Cultural adaptation, content validity and inter-rater reliability of the "STAR Skin Tear Classification System"1

AIMS: to perform the cultural adaptation of the STAR Skin Tear Classification System into the Portuguese language and to test the content validity and inter-rater reliability of the adapted version. METHODS: methodological study with a quantitative approach. The cultural adaptation was developed in three phases: translation, evaluation by a committee of judges and back-translation. The instrument was tested regarding content validity and inter-rater reliability. RESULTS: the adapted version obtained a regular level of concordance when it was applied by nurses using photographs of friction injuries. Regarding its application in clinical practice, the adapted version obtained a moderate and statistically significant level of concordance. CONCLUSION: the study tested the content validity and inter-rater reliability of the version adapted into the Portuguese language. Its inclusion in clinical practice will enable the correct identification of this type of injury, as well as the implementation of protocols for the prevention and treatment of friction injuries.


Introduction
Skin tear, as it is known internationally, is a traumatic wound resulting from friction or friction and shear, leading to the separation of the epidermis from the dermis or the separation of both from the underlying structures (1) . Fragile skin is commonly associated with frail and dependent people, or those with impaired mobility and nutrition, such as elderly people (1) , those in the terminal phase of life (2) and newborns (3) .
The theme is little known in Brazil, where skin tears were called lacerations (4) . Although this seems to be mere semantics, the fact of not having it own nomenclature is an important barrier to the implementation of specific preventive measures, as well as appropriate treatment techniques. To use standardized language and a reliable classification instrument are fundamental for evaluating the wound and planning the care, as well as being essential for the development of studies and the use of their results in clinical practice (5) .
Payne and Martin were the first to propose the skin tear nomenclature and classification system (1) .
Although this classification is the most widely used, its measurement properties have not been demonstrated (1) .
As there was no universally accepted nomenclature or classification instrument (6) Carville et al. (7) redesigned the Payne and Martin instrument in the light of evidencebased practice. The result was the STAR Skin Tear Classification System (STAR), a simple and easy to use, yet comprehensive instrument, with confirmed content validity and inter-rater reliability, and standardized terms and definitions (7) .
The STAR consists of a treatment guide, classification system and glossary. The treatment guide has six topics related to the care of the wound and the surrounding skin. The classification system evaluates the presence/ absence of the skin flap and its viability. It contains five photographs, each relating to a description of a skin tear category. Finally, on the back of the instrument there is a glossary that provides the tear skin definitions and technical terms related to the theme (8)(9) .
Aiming for the systematization of the knowledge concerning this type of injury, the objective of this study was to carry out the cultural adaptation of the STAR (treatment guide, classification system and glossary) for the Portuguese language of Brazil, as well as to test its content validity and inter-rater reliability.

Methods
To perform the cultural adaptation and test the content validity and inter-rater reliability of the STAR, authorization was sought and received from Prof. This study is a methodological study with a quantitative approach, in which the cultural adaptation was carried out and the measurement properties of content validity and inter-rater reliability were tested (10-11) .
-The cultural adaptation was developed in three phases: Translation: conversion of the STAR into Portuguese.
Two translations of the STAR into Portuguese were performed: One by a Brazilian connected to the health area and another by a Brazilian layman. Both translators were fluent in English.
-Evaluation by a committee of judges: content validation of the Portuguese version(10). The committee of judges was composed of six Brazilian nurses (stomatherapy or dermatology specialists), fluent in English and knowledgeable of the concepts to be analyzed. Based on the analysis of semantic, idiomatic, cultural and conceptual equivalence (11), the judges initially analyzed the instrument as a whole, determining its scope, that is, whether each concept was adequately covered by the set of items and all dimensions were included. They also analyzed the individual items of the instrument, checking their clarity and relevance. Regarding clarity, the judges reviewed the wording of the items and whether they were written so that the concepts involved were understandable and adequately expressed. Regarding relevance or representativity, they evaluated whether the items really reflected the concepts involved, whether they were relevant and whether they were adequate to achieve the aims proposed. The judges also evaluated the wording of the items to ensure that the concepts could be understood by users and whether they reflected those originally proposed, based on knowledge of the pathophysiology of the injury, always having the freedom to suggest adjustments and improvements needed in each item. After receipt of the analysis of the judges, Strazzieri-Pulido KC, Santos VLCG, Carville K. the authors examined the equivalence between them and the controversial aspects were discussed in order to reach a consensus. Concordance of 80% was adopted for the equivalence analysis between the evaluations.
The researchers discussed the controversial aspects of these evaluations and obtained a Portuguese version, with validated content, which was back-translated.  For the analysis of the socio-demographic and clinical data, measures of central tendency were used. The inter-rater reliability was evaluated by weighted kappa (wk). Negative values represent concordance lower than that expected due to chance or inconsistency of the test (12)(13) . The concordance is excellent if wk>0.8 and poor if wk<0.2 (14) . The assignment of the values related to the weights is subjective and depends on the context (13) . Figure 1 presents the interpretation of the distribution of the wk according to the level of concordance. Tests with a significance level less than 5% (p <0.05) were considered statistically significant.

