Psychometric Properties of the Patient Advocacy Scale for Intensive Care Nurses

Patient advocacy exercised by an intensive care nurse refers to the defense of the patient’s interests and rights and can be measured if, for that, there are valid, reliable and reliable instruments, that is, which have their psychometric properties proven. Therefore, the patient’s advocacy processes would be permanently monitored and evaluated, also, the practice would become more visible. Perform psychometric validation of the Patient Advocacy Scale for Intensive Care Nurses (EAPEnf-ICU). Instrumental research for instrument validation, with exploratory and confirmatory factor analysis. Carried out from January to June 2021 with 377 Brazilian intensive care nurses, selected by non-probabilistic convenience sampling. Data were collected by means of Google Forms, organized on Excel® 2010 software and analyzed on R software. The Ethics Committee approved the study. Participants received information about the research, agreed to respond to the questionnaire and were guaranteed anonymity. From the scale structural exploration by analyzing exploratory, confirmatory and internal consistency of the measurement instrument, the final version of the EAPEnf-ICU was composed of 54 items, distributed in 5 dimensions/factors: Factor 1—Clinical and organizational advocacy in intensive care; Factor 2—Barriers associated with the intensive care clinical and organizational complexity; Factor 3—Attitudes to promote the autonomy of patients and family members in intensive care; Factor 4—Barriers associated with divergences and ethical-professional limits in intensive care, and Factor 5—Intensive care nurse’s personal and professional background. Findings point that the instrument presented in this study is valid and reliable for assessing the aforementioned construct, as it presents theoretical and empirical consistency, identifying five dimensions related to the exercise of patient advocacy by the intensive care nurse. Plain Language Summary Patient Advocacy Scale for intensive care nurses The patient advocacy, exercised by an intensive care nurse, refers to the defense of your interests and rights and can be measured if, there are valid and reliable instruments. The aim of this study was validation of the Patient Advocacy Scale for Intensive Care Nurses (EAPEnf-ICU). Methodological study for instrument validation carried out from January to June 2021 with 377 Brazilian intensive care nurses. Data were collected by means of Google Forms, organized on Excel® 2010 software and analyzed on R software. The Ethics Committee approved the study. The final version of the EAPEnf-ICU was composed of 54 items, distributed in 5 dimensions/factors. Factor 1 “Clinical and organizational advocacy in intensive care”; Factor 2 “Barriers associated with the intensive care clinical and organizational complexity”; Factor 3 “Attitudes to promote the autonomy of patients and family members in intensive care”; Factor 4 “Barriers associated with divergences and ethical-professional limits in intensive care”; Factor 5 “Intensive care nurse’s personal and professional background.” Findings point that the instrument presented in this study is valid and reliable for assessing the patient advocacy in Intensive Care, by the intensive care nurse.


