Application of the refinements of ICF linking rules to the Visual Analogue Scale, Roland Morris questionnaire and SF-36 Aplicação do refinamento das regras de ligação da CIF à Escala

Resumo A Escala Visual Analógica (EVA), o Questionário de Incapacidade de Roland Morris (RMDQ) e Questionário de Qualidade de Vida SF-36, amplamente utilizados, tiveram seu conteúdo conectado à CIF por regras propostas em 2002 e 2005. Em 2016 foram refinadas e ainda não foram aplicadas. Aplicar as regras de conexão de conteúdo refinadas para os instrumentos EVA, RMDQ e SF-36. Dois profissionais de saúde identificaram os conceitos significativos e vincularam às categorias mais específicas da CIF, um terceiro arbitrou divergências. O grau de concordância foi dado pelo coeficiente kappa. Houve alto grau de concordância (Kappa=0,93 p<0,001). O conceito principal da EVA foi conectado à categoria b280, os 24 conceitos principais do RMDQ, à categoria b28013 e os 27 adicionais a outras categorias. O SF-36 teve 36 conceitos principais e 30 adicionais identificados, do total, 17 não foram definíveis pela CIF. Dos conceitos conectados dos 3 instrumentos 39 referem-se à Funções do Corpo, 57 à Atividades e Participação e 4 à Fatores Ambientais. O refinamento das regras propiciou mais clareza no processo de identificar, relacionar o conteúdo dos instrumentos à CIF e expor os resultado e aumentou o número de conceitos identificados e categorias contempladas pelos instrumentos. Palavras-chave CIF, Regras de ligação, Inquéritos e Questionários Abstract The Visual Analogue Scale (VAS), Roland Morris Disability Questionnaire (RMDQ), Short Form Health Survey (SF-36) are broadly used and had their content linked to ICF by the linking rules of 2002 and 2005. In 2016 were refined and were not applied yet. To apply the refinements of ICF linking rules to VAS, RMDQ, and SF-36. Two health professionals identified the meaningful concepts and linked to the most precise ICF categories and a third triggered in divergences. The degree of agreement was calculated by kappa statistic. There was almost perfect agreement (Kappa=0.93 p<0,001). The main concept of VAS was linked to ICF category b280, the 24 main concepts of RMDQ linked to b28013, and 27 additional linked to other categories. The SF-36 had 36 main concepts and 30 additional concepts identified which 27 were definable by the ICF and 17 do not. From the total of ICF linked concepts, 39 refer to Body Functions, 57 to Activities and Participation and 4 to Environmental Factors. The refinements of linking rules propitiated more clarity in the process to identify, to link instruments content with ICF and to expose the results. Thus, increased the number of identified and linked concepts as well as the categories in the instruments.


Introduction
The chronic pain has been reported with high prevalence and due to multiple factors [1][2][3] . Thus, it is necessary to consider the physical, psychosocial, and environmental factors involved in this health condition 4 .
A universally accepted framework that encompasses a biopsychosocial model is the International Classification of Functioning, Disability and Health (ICF), which was created by the World Health Organization (WHO) and standardizes the language on the health states of individuals referring to categories related to Body Functions, Body Structures, Activity and Participation, and Environmental Factors 5 .
The use of the classification is recommended in conjunction with clinical practice tools to valorize and to complement the information already collected, encouraging professionals to select more appropriate instruments for their clinical proposals [6][7][8] . These tools, however, have been used in a perspective strongly anchored in the biomedical model of health, even when they collect data regarding social participation and environment conditions 9 . This approach results in decontextualized practices focused on specific interventions and monitoring of signs and symptoms.
The ICF content, when connected to these instruments, will contribute to the operationalization of an expanded health concept, allowing the interpretation of the data in order to attribute the same status to the physiological components and contextual factors, considering privileging their interaction and interdependence, as product and producer health and its related states.
For better accuracy in the process is recommended to link both instruments and ICF content, making ICF a tool capable of translating the information contained in the instruments 10 . In view of this, in 2002 the first Linking Rules 11 were proposed, in 2005 updated 6 , and in 2016 refined 10 .
