The technological influence on health professionals' care: translation and adaptation of scales1

Objectives: in this study, two research tools were validated to study the impact of technological influence on health professionals' care practice. Methods: the following methodological steps were taken: bibliographic review, selection of the scales, translation and cultural adaptation and analysis of psychometric properties. Results: the psychometric properties of the scale were assessed based on its application to a sample of 341 individuals (nurses, physicians, final-year nursing and medical students). The validity, reliability and internal consistency were tested. Two scales were found: Caring Attributes Questionnaire (adapted) with a Cronbach's Alpha coefficient of 0.647 and the Technological Influence Questionnaire (adapted) with an Alpha coefficient of 0.777. Conclusions: the scales are easy to apply and reveal reliable psychometric properties, an additional quality as they permit generalized studies on a theme as important as the impact of technological influence in health care.


Introduction
The contemporary world is clearly technological, entailing advances in many of the areas that affect human life. In that sense, health care may represent the area where these advances are more visible and provoke greater expectations. Nevertheless, this technological progress "has aroused different concerns and questions on the benefits, risks and the relations constructed among workers, patients and technology use" (1) . Some even appoint that, together with this scientific progress, the dissatisfaction with health professionals seems to increase. At the same time as the scientific knowledge and new treatment and diagnostic techniques evolve, the patients' dissatisfaction with health care grows, which seems to point towards difficulties to achieve a harmonious relation between scientific progress and the prioritization of what is human in health care.
In several countries, today, professional organizations elaborate large-scale opinions and studies to defend the users and the quality of health care, with outcomes that are highly critical of health professionals and seem to want to alert to the fact that these workers have lost their ability to "take care" (2)(3)(4)(5) .
One of the most appointed aspects in the bibliography is the excessive technicality of man's action in the 21 st century, or the extreme rationalization of the contemporary technical civilization which, according to Silva and Ferreira (6) , exerts cultural and social control on human beings, sometimes leading to the rational automatism that replaces individual and group decision making. This attitude, associated with the strong influence of the so-called "biomedical model", strongly in vogue in the health care of the past centuries, may have lead to a "mix-up" of values which, according to the authors, managed to deviate the health practices from their core objective, which is the human being.
When centering almost exclusively on the diagnosis, disease and treatment forms, the health professionals often leave ill people in the hands of a depersonalizing solitude, a fact that irreparably impairs the quality of health care, affecting it exactly in one of its paramount characteristics, which is the therapeutic relationship or the relationship patient/health professional. In 1997, in one study (7) , the participants (14 nurses) displayed a positive view on the benefits of technology and their trust in the potential the "machine" offered. In addition, in the study Describing the Influence of Technologies on Registered Nurses' Work (8) , In that sense, the study The effect of technology on the caring attributes of an international samples of nurses (9) is particularly noteworthy, as it raises the possibility of a new approach to this problem, which had thus far only been discussed based on phenomenological or qualitative methods.
In view of these findings, the translation, adaptation and validation of two data collection tools -TIQ (Technological Influence Questionnaire) and CAQ (Caring Attributes Questionnaire), seems to be very useful to start a study on this theme. Although these tools have only been applied to Nursing professionals, according to the authors, as the technological influence and its possible effects on care will certainly affect all health care providers, it would be interesting to apply those tools to various professional groups in that are.
Nevertheless, due to difficulties to get access to some professional groups for sampling purposes, according to established tool validation rules, in this study, the researchers chose to apply the tools to nursing professionals, medical professionals, nursing students and medical students only was chosen. Therefore, some items in the scales were slightly adapted.

Method Caring Attributes Questionnaire (CAQ)
The introduction of scales that permit the quantitative assessment of caring attributes is very recent. Great efforts in that sense only started to be made as from the 1980's. The Caring Attributes Questionnaire (CAQ) (9) has been used in different countries on distinct continents, with very solid psychometric characteristics,

Technological Influence Questionnaire (TIQ)
The Technological Influence Questionnaire (9) is a single-factor tool that consists of 14 items, using an agreement scale.
The CAQ and TIQ are applied individually and, although originally directed at Nursing professionals, adaptations were made in this study to be applicable not only to nursing professionals, but also to medical professionals and nursing students and medical students.

Scale translation and validation procedures
The translation and adaptation of scales require strict procedures that go far beyond simple translation.
The cultural contexts need to be heeded, whether of the original culture or the target culture of the test, thus implying not only the translation, but a global adaptation to the new situation. Thus, the goal is for the test to similarly measure the original construct, even if that demands adjustments to the particularities of the study population (10)(11) .
In line with research experts' orientations, the CAQ and TIQ were translated to Portuguese in five steps.
In the first, two bilingual professionals translated the scale from the original language, in this case English, to Portuguese. These experts were asked to use simple language but to, beyond a literal translation, attempt to capture the meaning of the different items. After the individual translations, they were asked to analyze both translations and solve the discrepancies found in order to elaborate a single document.
After this phase, two other bilingual experts elaborated the back translation (11) , then comparing the results. Next, two experts fluent in English developed an independent review. They were knowledgeable on the study objectives and the target population and were asked to compare the back-translated version (in English) with the original scale.
To solve possible difficulties in the understanding of some of the items, a pretest was applied to 12 individuals from the health area (nursing and medical professionals and students).