Results
The results are presented according to the steps outlined in the methods.

Cultural adaptation
Translation Two Portuguese versions of the STAR were obtained.

Photographic set
Considering the five categories of friction injuries, each classified by 107 nurses, 535 observations were obtained.      The frequency of correct responses was higher than that of the errors in categories 1a (3/60.0%), 2b (4/80.0%) and 3 (6/60.0%) and, as in the application with the photographic set, categories 1b and 2a presented higher frequencies of errors (3/60.0% and 4/80.0%). Table 4 presents the levels of concordance between the responses of the nurses and the gold standard. The results in Table 4

Discussion
In the cultural adaptation, the difficulty that the All these considerations led the researchers to consider that the content validity of the instrument adapted for Brazil was confirmed.
The inter-rater reliability was analyzed through two strategies: photographic, as in the original study (7) and clinical (11) .
The worse performance of the instrument with the photographs can be attributed to the absence of a photographic set for comparison. The nurses correlated the five photographs of the instrument to their corresponding categories, therefore, there were no other photographs as the basis for comparison, unlike the original study by Carville et al. (7) . For the clinical application of the instrument, however, the nurses could associate the photographs of the adapted instrument with the wounds found on the patients. This detail can provide an advantage in the use of the instrument, expanding the explanations for the differences in the statistical analysis.
In the application of the original instrument, performed by Carville et al. (7) , 26 trained nurses classified 25 photographs of friction injuries. The concordance levels were below 65.0% for categories 1b, 2a and 2b.
One reason for the low concordance was the lower quality of the photographs of these categories compared to the photographs of category 1 and 3. The application of the instrument exclusively with photographs constituted a limitation of the original study (7) .
In the present study, in addition to the absence of a photographic set for comparison, a negative impact due to the quality of the photographs was also observed.
However, since the assembly of a photography database for comparison would require another validation work, not only of "content" (images) but also of the equivalence to the original photographs, which would constitute another study, it was decided to also perform the application of the instrument in the clinical practice.
It was considered that, for the evaluation of wounds, the in vivo model provides the advantage of a three- For the construction of the original instrument, the concordance obtained for each friction injury category was much higher than in the present study, in both applications (photographic and clinical), ranging from 83.0% to 97.0%, with 93.0% for the photographic set (7) . Only category 1b obtained concordance below 90.0%, with 85.0%.
In the present study, category 1b did not present good performance in either application. The percentage of correct responses was only 34.6% in the application with photographs and slightly better (40.0%) in the clinical practice. In the version of the instrument published in 2007 (8) and used in this study, this category was represented by an injury with non-invasive sutures. As well as the low quality of this photograph compromising the results obtained in the STAR project, this probably also compromised the concordance obtained here.
Regarding the difference between the results of both studies, another aspect to be considered is related to the fact that the original authors (7) verified the inter- In the first version of the original instrument, the concordance obtained for category 3 was above 90.0% and one hypothesis is that the evaluation of the skin flap is more difficult than that of the wound bed (7) .
It can be said that the moderate levels of concordance obtained (wk = 0.596) in relation to the clinical application of the Portuguese version, confirmed the inter-rater reliability.
The cultural adaptation of the STAR and the performance of tests to validate its contents and interrater reliability are the first initiatives to arouse the interest of Brazilian healthcare professionals regarding the problem involving friction injuries. It is hoped that this study will inspire other initiatives in order to identify environments, situations and people at risk, as well as risk factors, incidence and prevalence of these injuries, characteristic to Brazil. In addition, the inclusion of the evaluation of these types of wounds in daily clinical practice will enable the implementation of protocols for their prevention and adequate treatment. Visualizing the future, it is expected that friction injuries will receive the same attention, dedication and care that is currently given to pressure ulcers, and that this will also be seen as an indicator of the quality of the healthcare services.
Regarding the present study, although two measurement properties of the instrument were confirmed, the validation process does not end here.
Validation is an ongoing process. The more evidence that can be gathered related to the instrument measuring what it purports to measure, the greater the confidence in the results of its use (11) .
Other limitations in addition to those mentioned here should be highlighted. It was not possible to make the fragility of the skin, also present at the other extreme of age, at the end of life and in some conditions such as cachexia, characterize it as the preponderant factor (2) .

Conclusion
This study led to the conclusion that the