Scales could be developed based on simple clinical ratings of International Classification of Functioning, Disability and Health Core Set categories

Abstract

Objective

The aim of the study was to examine whether clinical ratings of International Classification of Functioning, Disability and Health categories can be integrated into parametric scales, which provide a reliable estimation of specified patient problems and rehabilitation goals using the example of mobility of the upper and lower extremities in the acute hospital situation.

Study Design and Setting

Psychometrical study based on data from a prospective multicentric cohort study in patients with musculoskeletal conditions in the acute hospital.

Results

Two hundred thirty-four patients were included (mean age 56, 50% female. Forty-four percent with a diagnosis involving lower extremities, 18% with a diagnosis involving the hip, 18% with a diagnosis involving upper extremities, 16% with a diagnosis involving the spine). After adjustment for differential item functioning two separate scales for upper and lower extremity mobility could be constructed. The constructed scales had 10 (upper) and eight (lower) items displaying adequate to good fit.

Conclusion

The results of this study indicate that it could be possible to develop scales based on categories of the International Classification of Functioning, Disability and Health Core Sets. This may be a promising approach for areas where psychometrically sound measures are not available.

Introduction

Human functioning is a central concept in medicine relevant to the individual and the society. Its measurement is therefore essential for research on effective patient care [1]. In rehabilitation practice and research it is the basis for the assessment of patients problems and the setting of rehabilitation goals, the assignment to services and interventions, intervention management, and evaluation of intervention outcomes [2], [3].

Measures of functioning that are useful in the hand of health professionals across health care systems and services need to be simple and based on a universal language of functioning. Measures that are useful for the specification and evaluation of rehabilitation goals need to be tailored to patients’ problems. In rehabilitation research, measures should be etiologically neutral to allow comparisons across health conditions and they should closely reflect rehabilitation goals to be valid and assure responsiveness.

No currently available measures fulfill these requirements. In general, current measures are not simple and are thus often used only for research or health statistics but not in daily clinical practice. Lacking a universal language or classification of functioning current measures vary widely with respect to included items [4], to assumed underlying constructs and accordingly, for example, to the definition and naming of scales and subscales [5].

Currently widely used measures in rehabilitation such as the Functional Independence Measure [6] or the Barthel Index [7] have been developed to measure, for example, resource utilization or to assign patients to appropriate levels of rehabilitation services. However, they are neither comprehensive assessments of functioning [8] nor invariant across different health conditions [9]. Most importantly, they are not specific enough when assessing patients problems and defining rehabilitation goals. Also, most of the current measures used in rehabilitation are categorical measures and hence not qualified for parametric statistics such as calculating means, standard deviations, or even change scores [10].

Thus, there is the need to develop a practical approach which enables health professionals and researchers to develop interval scaled clinical measures which are simple to rate and interpret and useful to specify and follow patients problems and rehabilitation goals in clinical practice. For this, we need first a unifying framework and classification which comprehensively covers the human experience in relation to functioning irrespective of the underlying condition and in relation to a wide range of personal and environmental factors. Second, we need to show whether clinical ratings of the categories of such a universal classification can be integrated into parametric scales, which provide a reliable estimation of specified patient problems and rehabilitation goals.

The International Classification of Functioning, Disability and Health (ICF) [11] has now emerged as the accepted universal language and framework for rehabilitation [12] to describe and classify functioning, health and disability. The ICF is organized into four components, Body Functions, Body Structures, Activities and Participation, and Environmental Factors subdivided into over 1,400 different, hierarchically arranged categories. With both the opportunities of ICF and newly developed methods of item response theory combined it is now possible to examine whether it is in principle possible to develop clinical measures which are tailored to patients problems by simply integrating clinical ratings of relevant ICF categories with the ICF qualifier. The ICF provides qualifier which range from 0 to 4 for each category to account for the severity of the problem. The ICF qualifier can be used as a direct measurement of human functioning by an expert based on given definitions and defined anchors.

When developing clinical measures tailored to patients problems and rehabilitation goals one can rely on parts of the ICF which comprehensively cover aspects of functioning for specific health conditions [13] or health care situations such as the acute hospital or early acute rehabilitation facilities [14]. Those parts, the so-called ICF Core Sets, were developed in an international effort supported by the World Health Association.

If then, for example, a scale is needed to assess mobility of the upper or lower extremities in the acute hospital, ICF categories can be selected out of one of the ICF Core Sets for the acute hospital. By applying item response theory, one can then examine whether the selected categories indeed cover a common underlying trait (such as mobility), thus forming a scale, which of the selected categories have a reasonable fit in relation to the assumed trait, and whether the selected categories cover the spectrum of ability one is likely to encounter in typical patients.

The objective of this paper is to examine whether clinical ratings of ICF categories can be integrated into parametric scales that provide a reliable estimation of specified patient problems and rehabilitation goals using the example of mobility of the upper and lower extremities in the acute hospital situation.

The specific aims were

• to examine whether a subset of categories relevant for the specified problems and selected from the Acute ICF Core Set for Musculoskeletal Conditions constitutes a scale with Rasch properties invariant to selected person factors and

• to identify which categories of the subset can and should be included in a clinical scale to enable simple and valid measurement in clinical practice.