Development of a Questionnaire for the Reflective Practice of Nursing Involving Invasive Mechanical Ventilation: Assessment of validity and reliability

Abstract Aim To develop the Questionnaire for Reflective Practice of Nursing Involving Invasive Mechanical Ventilation (Q‐RPN‐IMV), a Japanese self‐evaluation instrument for ward nurses’ IMV practices. Design Cross‐sectional survey. Methods Participants were 305 ward nurses from five hospitals in Japan with nursing involving invasive mechanical ventilation. Items concerning the process of nursing practice, including the thought process related to ventilator care, were collected from the literature and observation and interviews with five IMV specialists. Construct validity, concurrent validity, internal consistency and test–retest reliability were tested. Results Initially, 141 items were collected and classified into three domains (i.e., observation, assessment and practice). Examination of exploratory factor analysis yielded five factors in the observation domain, six factors in the assessment domain and six factors in the practice domain. The data exhibited internal consistency, stability and concurrent validity. Items of practical content, including thoughts on ventilator care, are useful for preparing educational programmes.


| INTRODUC TI ON
Worldwide, the number of people who require ventilator-assisted care has been increasing and this population has unique and variable care needs (Rose et al., 2015). This is especially so in Japan where the rate of ageing of the population is one of the highest in incurable diseases, and the required care varies implicitly depending on the situation.
Invasive mechanical ventilation-associated medical accidents and complications, some of which can be fatal, have become increasingly frequent. Such accidents occur not only due to inadequate ventilator maintenance, but also due to a lack of knowledge and patient observation by medical staff (Aiken, Clarke, Cheung, Sloane, & Silber, 2003;Aiken, Clarke, Sloane, Sochalski, & Silber, 2002;Bond & Raehl, 2007).
In Japan, there are no respiratory therapists with specialist training, unlike the United States; thus, nurses from various backgrounds provide care regarding oxygen therapy, respiratory physiotherapy and artificial respiration therapy to people in hospitals or homes who require therapy (Uzawa, 2006). Therefore, nurses of all skill levels must be able to administer high-quality IMV care safely in various settings. Accordingly, nurses often use tools to record ventilator settings chronologically and use checklists to confirm IMV care items, such as checking the cuff pressure.
In addition to these tools, guidelines (on sedation and ventilator weaning bundles, automated weaning systems and preventing ventilator-associated pneumonia) have supported decision-making among nurses and resulted in reductions in patient time spent on mechanical ventilation and improved survival rates (Girard & Ely, 2008). Despite these advances, current research highlights significant variation and inconsistency in clinicians' assessment and practices, which have been shown to have an adverse impact on safety and patient outcomes (Rose & Nelson, 2006). Although some variation is to be expected, it becomes a problem when application of interventions (i.e., weaning protocols) leads to ineffective outcomes for people (Ericsson, Whyte, & Ward, 2007). This suggests that clinical guidelines alone cannot ensure optimum nursing practice for mechanical ventilation.
When nurses provide care during mechanical ventilation, they do not only focus on the criteria provided in guidelines but on patient-centred information collected from objective physiological and subjective criteria. Especially, less experienced nurses require more encounters with cues (Kydonaki, Huby, Tocher, & Aitken, 2016). Although inexperienced nurses can only record items on existing checklists, they may not understand the implications of those items and may be unable to assess the status of ventilated people. Therefore, these nurses experience uncertainty, stress and anxiety when administering IMV care.
Professional advice and effective and continuous educational programmes are required for nurses who administer IMV care without supervision. In Canada and some other countries, paid personal support workers with minimal appropriate training and education often provide IMV care (Brooks, Gibson, & DeMatteo, 2008).
Although IMV care requires advanced nursing practice skills (Myers, 2013), basic nursing education does not address these skills.
Moreover, most nursing educational programmes generally focus only on imparting knowledge on the techniques related to the individual aspects of IMV (Bloos et al., 2009). Consequently, people with various conditions may receive IMV care from nurses who have not received an integrated education that includes all the IMV-related skills. Benner, Sutphen, Leonard, Day, and Shulman (2009) explained that "to take action in a given patient-care situation, the nurse must have a fluent grasp of the relevant medical information and be able to translate it into practical knowledge." The nursing process, a series of premeditated nursing actions that maintain the best medical environment for people, is a standardized process used to achieve nursing goals. This process comprises four components: assessment, planning, implementation and evaluation.
In addition, nurses provide qualitatively and quantitatively adequate nursing care for restoring the health of people whose conditions have changed. Using the nursing process, nurses can plan a course of action to improve a patient's condition (Iyer, Taptich, & Bernocchi-Losey, 1986).
Given that nursing care is based on the nursing process, the nursing practices associated with IMV should also be analysed accordingly. Therefore, a stepwise and structured education programme based on the nursing process should be provided to nurses, particularly those with little experience in IMV care.
Based on the above literature review, this study aimed to develop an instrument that could clarify the practical nursing process of ventilator care practiced by general nurses in general wards in Japan, where individual nurses need to practice artificial respiratory care according to the situation on each occasion.
In this study, we observed the provision of IMV care by IMV specialists and analysed their assessment processes. Subsequently, we developed a Questionnaire for the Reflective Practice of Nursing Involving Invasive Mechanical Ventilation (Q-RPN-IMV) that focuses on the assessment, implementation and evaluation components of the nursing process. The reliability and validity of this questionnaire were also analysed.
This might enable ward nurses to self-evaluate their own IMV practices and assess the effectiveness of any educational programmes for ward nurses. It could also be used in other countries where nurses working in hospitals or at home will be required to practice IMV care.

