A linear model to identify which aspects of functional health were associated most with Rankin scores (after ordinal transformation) is presented in Table 3⇓. Seventy-eight percent of the total variation of Rankin scores could be explained by the fitted model. ADL disability and, to a much lesser extent, instrumental disability and living arrangement were the most important factors; together they explained 76% of the total variance of Rankin scores. Separate analyses in our sample did not show a significant difference of regression results between patients with and without the ability to communicate.
In this study, we examined the association between functional health indicators and Rankin Scale scores after stroke. On average, mildly limited stroke patients most often had Rankin grades 0, 1, and 2. Moderately limited patients were more likely to be clustered in scoring category 3, whereas in general the severely affected and institutionalized patients were more often classified as having Rankin grades 4 and 5.
Univariate analysis showed pronounced associations between (complex) daily activities, mobility, and living arrangement on the one hand and Rankin grades on the other. Multivariate analysis largely supported these findings. Disability in ADL and, to a considerably lesser extent, instrumental disability and living arrangement were found to be the most important factors associated with Rankin scores. The other indicators, including the cognitive and social health domains, did not explain variation in Rankin scores. In other words, Rankin scores appear to reflect primarily a global level of disability rather than handicap.
During interviewer training, the importance of considering patients’ level of social role functioning in generating Rankin scores was stressed. Additionally, particular attention was given to the conceptual difference between disability and handicap in general as well as the scoring of disability instruments and the Rankin Scale in particular. Nevertheless, the present results indicate that to broaden the scope of the Rankin Scale (or of the interviewer) to include more handicap-related health aspects one probably has to introduce more nuances in the globally defined Rankin scoring categories and introduce a standardized observational procedure.
Before discussing the usefulness of the Rankin Scale in stroke outcome studies, two questions related to the concept of handicap need to be addressed: (1) What precisely is meant by handicap? and (2) Should we focus on this concept in stroke medical-intervention studies?
As has been noted previously, impairment and disability are currently better conceptualized than handicap.10 The ICIDH scheme defines handicap as the disadvantageous situation of an impaired or disabled patient that prevents him from fulfilling survival roles. However, the ICIDH indicators of the handicap dimension do not refer explicitly to the patient’s circumstances and situation but rather to his functional abilities, activities, and competence (eg, physical independence, mobility, and the ability to work).11 12 As long as the handicap section of the ICIDH contains such confusing disability-related elements, problems will remain in constructing indexes to measure distinct aspects of handicap and disability.
The extent to which an impaired and disabled stroke patient fulfills normally expected social roles and relates adequately with his social environment is of major importance for his overall quality of life. For a patient whose medical treatment has reached its limits, the quality of life can probably be improved only by reducing handicap.13 Assessment of the individual patient’s level of handicap is therefore particularly relevant in rehabilitation settings. Other areas in which the concept of handicap can be used meaningfully are evaluation of a patient’s need for nonmedical care services, care planning, and healthcare policy. Handicap, however, is extremely broadly defined and therefore difficult to measure. In this context, Wade14 outlines three major problems: (1) Handicap has to be assessed with reference to specific culturally based expectations of the individual patient, so there is no absolute standard for judging handicap; (2) Since handicap arises from the interaction between disability and environment, it is also necessary to assess the patient’s environment (eg, public attitudes and economic aspects); and (3) In contrast to disability, which refers to skills and behaviors, handicap cannot be observed directly. In view of these problems, we feel that at present handicap is not specifically the field of the physician and that, in the evaluation of medical treatments for stroke patients, physicians should generally not focus on handicap as a primary outcome but on a more tangible manifestation of disease in terms of disability.
Despite its modifications, the Rankin Scale should not be viewed as a pure handicap measure but rather as a global functional health index with a strong accent on physical disability. That is not to say that the Rankin Scale has no value in stroke outcome research. To the contrary, the index can be used as an efficient tool to describe global functional health levels. The index is also useful as a simple and time-efficient functional outcome measure in large-scale multicenter trials. In the latter case, for example, Rankin scores can be divided into two frequently used scoring categories, grades 0 to 3 (reflecting mild to moderate disablement) and grades 4 and 5 (referring to severe disablement), such that an intervention effect can be expressed in terms of “relative risk.” Our results support the validity of such a dichotomy.
To summarize, although there are significant associations between various functional health indicators and Rankin grades, our study indicates that physical disability is by far the most important health factor associated with Rankin grades. Nevertheless, the index can be used as an efficient global functional outcome index after stroke. At this moment, there is no clear need to assess handicap as the primary outcome in medically oriented stroke intervention studies.
This study was funded by the Netherlands Heart Foundation (NHS 40.004) and “Ontwikkelingsgeneeskunde” (OG 1991-037). We wish to thank Professor Dr H. van Crevel and Professor Dr M. Vermeulen of the Department of Neurology, Academic Medical Center, Amsterdam, the Netherlands, for supporting this study.
- Received March 23, 1995.
- Revision received July 25, 1995.
- Accepted July 27, 1995.
- Copyright © 1995 by American Heart Association
World Health Organization. International Classification of Impairments, Disabilities and Handicaps. Geneva, Switzerland: World Health Organization; 1980.
Van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJA, Van Gijn J. Interobserver agreement for the assessment of handicap in stroke patients. Stroke. 1988;19:604-607.
Wolfe CDA, Taub NA, Woodrow EJ, Burney PGJ. Assessment of scales of disability and handicap for stroke patients. Stroke. 1991;22:1241-1244.
Bloch RF. Interobserver agreement for the assessment of handicap in stroke patients. Stroke. 1988;19:1448. Letter.
Siegel S, Castellan NJ. Nonparametric Statistics for the Behavioral Sciences. 2nd ed. New York, NY: McGraw-Hill Book Co; 1988:303-310.
Task Force on Stroke Impairment, Task Force on Stroke Disability, and Task Force on Stroke Handicap. Symposium recommendations for methodology in stroke outcome research. Stroke. 1990;21(suppl II):II-68-II-73.
Nordenfelt L. On Disabilities and Their Classification: A Study in the Theory of Action Inspired by the International Classification of Impairments, Disabilities and Handicaps (ICIDH). Linköping, Sweden: University of Linköping, Studies on Health and Society, SHS 1; 1983.
Wade DT. Measurement in Neurological Rehabilitation. New York, NY: Oxford University Press; 1992:89-96.