Introduction
In the Intensive Care Unit (ICU) environment, the nurse plays numerous roles, such as direct patient care, interaction with family members, and care management.Considering the specific ICU characteristics, it is seen by patients and their families as a hostile environment, but which presents high technological power, necessary in view of the hospitalized patients' high-severity clinical conditions and risk of death (Manoel et al., 2022;Nascimento et al., 2021).
To act effectively in this unit, nurses need autonomy to perform their duties, and should have professional skills and competencies such as: leadership, technical-scientific knowledge, teamwork, decision-making, care/human resources and materials management, besides communication, continuing/permanent education, and humanization.Professional performance based on these requirements has an impact on care quality and leads to profession strengthening (Jesus Moraes & Rodrigues, 2021).
Based on these skills, nurses will be able to analyze different situations in their daily care, including those that can harm the patient, which involves their ethical-moral sense, in order to play a fundamental role as advocates/defenders.The process in which the patient is given support in the health arena is called health advocacy, topic on which this study is based (Luz et al., 2019;Manoel et al., 2022).
Advocacy has a number of definitions, depending on the place and context in which it is used.In nursing, it can be generally defined as initiatives related to intervention on behalf of the patient, protecting their rights, and protection and comfort to patients who cannot communicate.The several interpretations can make it difficult for the nursing team and the multidisciplinary team to understand the advocate's role (Abbasinia et al., 2019;Neves et al., 2021).
This study is based on the definition proposed by Hanks (2010), which points out that patient advocacy exercised by the nurse aims to help patients to receive the necessary health care, defend their rights, guarantee care quality, and serve as a link between patients and healthcare environment.In this sense, health advocacy is mainly associated with the recognition, by nurses, of their role in defending the patients' interests and rights, considering their beliefs and actions in relation to the care they provide to patients (Menezes et al., 2021;C. P. Vargas et al., 2022).
The exercise of patient advocacy by intensive care nurses has recently been addressed in the scientific literature.However, a recent integrative review study on the subject, which covered the scientific production in the period from 2010 to 2020, has found only 18 articles dealing with the topic, which highlights the need to advance the discussion on the matter (Menezes et al., 2021).Also, results pointed to the lack of valid and reliable instruments-national or international-that assess the individual differences of intensive care nurses regarding motivation to defense/advocacy for the patient's interests in the ICU context (Menezes et al., 2021).
It is essential to highlight this type of gap when proposing and validating new instruments, as it was the case in this study.Then, with the purpose of evaluating the exercise of patient advocacy by intensive care nurses, the Patient Advocacy Scale for Intensive Care Nurses (EAPEnf-ICU) (M.Vargas et al., 2022) was created.However, for a scale to be considered valid and reliable, it needs to go through different stages, including psychometric validation (Oliveira et al., 2018).
The validation process consists of the procedures for developing the items that compose the instrument, as well as the stages of statistical analysis of the data obtained by applying the instrument to a target population sample.In this sense, this study intends to answer the following research question: is the EAPEnf-ICU reliable and does it present theoretical and empirical consistency to be validated as an instrument that measures the exercise of patient advocacy by the intensive care nurse?With this in mind, this study aims to describe the psychometric validation of the EAPEnf-ICU.

Research Methods
This is an instrumental research with a cross-sectional design (Ato et al., 2013), carried out from January to June 2021, directed at factorial validation of the EAPEnf-ICU (M.Vargas et al., 2022).The Patient Advocacy Scale for Intensive Care Nurses (EAPEnf-ICU) was a instrument elaborated on 2020, in Brazil, from a Survey and Integrative Literature Review.The validation of face and content was performed by expert judges and by pre-test (M.Vargas et al., 2022).
To this end, the instrument was applied to a sample of Brazilian intensive care nurses, which enabled the execution of stages six, seven, and eight described by DeVellis ( 2012), aimed at concluding the evaluation of the instrument measuring properties using statistical tests.

Settings, Sample, and Data Collection
Non-probabilistic convenience sampling method was used to select the participants according to their availability.However, despite the non-probabilistic sample, we sought to use a minimum sample size to ensure reliability and possible result generalization to the study population.To guarantee the minimum sample size, it was necessary to estimate the population of Brazilian intensive care nurses.In the absence of official data, the total number of existing hospital beds was adopted (45,848 beds from public and private sectors) (Associac xa˜o de Medicina Intensiva Brasileira, 2020), divided by 10 (recommended ratio of 10 beds per nurse) and multiplied by 4 (considering the four work shifts-morning, afternoon, night 1 and night 2), according to the National Health Surveillance Agency, which provides for the minimum requirements for ICU operation (Ageˆncia Nacional de Vigilaˆncia Sanita´ria, 2012).Thus, a simple random sample with a 95% confidence interval of at least 377 respondents was defined.
The minimum time of 1 year working as an adult intensive care nurse was adopted as an inclusion criterion, following the perception that this is the necessary time to adapt to routines and the work environment.Thus, 382 intensive care nurses working in adult intensive care units-both from the Unified Health System (SUS) and from private institutions-participated in this study, representing the 26 Brazilian states and the Federal District.Regarding the participants, five were excluded for not completing the instrument in full, thus resulting in a final sample of 377 participants.
Participants were recruited through e-mails addressed to professionals, institution's service heads, researchers and professors from different cities in the country, requesting that the access link to the instrument should be disclosed on their professional contact networks.With a view to expanding the sample and achieving representativeness of all regions of the country, a page on a social network platform (Instagram) was created, which allowed boosting access to the form by Instagram professional users.
Data collection took place from January to June 2021, using an online form (Google form).Nursing professionals were given information about the study when accessing the form page, and were also invited to participate in it.For this purpose, they should sign the Informed Consent Form (ICF), in accordance with resolution 466/ 2012 (Ministe´rio da Sau´de, 2012), and only upon their signature the form was made available for completion, which took about 20 min to be filled out.