The refinement of the linking rules aimed to increase the transparency and reliability of the instrument content linking process proposed five improvements. The first concerns the preparation of information for linking. The ICF Linking Rules from 2002 and 2005 stressed that the meaningful concept had to be identified before starting the linking process. Thus, the process presents the risk of applying the linking rules mechanically and out of its context or purpose in which the information is collected.
The refinement proposes to identify the "purpose of information" to be linked to a category before identifying the meaningful concept, answering the questions: "is this the key piece of information? What is this item about?". Thus, the refinement of the rules proposes not only the identification of significant concept but let this be separated into main concept and additional concept according to purpose and perspective of information to be linked to the ICF 10 .
Another refinement is to take into account the perspective on the information collected. It proposes the documentation of the perspectives from which the information is collected, by the data collection instrument or by the collection mode. The most prominent perspectives are: descriptive, involving capacity and performance in performing a given task or activity; appraisal of individual satisfaction in relation to a given situation, asking himself to what extent personal expectations and hopes have been achieved; and, the perspective of need or dependency, which refers how many assistance devices are needed to perform certain activities or tasks.
In the 2002 and 2005 linking rules there was no information on the categorization of the response options captured in the connection process, in addition to the meaningful concepts 6,11 . The refinement proposes to identify and document the categorizations of response options, such as: intensity, frequency, duration, confirmation or agreement, and qualitative attributes. Given that this is relevant only to questionnaires, assessments, or tests that contain response options. It is worth mentioning that there are response options where the link is not possible or where the answer option also contains meaningful concepts 10 .
The rules proposed in 2002 and updated in 2005 recommend that the qualifiers 8 (unspecified) and 9 (not applicable) do not be used 6,11 . The use of these qualifiers is recommended in the refinement, as experience has shown that not using this categorization leads to loss of information. If a concept is linked to a qualification 8 or 9, additional information not specified or not covered by the ICF should be documented together with the category of the ICF 10 .
The last propose guides that information not contained in the ICF should be attributed to the abbreviation nc (not covered), although the linking process can be applied to any type health information, not always feasible to link information to an ICF category. This may be the case as the information to be linked is beyond the scope of the ICF or may be very specific to be linked to the ICF 10 .
Many instruments are used in research and clinical practice to evaluate the functionality, quality of life and pain. Those that appear most in the literature are the Visual Analogue Scale (VAS) 12 , Roland Morris Disability Questionnaire (RMDQ) 13 and Short Form Health Survey (SF-36) 14 , used to assess patients' pain, functional capacity, and quality of life, respectively. These instruments have already had their content linked to the ICF by studies that applied the linking rules of 2002 and 2005, such as Prodinger et al. 15 , Sigl et al. 16 , Fréz et al. 17 , Schepers et al. 18 , Geyh et al. 19 and Cieza and Stucki 7 .
Cieza et al. 10 suggest that researchers in the field apply the updated ICF linking rules and comment about the experiences during the utilization process. It is in this sense that the present study aims to contribute. To date, no studies have been found in the literature conducted with the application of rules refinement.
Considering the magnitude of the qualitative changes brought by the new linking rules, and the relevance degree of the instruments adopted in this study in the clinical practice of health professionals, the authors consider it important to apply the new rules to them.
Therefore, the objective of the present study was to apply the refinements of ICF linking rules to VAS, RMDQ, and SF-36 proposed in the literature.

Method
The process of linking the content of the questionnaires occurred through the application of the refinement of the linking rules proposed by Cieza et al. 10 . Two health professionals who participated in this process undertook a course on ICF certified by the WHO collaborating center of Brazil and had experience in the linking process with the previous rules. They were also included in a study group on ICF including the classification by the tool available and recommended by the WHO, the e-learning tool. A third physiotherapist researcher with the same background was triggered in cases of divergence.
