Mapping Patient-Specific Functional Scale (PSFS) Items to the International Classification of Functioning, Disability and Health (ICF)

Discussion

Our findings indicated that items selected by patients completing the PSFS predominantly represented the activity and participation components of the ICF, with low representation of impairment of body functions and structures. Our findings showed that the highest proportion of PSFS items nominated by patients represented the activity component (80%); the participation component was moderately represented when considered alone or in combination with the activity component (12.6%). Impairment of body functions and structures was represented by 7.4% of the PSFS items nominated by patients with musculoskeletal disorders. The lower proportion of impairment-based items indicated that the breadth and depth of the ICF and its components were not fully represented by the PSFS alone.

The high representation of activity-based items could have been a consequence of the standardized phrasing used in the administration of the PSFS. Patients were asked to describe 3 to 5 activities; this phrasing might have caused bias away from impairment- or participation-based items. Jolles et al¹⁰ also raised this point when investigating the content validity of the PSFS, concluding that the PSFS only partially met their content validity standards. The creators of the PSFS did not identify why the word “activity” was used,¹¹ and it is possible that an instrument with different terminology would encourage patients to think beyond activities to include participation or impairment. However, conversely, different terminology might result in weakened representation of activity-based items, and further research would be needed to ascertain whether the change in wording affected the properties of the PSFS. In addition, different terminology might present difficulties in conveying the differences in the components to patients, considering the difficulties encountered in differentiating between participation and activities within the ICF model.^2,19

Although the PSFS strongly represented the activity component of the ICF and partially represented the participation component of the ICF, impairment of body functions and structures was not strongly represented (except in the neck). This finding indicated that the PSFS should be supplemented by complementary outcome measures addressing impairments. Other researchers have noted that physical therapists are more inclined to use outcome measures of impairments, such as pain and range of movement^6,25; the addition of the PSFS would complement these measures by representing the activity and participation components. Brockow et al⁶ reported that the dominance of impairment-based outcome measures also appeared in clinical research. They found that in 836 outcome measures reported in clinical trials for lower back pain, chronic widespread pain, osteoarthritis, osteoporosis, and rheumatoid arthritis, impairment of body structures and functions was most commonly represented.⁶ For lower back pain, chronic widespread pain, and osteoarthritis, the categories most commonly used were sensation and pain; the categories most commonly used for osteoporosis and rheumatoid arthritis were structures related to the trunk and other musculoskeletal structures related to movement, respectively.⁶ By using the PSFS in conjunction with commonly used impairment-based outcome measures for individual patients, physical therapists would capture a wider range of the ICF components. The dominance of activity- and participation-based items in the PSFS might improve coverage of the breadth of the ICF.

We investigated the consistency of representation of ICF components across body region subgroups. The findings from this analysis revealed that patients with injuries in the neck selected a somewhat lower proportion of activity-based items and a higher proportion of impairment-based items than patients with injuries in the other body regions. Most of the impairment-based items were associated with the body function item “turning the head” (ICF code b710: mobility of joint functions). Overall, the neck was shown to have the broadest representation of ICF components because it was associated with the most impairment-based items and 1 of the highest proportions of participation-based items. Differences between body regions were not marked, indicating that the PSFS ably represented all body regions affected by musculoskeletal complaints seen in outpatient physical therapy practice. We saw no notable differences between age ranges with regard to the proportions of items representing each ICF component.

This research has highlighted the difficulties encountered in differentiating between activities and participation within the ICF model. Earlier drafts of the ICF model described activities and participation as 2 separate components (Fig. 1); however, in the 2001 version of the ICF, activities and participation were grouped together.^2,19 Annex 3 of the ICF² suggested 4 ways of classifying items as either activities or participation but did not provide rules for doing so. Dixon and colleagues^21,26 and Perenboom and Chorus²⁷ also found that it was difficult to differentiate between activities and participation when they mapped some frequently used outcome measures into the 3 components of the ICF. A fourth group was created in those studies to capture items that overlapped activities and participation. Other investigators have explored possible ways of separating activities from participation.^28,29 Badley²⁸ classified items into the categories of acts, tasks, and societal involvement but failed to classify tasks as either activities or participation. Jette et al²⁹ differentiated the items in the Late-Life Function and Disability Instrument as distinct activities or participation. However, the nature of the items in that particular outcome measure was such that the researchers could clearly classify items into the ICF components; this method may not produce the same results when applied to other outcome measures.

Other researchers attempting to separate activities and participation used quantitative methods to calculate rates of agreement in differentiating between the components.^21,26,30,31 However, because of the lack of clear differentiation between these 2 components, we chose to take a qualitative approach in the present study; this approach involved discussions by researchers to decide into which components items could be classified.

We differentiated activities and participation within the context of how they were described.³² A limitation of this method is that items were categorized depending on how patients described the items and how therapists interpreted and recorded them. Collapsing activities and participation under a single taxonomy, as is the case for the ICF model from 2001 to the present,^2,19 avoids these classification difficulties.

The PSFS is intended to serve as an efficient, widely applicable, and responsive measure of change over time at the individual level.^11,14 Patient-specific indexes, by definition, focus on issues most relevant to a patient, explicitly from the patient's perspective. In the present study, we attempted to classify patient-nominated PSFS items into the chapters and categories of the 3 main ICF components—body functions and structures, activities, and participation; however, not all aspects of the ICF were represented in our data. The ICF is intended to contain categories for all aspects of human health and well-being; some of these categories were not relevant to the group of patients whose data we studied (patients receiving musculoskeletal physical therapy). The PSFS is not intended to comprehensively represent the breadth and depth of the ICF or even the much more concise Core Sets because it is limited to only 3 to 5 items (median and mode of 3 in our data).

Another point to consider is that the main ICF components—body functions and structures, activities, and participation—should be considered within the context of personal and environmental factors (Fig. 1). Personal and environmental factors may influence which items patients nominate; however, with its focus on activities, the PSFS does not take this possible influence into account, generally failing to represent personal and environmental factors. Silva Drummond et al³³ reached a similar conclusion in linking the Disabilities of Arm, Shoulder, and Hand instrument to the ICF, stating that a limitation of the content validity of the Disabilities of Arm, Shoulder, and Hand instrument was its failure to consider environmental factors.