Measurement
The initial part of the questionnaire contained the sample characterization items (i.e., sociodemographic questionnaire), such as date of birth, city and state of residence, sex, number of children, courses taken, time working in the ICU, number of employment relationships, and ICU specialty and complexity, among others.After filling out the questionnaire, participants had access to an introductory text that presented the operational definition of patient advocacy, according to Hanks (2010), which allowed all study participants to be introduced to the concept in a standard manner.
After participants became aware of the concept, there was presentation of 57 questions that corresponded to the advocacy measuring items, measured using a fivepoint Likert scale, with the following options: 1-''totally disagree''; 2-''disagree more than agree''; 3-''do not agree and nor disagree''; 4-''agree more than disagree,'' and 5-''totally agree.''These questions were firstly organized into three dimensions, namely: Nurse's attributes, which sought to know nurses' beliefs about the necessary attributes for the exercise of patient advocacy (7); Patient advocacy actions (33), and Barriers to patient advocacy (17), dimensions that underwent changes after exploratory factor analysis.

Data Analysis
The data were organized in a table using Excel 2010 Ò software and subsequently submitted to statistical tests to ensure the measure construct validity using R software.
Initially, descriptive statistics were calculated to characterize the sample, using frequency, mode, and standard deviation analyzes.After performance of descriptive statistical analysis, the database composed of 377 participants was divided into two groups: bank 1 was composed of data from 189 participants, and bank 2 of data from 188 participants.In order for there to be representation of all Brazilian states and the Federal District, the participants were first grouped by state, and then divided randomly and proportionally into two groups.
First, an exploratory factor analysis (EFA) was conducted to examine the factor structure of the scale.This procedure made it possible to analyze the dimensionality of the instrument as well as the factor loadings of the scale items.This is an elementary approach aimed at obtaining initial indications of construct validity based on the internal structure of the measurement.
In the analysis of the measuring properties, evidence of factorial validity of the measure structured by 57 items was sought through exploratory factor analysis (EFA) of bank 1 data (n = 189) using a Pearson correlation matrix, performed using R (R Core Team, 2021) software with the aid of the psych (Revelle, 2021) and EFA packages (Navarro-Gonzalez & Lorenzo-Seva, 2021).Sample adequacy and data factorability were verified using the Kaiser-Meyer-Olkin (KMO) test and Bartlett's test of sphericity.In this study, the Kaiser-Mayer-Olkin (KMO) test was used to assess the suitability of the data for factor analysis (Kaiser, 1974).KMO values range from 0 to 1, so that values between 0.6 and 1 indicate an appropriate sample size to continue with the factor analysis (Tabachnick & Fidell, 2012).KMO values below 0.5 are unacceptable and indicate that the sample cannot be factored (Hair et al., 2006).
The Minimum Rank Factor Analysis was used for factor extraction, with Oblimin rotation, as well as the parallel analysis based on the Minimum Rank Factor Analysis (Timmerman & Lorenzo-Seva, 2011).The number of extracted factors was based on the Optimal Implementation Parallel Analysis (Timmerman & Lorenzo-Seva, 2011), with a 500-database simulation, which is based on the comparison of the average variance explained by the items displayed on the real bank versus those on the simulated banks.For item retention, factor loading values equal to or greater than |0.40| were used as a criterion.
Once the factor structure of the scale was known, Confirmatory Factor Analysis (CFA) was conducted.This is a more focused statistical approach that aims to assess whether the factorial structure obtained through EFA can be confirmed in a new sample.To this end, the factorial structure is tested using a statistical model that considers different indicators as criteria for fitting the data (Hu & Bentler, 1999).
With the aim of finding more evidence of construct validity, confirmatory factor analysis of bank 2 data (n = 188) was performed.In this process, four models were compared, specifically: (1) model found in the exploratory factor analysis (five-factor model with related factors); ( 2) five-factor model with unrelated factors; (3) one-factor model; (4) second-order hierarchical five-factor model.This analysis was performed on R17 software, using the lavaan (Rosseel, 2012) package, with the aid of the Diagonally Weighted Least Squares (DWLS) estimator.Model fitting was evaluated using the following combination of fitting statistics: ratio of chi-square (x 2 ) to degrees of freedom (df), Tucker-Lewis Index (TLI), Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR).Following the recommendations by Brown (2015), cutoff values lower than 5 were used as parameters of good model fit for the ratio of the chi-square to the degrees of freedom (x 2 /df \ 5); values close to 0.90 for CFI and TLI, and 0.08 for RMSEA and SRMR.When comparing the models tested, the best model would be the one with the best fit indexes.
For the analysis of internal consistency, given the nature of the data obtained from the scale, Cronbach's alpha and McDonald's omega coefficients were used, consider values above .70appropriate (Hair et al., 2006).