According to the methodological guideline 10 , the purpose of the information to be linked was identified before the meaningful concept identification. Thus, the two health professionals identified the meaningful concepts contained in the instruments considering the context, statements, and response options to select any other additional concept. Then, the meaningful concepts were compared to obtain a consensus.
Another update made in the present study was to identify and describe the perspectives adopted in the questionnaires. Questions were identified and described as descriptive, appraisal, and need or dependency. The categorizations of the response options were also identified and documented as intensity, frequency, duration, confirmation or agreement, and qualitative attributes.
In sequence, each meaningful concept was linked to the most precise and specific ICF category. In cases of a relationship between the concepts by conjunctions, they were also recorded. The meaningful concepts related to physical health, mental health, health in general, disability in general, functioning, and child's development that did not provide sufficient information for selecting an ICF category were classified as not definable and assigned as nd-ph, nd-mh, nd-gh, nd-dis, nd-func, and nd-dev, respectively. When a meaningful concept was identified as a personal factor, it was assigned as pf, and, when the meaningful concept was not covered by ICF, it was assigned as nc (not covered), such as diagnosis or health condition (nc-hc) and quality of life in general (nc-qol) 10 . These situations are called special cases.
The process of linking content was organized in Microsoft Excel 2007. The degree of agreement between the two health professionals regarding the identified and linked concepts was calculated by means of the Kappa statistic 20 . All the pre-determined criteria for verifying inter-rater concordance by Cohen's Kappa coefficient were followed in this study, ie the variable was nominal in nature (ICF categories) and the comparison was made only between two examiners and there was no missing data [21][22][23][24][25][26] . In addition, it is the test recommended in the rules of connection and has been frequently reported in the literature as the chosen one to evaluate the degree of agreement between the professionals in the identification of significant concepts and in the connection of these to the categories of the ICF 6,10,16,[27][28][29] . Kappa values generally range from 0 to 1 with 1 indicating perfect agreement and 0 indicating no additional agreement beyond what is expected by chance alone. Kappa coefficients above 0.61 are regarded as good 18 . The data analysis was performed with SPSS 22.0.

Results
For the presentation of the results, the model proposed by the methodological guideline was used. A table was generated containing the information extracted from the questionnaires together with the results of the linking process, valuing the perspective adopted in the information, the classification of the response options, and the categories from the main concepts and additional concepts. The results of the instrument content linking are described in Charts 1 and 2.
There was almost perfect agreement among the researchers who linked the items of the RMQD and SF-36 and the ICF codes (Kappa In-dex=0.93 p<0,001). Table 1 shows the 10 meaningful concepts of the SF-36 and 1 of the RMDQ that were linked to the special cases. Table 2 shows the total number of concepts identified in ICF instruments and categories linked to these concepts. Regarding the RMDQ, it was considered that all items were guided by the concept of back pain, which was considered the main concept of all items and was linked to a category (b28013). In 23 of the 24 RMDQ items, it was possible to identify 27 additional concepts, most of them linked to Activities and Participation categories followed by Body Functions and Environmental Factors (Chart 1 and Table 2).
In relation to the SF-36, it was observed that, of the 36 main concepts, 13 were not definable by the ICF, and, of the 30 additional concepts identified, 4 were not definable (Table 1). This fact shows that the ICF does not contemplate comprehensive concepts regarding general health, physical health, mental health, and health conditions.
Questions 4 and 5 of the SF-36 brought relevant information from the heading, and, from this information, a main concept and additional concepts for each question were identified. Each item of these questions allowed the identification of additional information.
These results enable a comparation between the application of ICF Linking Rules to 2002, 2005 and 2016 when applied to the same instrument which allows further discussion. Table 3 presents the numerical results of the linking to the three instruments with ICF of the current study and other previous studies. The refinements of linking rules applied in the present study propitiated an increase of identified and linked concepts to ICF to SF-36. For both SF-36 and RMDQ there was an increase in the number of categories. For VAS, the results were the same in the present study and for Scheuringer et al. 30 .