Validation procedures and criteria of the CAQ and TIQ scales
To assess the psychometric properties of the scales and analyze the results of their application, maintaining the author's method was considered to be most correct to make it easier to compare the results. Hence, after inverting the scores of items that were formulated in the opposite sense, validity and reliability tests were applied, based on a set of criteria that follow the best practices. Thus, the data resulting from exploratory factor analysis were crossed with Cronbach's alpha and item-item and item-total correlation coefficients. In that sense, the following criteria were set (12) .
-For the factor analysis, the principal component extraction method was used, adopting four main components to respect the structural organization of the original scale, followed by the rotation of the factors to obtain a clearer and more objective factor solution, thus maximizing the factor loadings of the items (12) . Like the scale author, the researchers chose the Varimax rotation

method.
To determine what factors and items to retain, different authors' recommendations were followed (12)(13)(14)(15)  If that happens and if the difference between them is not equal to or higher than 0.15, the elimination of the item should be considered; d) the percentage of the variance the retained factors explain should be at least 40% and e) each factor can contain no less than three items.
To complement the reliability analysis, 0.60 was set as the minimal internal consistency ratio (Cronbach's alpha), the item/item-total correlation should not be lower than 0.3 and the internal consistency of the factor should not increase if the item were eliminated.

Results
Considering the preset inclusion criteria and to reduce the universe of care providers to be included in the target population, stratified sampling was chosen to obtain a representative sample, according to some pre-identified variables of the study population and nonprobabilistic convenience -snowball -sampling. Hence, On the other hand, even if not that reliable, the significance of Bartlett's sphericity test (coefficients associated with p<0.05) shows that the variables can be correlated (13) . Table 1 shows the results of an exploratory analysis using the principal component extraction method, revealing four principal components, aiming to respect the structural organization of the original scale, followed by the Varimax rotation method. As a whole, the intended four-factor organization justifies 41.685% of the total variance (factor 1-13.68; factor 2-11.121; factor 3-8.683 and factor 4-8.198).     Concerning the item/total correlation coefficients obtained, all coefficients comply with the preset criteria.   (13) . Thus, the distribution of the items among the three factors obtained through the principal component analysis (Table 3) differs from the author's original idea but deserves more careful analysis. Hence, factor 1 emerges with six items (4, 6, 7, 8, 9 and 11), all of which are formulated in the negative form and require inverted scores. The saturation levels range between 0.390 for item 6 ("due to using more technology, the health professionals feel frustrated when a patient dies") and 0.760 for item 9 ("I am in doubt about the benefits of the technology for my (future) profession").

Reliability and internal consistency analysis
In this study, factor 2 emerges with four items (10, 12, 13 and 14), formulated in the positive sense to value the health technology. These items display high saturation levels, superior to 0.60.

Reliability and internal consistency analysis
As mentioned in the validity criteria for the reliability and internal consistency analyses, in this study, Cronbach's alpha coefficients were analyzed and Pearson's correlation coefficients between each item and the scale it belongs to were determined. Only item/total correlations with r>0.3 were considered.
From a careful analysis of the items, it was verified that item 6 not only presents a low communality value (0.220), but that the r-coefficient is inferior to the intended 0.3. Therefore, the item was excluded, which even increases the alpha coefficient.
For factor 1, for which the scale designation "negative influence" will be used, which now consists of five items, the alpha coefficient was 0.80, which is considered good. As for the item/total correlation, all coefficients are superior to 0.5.
For factor 2, for which the scale designation "positive influence" was maintained, in the factor analysis, a fouritem structure was obtained. The detailed analysis of each item's behavior showed that all items comply with the preset criteria (   Figure 1).
In addition, as opposed to the original scale with a onedimensional structure, in this study, an organization in two dimensions was found, with an apparent separation between positive and negative aspects. Once again, the cultural differences and the now more heterogeneous population seem to justify the differences found. These changes resulted in a Cronbach's alpha coefficient of 0.777 for the total scale, higher than the coefficient of the original scale (0.75).

Escala dos Atributos do Cuidar (adaptada)
Advocacy Estou a cuidar quando trato a informação do doente de forma confidencial  Questionnaire started to construct a tool to study the relation between these two variables. After analyzing the results, it seems that this objective was positively achieved, as the scales demonstrate very reasonable psychometric properties. The analysis by the experts and the groups used shows that the tool is easy to understand and complete. The only less positive aspect was the loss of some original items, probably due to the heterogeneity of the selected population.
The validation of these scales entails the possibility of further studies on an essential aspect of health care, which is the relation between the technological influence and the health professionals' care. As mentioned in the introduction, different theoretical theses exist that appoint the technological influence as something negative for care delivery, but the few scientific studies that exist point in the opposite direction. Therefore, the validation of these tools and their general application will permit a deeper look into this theme.