Ethical Considerations
This study was approved by the Human Research Ethics Committee at the Institution of origin of the author and followed the guidelines set out in the National Health Council Resolution 466/2012, in compliance with Brazilian legislation (Ministe´rio da Sau´de, 2012).All participants were instructed on the research and signed the Informed Consent Form (ICF).

Results
Of the 377 intensive care nurses participating in the final sample, 75.5% of them were female; in the 31 to 40 age group (55.8%); with children (54.5%); presenting religious beliefs (87.2%), and most of them had family experience with a terminal illness (63.5%).
With regard to experience and professional qualification, working time in an adult ICU presented a 9-year mean and an 8-year median, and 81.1% of professionals had specialization in that area of knowledge.In relation to the work environment, 57.9% of the participants worked in public institutions, with 25-bed mean and a 20-bed median; predominance of general units (71.8%); effective employment (88.3%); and 37-hr weekly workload mean and 36-hr weekly workload median.
Besides, participants reported the existence of an Ethics Committee at the institution (78.9%); experience or knowledge of palliative care (83.2%), as well as holding meetings in the workspace, whether administrative, scientific, or between teams (89.1%).
Regarding the family member presence in the ICU, it was found that in most units (48.4%) this occurs through either an extended visit or a restricted-hour visit (44.1%).

Construct Validity
Initially, using database 1, EFA of the 57 items of the instrument was performed.The results showed that the sample and data correlation matrix were adequate for this analysis: KMO = .88;Bartlett's test of sphericity, x 2 (1,596) = 10,992.5,p \ .001.
From EFA, it was observed that items 19, 40, and 41 had factor loading values lower than |0.40|.Given the parameter used for item retention ( ʎ ø |0.40|), these items were excluded, resulting in a scale composed of 54 items at this stage.The parallel analysis suggested the extraction of five factors (Figure 1), and the explained variance observed for factors was as following: 33.60% for Factor 1; 15.66% for Factor 2; 6.48% for Factor 3; 4.22% for Factor 4, and 3.81% for Factor 5 (the value of the average variance explained for the simulated Factor 5 was 3.71%, the lowest index in relation to the real bank data).The factor loadings carried in the five factors are shown in Table 1.The internal consistency indexes of the scale considering each factor were: Factor 1 (a = .97,v = 0.98); Factor 2 (a = .91,v = 0.94); Factor 3 (a = .90,v = 0.94); Factor 4 (a = .89,v = 0.94); Factor 5 (a = .83,v = 0.89).
After performing EFA, CFA was carried out.Table 2 presents the fit indexes found when comparing the models tested by CFA.
When comparing the fit indexes of the models presented in Table 2, it is observed that model 1 (five-factor structure with related factors, found in the exploratory analysis) showed better data fit than the other models tested.Thus, the CFA fit indexes for the five-factor model (model 1) performed with the second database were: x 2 (1,315) = 683.34,p ..05; x 2 /df = 0.51; CFI = 1.00;TLI = 1.13;RMSEA = 0.01 (90% CI = [0.00,0.02]); SRMR = 0.09.The confirmatory model factor loadings varied between 0.89 (item 34) and 0.48 (item 7), and are presented in Table 3.Furthermore, a graphical representation of the five-factor model (model 1) can be seen in Figure 2.