In order to increase the transparency and reliability during the link process we followed the suggestion to put results in a table containing the information extracted from the questionnaires together with the results of the linking process, valuing the perspective adopted in the information, the classification of the response options, and the categories from the main concepts and additional concepts.

Discussion
The study aimed to report the experience of applying the linking rules refinements proposed in the literature. This proposal 10 , which suggests a reflection before the identification of concepts, has guided us to the identification of a main concept related to each item of the instruments and to the identification of additional concepts and additional information relating to items.
The previous rules "a" and "e" 6 emphasized that the information should be identified before starting the linking process. However, there was a risk of information being extracted mechanically without considering the context or the proposal of information collected 10 . In such a case, information could be masked going on identification of categories without definition of the most representative of the questionnaire items. The recent rules propose the identification of the information to be linked before the meaningful and additional concepts identification 10 .
After the application of rule two 10 , to identify the main concept, rule three was applied to identify any additional concept to the main information. These concepts could also be identified in the answer options. The organization of the main and additional concepts will allow comparison of health information in a more specific way. An example to illustrate this occurred in item 3 of the SF-36 when linked with the ICF by Cieza et al. 11 and linked by the current study. Although the same meaningful concept (Rigorous Activities, Running, Lifting Heavy Objects and Participating in Strenuous Sports) have been identified in both studies, in the study of Cieza et al. 11 , they were linked independently to the categories of the ICF, the Rigorous Activities concept was regarded as not definable, and the remaining concepts were linked to the categories from Activities and Participation. The present study identified rigorous activities as the main concept and the others as  it continues which refers to the type and level of need that the individual requires. In addition, it proposes that the items identified as descriptive perspective be distinguished in capacity and performance 10 .
intensity, frequency, duration, confirmation or agreement, and qualitative attributes 10 . One area that may benefit from this categorization is clinical practice, which may direct the choice of an instrument, now with respect not only to its content but also to the characteristics of the answers that best suit its purposes.
In the previous linking rules 6,11 , it was recommended that the results table should include significant concepts, ICF categories linked to concepts and information additional to meaningful concepts. However, the ICF linking rules refinement recommended that the table with the results should contain the name of the instrument or other identifier, verbatim health information, perspective adopted in the information, response options, classification of response options, main concept, additional concepts contained in the information, ICF category of the main concept, and ICF category of other concepts and annotation. The annotation item, proposed in the refinement, corresponds to the additional information of the 2005 rules 10 .
As for the structuring of the RMDQ and SF-36 instruments, it can be seen that while, in the SF-36, the intensity and frequency of the difficulties are diluted in the responses, in the RMDQ, this is contained within the items. Therefore, the additional information contained in the content link annotations of the RMDQ items refers to the intensity and frequency of the difficulties while the SF-36 additional information refers to examples and specifications of the items.

Superscripition Item
Perspectives adopted in information Be healthy nd-gh As anybody I know 11c Health nd-gh To get worse 11d Health nd-gh Excellent The question 2 of SF-36 has an annotation: "compared to one year ago". nc: not covered by ICF; nd: not definable by ICF; nd-gh: not definable -health in general; nd-ph: not definable -physical health; nd-mh: not definable -mental health. Source: The authors. In the present study, the VAS perspective was identified as evaluation, that for the RMDQ as descriptive of performance in 22 items and dependence in the 2 others, and that for the SF-36 as descriptive of performance in 25 items and evaluation in 11.
Regarding rule five, the categorization of the instrument response options can be described as The proposal to link three instruments used in research and clinical practice for the assessment of chronic pain was strengthened with the application of the linking rules refinements, giving greater clarity to the results expressed in Charts 1 and 2 with the columns of main significant concepts, additional concepts, and categories linked to them.
Observing previous studies and the present study, there was a gradual increase in both the number of concepts identified and categories linked throughout the publications to SF-36 instrument 6,11,15,[17][18][19] (Table 3) and to RMDQ 16 . It is believed that the identification of more meaningful concepts in the present study occurred due to the consideration of the information in the headings and answers of the questions. The number of categories linked in this study compared with the number found in previous studies had an increase, but it was not as expressive. This observation corroborates the refinement proposal objective to improve data capture and information nuances by improving data exposure.