Description of the Final EAPEnf-ICU
After grouping the items, according the EFA results, it was necessary to carry out a conceptual analysis to identify the dimensions of the construct that emerged in the results.Figure 3 presents the conceptual organization corresponding to each dimension.
From the structural exploration of the scale through EFA and CFA, as well as the analysis of the measure internal consistency, the final version of the EAPEnf-ICU was composed of 54 items, scored in a five-point range graduation, where 1 was attributed to ''strongly disagree'' and 5 to ''strongly agree.''The items are divided into 5 dimensions/factors (Chart 1).
As an analysis possibility, the total score was defined by the sum of the scores achieved in the assessment of the five dimensions, and it was necessary to invert the scores obtained in Factor 2-Barriers associated with the intensive care clinical and organizational complexity, and Factor 4-Barriers associated with divergences and ethical-professional limits in intensive care.
The scale score can vary from 54 (1 point for all items) to 270 points (5 points for all items) and the higher the score achieved, the greater the intensive care nurses' beliefs in relation to exercising patient advocacy.Thus, the intensive care nurses' beliefs in relation to patient advocacy can be considered from the following standardized score: between 54 and 108 points-very low; between 109 and 162 points-low; between 163 and 216 points-high, and equal to or above 217 pointsvery high.

Discussion
The main objective of the research, in the first instance, that is, the exploratory factor analysis (EFA) of the instrument, got results that showed evidence of the factorial validity of the EAPEnf-ICU measure This procedure is a crucial step in the development of a measurement instrument, given that this objective analysis verifies how the latent trait (patient advocacy) appears in a factorial structure consisting of items aimed at measuring the behavior manifested.More precisely, it is from the analysis that one explores how the items proposed for the scale are grouped into different semantic factors.For this purpose, it assumes that a set of statements highly related to each other constitute a dimension of the investigated construct (Nunnally & Bernstein, 1994).In the study results, the emergence of five factors related to each other was observed, most of which presented satisfactory values of factor loadings.
It should be noted that in the exploration stage of the measure factorial structure, three items were subtracted from the general structure of the EAPEnf-ICU, resulting in a 54-item scale.This is because the items did not meet the item retention criteria, based on the values of the factor loadings of each item, according to the criteria by Hinkin (1998).The procedure for excluding items is a recurrent practice in studies of measure properties (Costello & Osborne, 2005), and it is always recommended in cases such as this study, in order to guarantee the measure parsimony and objectivity, helping refine the instruments (Pasquali et al., 2010).
Regarding the factors, the EAPEnf-ICU Factor 1, Clinical and organizational advocacy in intensive care (22 items) addresses defense actions carried out by the intensive care nurse to benefit the patient at the individual (process of death and dying, stimulation of patient autonomy and independence) or collective (promotion of a safety culture, intervention between service limitations and patient needs, and care environment improvement, for example) level.
The actions reflected by the items of this factor are necessary because of these patients' vulnerability in the ICU context due to their critical health condition and the amount of technological apparatus present in the environment, in addition to the fact that they are far from their loved ones, which is caused by specific norms and routines of that unit (Manoel et al., 2022;Nascimento et al., 2021).
Intensive care nurses exercise their role as patient advocates when, to provide a quality care, they base their thinking and action on scientifically recognized clinical criteria.Having this knowledge will allow the professional to participate, along with the multidisciplinary team, in the evaluation and clinical decision-making, besides allowing them to question behaviors that they consider to be inadequate in the clinical management of the patient case (Kurhila et al., 2020;Manoel et al., 2022).
Despite the measures and actions adopted to restore the patient's health, the nurse will experience the process of death and dying on several occasions.Due to the great proximity with the patient, this professional will be able to know their wishes in relation to the therapeutic treatment, and will be able to act as an interlocutor between the patient, family, and the rest of the health team, arranging the involvement of family members in the evaluation of health care (Barnes et al., 2020), promoting patient autonomy and advocacy, and ensuring that their rights and wishes are respected, thus avoiding the maintenance of treatments considered futile and useless (Godinho et al., 2020;Manoel et al., 2022).With regard to Factor 2 (11 items), Barriers associated with the clinical and organizational complexity of intensive care, there is a reflection on the barriers to the exercise of advocacy in the intensive care environment.
One identified barriers to implementation of patientand family-centered care in the ICU classified into four categories: barriers referring to lack of understanding about patient-and family-centered care; organizational barriers; individual barriers, and inter-professional barriers.Many of the barriers mentioned are reflected in the items that form Factor 2, such as: nurses' work overload, inadequate work environment, overcrowding, absence of family during visiting hours, lack of professional motivation, and communication barriers with doctors (Kiwanuka et al., 2019), which suggests relationships between patient/family-centered care and the exercise of advocacy.
In the Brazilian reality, it was evidenced that physical and emotional fatigue generates less patient advocacy or less advocacy efficiency by nurses in decision-making processes.Fatigue was related to workload, ethical problems experienced, lack of autonomy, and managerial and institutional support, among others, which contribute to the occurrence of moral distress in decisions with emotional, personal, and professional consequences of illness, dissatisfaction, absenteeism and, consequently, impaired patient care (Carvalho et al., 2020).
With regard to Factor 3 (8 items), Attitudes to promote the autonomy of patients and family members in intensive care, it is related to actions that aim to promote the means for the patient to be able to make decisions, and thus exercise their autonomy, and also advocate for  the family member presence during the ICU hospitalization process.
The critical condition patients' family is fragile and unstructured, in need of support to overcome challenges and feelings of concern, anguish and fear; the nurse's role is fundamental to guide, clarify (Poerschke et al., 2020), or promote the partnership between family and team (Hinkin, 1998), which are so important for the best and most prudent decisions.
When crucial decisions fall on the family, qualified information and adequate interaction between family and professionals are fundamental instruments for such decisions (Castro et al., 2018), which have shown to be valued in the items of this factor.
Factor 4, Barriers associated with disagreements and ethical-professional limits in intensive care (5 items), in turn, groups items related to conflicts between health teams working in ICU, as well as ethical issues that become barriers to the exercise of advocacy and consequently can lead to disqualification of the care provided to patients and their families.In this dimension, conflicts within the nursing team or with other health team professionals were highlighted.
Even if nurses realize unnecessary interventionsunable to change the prognosis or improve the patient quality of life-, impotence, and lack of control or authority to influence or avoid such conduct are sources of conflict, especially when they do not feel heard by doctor colleagues (Manoel et al., 2022;McAndrew & Hardin, 2020).
Among the most frequent conflicts faced by intensive care nurses it is the frequent use of resources considered useless for certain cases, as well as privacy violation, situations identified as moral dilemma and cause of moral distress, that is, they represent impossibility of taking appropriate action because internal or external factors (Pereira et al., 2020;Pishgooie et al., 2018).
This reinforces the complexity of exercising patient advocacy in the ICU, in a process that can be time-consuming, exhausting, and frustrating for nurses, even more so when these professionals believe that their knowledge and opinions are not taken into account when making decisions related to patient therapeutic conduct.In this sense, the items proposed for this factor seek to assess the extent to which nurses perceive or experience ethicalprofessional conflicts in the ICU context.
Finally, Factor 5, Intensive care nurse's personal and professional background (7 items), assesses the intensive care nurse's personal/professional background and its impact on the patient advocacy process.
Studies point out that nurse professional skills and personal values are among the aspects that influence patient advocacy.From this perspective, nurses that are older, have more professional experience, have graduation in intensive care, do not present health conditions, and with a favorable economic situation, exercise a stronger patient advocacy.Still, those who received some type of training regarding advocacy in end-of-life situations, end up doing it more actively and, as a result, show fulfillment, promotion, and satisfaction at work (Beigzadeh et al., 2016;Davoodvand et al., 2016;Forsberg et al., 2015;C. P. Vargas et al., 2022).Such findings reflect the set of items proposed by this dimension.
Thus, after the initial exploration of the factorial structure of the EAPEnf-ICU, it was observed whether the organization found in the first stage would be confirmed in an unprecedented sample of the measure target population (i.e., Brazilian intensive care nurses).For this, four different models were compared in order to verify whether the proposed model (five-factor model with related factors) would be better than the alternative models.The results of this stage demonstrated that the 54 items of the EAPEnf-ICU were structured into a matrix composed of five related factors, confirming the structure found in EFA.Furthermore, the quality indexes of the model proposed in this CFA were excellent, according to criteria suggested by Brown (2015).
Taken together, EFA and CFA results attest that the scale is sensitive enough to capture the latent trait it intends to measure.In general, the findings of both approaches show that the measure proposed in this study has factorial validity, indicating that the EAPEnf-ICU items are loaded in a five-factor structure, advancing in the debate about patient advocacy.In other words, since validity is the ability of an instrument to assess what it is intended to measure (Borsboom et al., 2004) and factorial validity is a type of construct validity (Smith, 2005), it can be said that the new assessment measure is valid for measuring the exercise of patient advocacy by Brazilian intensive care nurses in the ICU context.
In addition, the results obtained in the calculation of Cronbach's Alpha and McDonald's Omega coefficients demonstrated that the measure presents good internal consistency/reliability indexes.This concept concerns the stability and precision of the instrument in assessing what it proposes to measure, that is, whether the results can be reproduced under the same conditions of application of the study (Willoughby, 1935).Thus, the results indicated that the scale presents sufficient evidence of accuracy, demonstrating that the structure found in both samples (bank 1 and bank 2) is consistent in the construct assessment (Hair et al., 2006).
With the application of this instrument in health services, it becomes possible to permanently evaluate patient advocacy actions.Research based on the use of the Scale will produce new knowledge about this object, deepening it in relation to its theoretical and practical scope for nursing.