In analyzing studies prior to refinement, which link the SF-36 instrument 6,11,15,[17][18][19]30 (Table  3) and RMDQ 16 , it was observed that only Cieza et al. 11 exposed the significant concepts identified in all instruments and the linking of all items with the ICF. Cieza et al 6 , Geyh at al. 19 , Schepers et al. 18 , Prodinger et al. 15 and Sigl et al. 16 showed the selected categories but did not show which instrument item they were linked to. Scheuringer et al. 30 linked the content of 120 instruments, among them the SF-36, but the results expressed the total of categories related to all the instruments. Fréz et al. 17 linked only the SF-36 domains to the ICF categories.
It is believed that the detail in the exposure of the results is related to the purpose of the study and to the number of related instruments the higher the number of instruments connected the less the possibility of detailing.
Rat et al. 31 and Milman et al. 32 use the already established SF-36 links to compare content with other instruments. Faria et al. 33 carried out a systematic review to identify categories of the Activities and Participation component in some quality-of-life instruments already linked to the ICF. This shows the importance of linking these questionnaires using rule refinement so that it can also serve as a basis of comparison for the establishment of other instruments.
Since the objective of this work was to report the experience of applying the refinement of the linking rules, the study presents the detailed results as suggested by Cieza et al. 10 , allowing users to access information on the content of these measuring instruments, which can assist clinicians and researchers in choosing the instruments that best suit their interests. This situation can be explained with the VAS. In the present study the category b280 was choose but depending on the context of the research a more specific category can be chosen related to pain, such as low back pain, the category b28013. Some possible limitations of the present study should be considered, even if important information has been obtained to the operationalization of the refinement of the linking rules and its dissemination. Differences were observed in the general findings of studies that used the same instruments using the same rules. This situation shows that the experience and clinical context should be considered when selecting the version that best meets your needs. It is also necessary that other professionals use the new linking rules in different clinical contexts to identify your contribution to the production of records of functionality and monitoring of health status.
A limitation for the discussion of these results was the lack of clear identification in the studies of which concepts were identified and their respective items in the instruments investigated. In such cases, those interested in using the studies as a reference must go the other way to identify what has been linked.
Among the studies included in this study, only Cieza et al. 11 presented clarity in the presentation of the findings, so it is believed that the recent insertion of a table for exposure of the results will help future works as well as the registration of the constructs capacity and performance recently considered. In this way, the professional/researcher that will use the instruments may know the items with association to the categories of the ICF, its specification or content.
It is known that international efforts have been applied to associate instruments validated with ICF categories. Cieza et al. 10 suggest that the continued use of the updated rules for linking instruments to the ICF may assist in the creation of a database containing sets of items linked to each ICF category, aiming at both the development of new instruments and the operationalization of certain ICF categories and instruments according to the perspective that best suits the clinical purpose. This exercise will also provide information to better specify the categories relevant to a future classification review.
The ICF use equips professionals in clinical practice as well as programs and public policies'   8 . Its use could also help to select and to build instruments more precise in content and in a quantitative source of information 10 . Thus, it is important that more studies like this are carried out to expand the databases and evaluate the influence of the refined linking rules on the process of linking instruments with different perspectives and different categories of responses.

Conclusion
The refinements of linking rules propitiated more clarity in the process to identify, to link instruments content with ICF and to expose the results. Thus, increased the number of identified and linked concepts as well as the categories in the instruments.
The refinement of the linking rules allows the standardization of the concepts of assessment instruments and the ICF with different levels of precision, with specification/indication of the perspective and the categorization of the response. The results expressed in a table with the main information about the linking process will help to increase the transparency and reliability as well as the comparison within studies.
The content linking of diverse instruments, such as the VAS, RMDQ, and SF-36 based on the refinement of the linking rules, will contribute to the alignment of the existing instruments with the ICF and with the selection of instruments that cover relevant parameters of interest to researchers and clinicians.