Limitations
The first study limitation relates to the non-probabilistic sampling, indicating results from a specific context and moment.A second limitation concerns the absence of other measures to observe the relationship between the scale and the convergent and divergent constructs.This comparative process of the relationship between EAPEnf-ICU and other measures that assess correlated and non-correlated constructs is extremely important to verify the measure convergent and discriminant validity.
Another limitation refers to the period of data collection and the development of the work itself, during the Covid-19 pandemic.Therefore, this situation had a direct impact on people's work and lives, resulting in limitations of physical access to the target population, as well as in the situations that have been faced by the respondents, leading to changes in sensitivity, emotional state, and overloads, among others.

Conclusion
As evidenced in these results, the EAPEnf-ICU in its final version, composed of 54 items, distributed in five factors, constitutes a valid and reliable instrument to carry out the assessment of the intensive care nurses' beliefs and actions in the exercise of patient advocacy, as it presents theoretical and empirical consistency.
Those results have important theoretical implications for the field of patient advocacy studies.The main implication concerns the fact that EAPEnf-ICU is an innovative instrument for assessing the intensive care nurses' beliefs and actions in advocating the ICU patients' health interests.
It is worth mentioning the lack, to this date, of a valid and reliable instrument to assess the nurses' beliefs and actions when advocating for the patient in the intensive care environment.In this way, the EAPEnf-ICU is a pioneer, and it is not possible either to make comparisons with other versions, or consider the reality of different countries.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Figure 1 .
Figure 1.Parallel analysis based on minimum rank factor analysis.

Figure 3 .
Figure 3. Definitions of the dimensions formed.

Table 1 .
Instrument Factor Loads and Commonality.

Table 2 .
Adjustment Indicators of the Structural Models Tested in CFA.