| ME THODS
This study was conducted to develop the Q-RPN-IMV, which covers the practical nursing process of ventilator care in the general ward, including the thought process. The study consisted of item development and validity and reliability testing.

| Item development
To develop the items for the Q-RPN-IMV, we first reviewed the literature in Japanese and international databases (i.e., PubMed, CINAHL and Japan Medical Abstracts Society database) using the keywords "mechanical ventilation," "bundle," "nursing" and "reflective practice." We also conducted observational and semi-structured interviews of five IMV specialists, two of whom were certified specialists in critical care nursing and three of whom were certified intensive care nurses.
In this way, it covered the process of nursing practice, including the thought process of ventilator care as an item. The literature review and interviews yielded 141 items for the Q-RPN-IMV, and these were classified into three domains (i.e., observation, assessment and practice). This list underwent statistical analysis using the survey data.

| Participants
The survey participants were all ward nurses from five acute care general hospitals in Central Japan (N = 305). These hospitals were a convenience sample and had 300-1,000 beds, which showed that they were typical urban, middle-to-large-sized general acute care hospitals. All ward nurses at the five hospitals were contacted about possible participation in this study. The sample size was determined using the general rule for factor analytic procedure that requires a minimum of three respondents per item (Kline, 1998).

| Measures
The questionnaire included the following: (a) demographic and professional characteristics; (b) the 141 items (i.e., observation, assessment and practice) for Q-RPN-IMV; and (c) the Educational Needs Assessment Tool for Clinical Nurses (Miura & Funashima, 2006).
The demographic and professional characteristics included nursing experience, sex and the number of IMV cases experienced. The 141 items (observation, assessment and practice) for the Q-RPN-IMV began with the stem "How often do you practice ventilator nursing care?" with answers recorded on a 5-point Likert scale ranging from 1 (never) -5 (always).
The Educational Needs Assessment Tool for Clinical Nurses was used for concurrent validity testing, because this scale is one that yields a higher score for items requiring education regarding professional nursing practice. Regarding items considered necessary for ventilator care, we considered that education was necessary for less frequent practice items; it was the best, currently available instrument among the items of ventilator care. The Educational Needs Assessment Tool for Clinical Nurses consisted of 25 items with five domains measured on a 5-point Likert scale, with higher scores representing items requiring education regarding professional nursing practice. Cronbach's alpha for this study was 0.94.

| Data collection procedure
The researchers explained this study to the nursing managers. When they agreed to participate, the nursing managers were asked to forward questionnaires to the ward nurses. Once the ward nurses agreed to participate in this study, they completed the questionnaires concerning nursing practice, placed their completed questionnaires in a sealed envelope and returned it to the researchers.
To evaluate test-retest reliability, ward nurses from five hospitals (N = 152) were asked to complete the questionnaire again after 2 weeks. During the 2-week interval, participants did not participate in any training, seminars or educational activities.

| Data analyses
The Q-RPN-IMV was tested for face, factor and concurrent validity. Internal consistency and reproducibility were tested to determine reliability. Each domain was subjected to an exploratory factor analysis using the principal factor method and varimax rotation to test the validity of the models, based on the postulated constructs (i.e., whether all the items for a single factor loaded >0.35) and to confirm that the item loadings were theoretically coherent. Initial factor selection was based on eigenvalues >1.0.
After removing items that did not have a loading >0.35 for a given factor, the models were tested via factor analysis, followed by confirmatory factor analysis (Polit & Beck, 2011). The Akaike information criterion (AIC), the comparative fit index (CFI) and the root mean square error of approximation (RMSEA) were used to evaluate the fit of the models to the data (Kääriäinen et al., 2011). With regard to the AIC, the smallest value represents the best classification. RMSEA values <0.05 indicate that a model has a close fit, whereas those between 0.05 -0.08 indicate a reasonable error when approximating a given structure (Browne & Gudeck, 1993). A CFI >0.9 was assumed to indicate an adequate fit. The final measurement models were selected by examining the four indices of fit and choosing the model with the best indices. Pearson's correlation coefficients for the Q-RPN-IMV and the Educational Needs Assessment Tool for Clinical Nurses were calculated to evaluate concurrent validity. We decided that there was a need for education for items with low frequency of practice. Cronbach's α coefficient was calculated for each of the subscales to assess internal consistency. Spearman's rank correlation coefficients, which measure the strength of agreement between repeated measurements (Fayers & Machin, 2000), were calculated to evaluate test-retest reliability. All statistical analyses were performed using IBM SPSS 22.0 and Amos 22.0 (IBM Corp., Armonk, NY, USA).

| Ethical considerations
Participants were informed of the purpose and methods of the study, risks and benefits of participation, confidentiality of their data and the voluntary nature of participation. Written informed consent was obtained before the interviews for item development and refinement study. The return of the filled anonymous questionnaire was taken as consent to participate in the main study. The study protocol conformed to the tenets of the Declaration of Helsinki (as revised in Edinburgh 2000), and the study process was reviewed and approved by the ethics committee of the Kobe University Graduate School of Health Sciences (approval number: 241).

| Participant characteristics
In total, 432 questionnaires were distributed and 331 returned (response rate: 76.6%). Among them, 305 questionnaires with less than 5% missing data of the total items were used (validity rate: 92.1%).
More than half of the participants were nurses who had nursing experience of more than nine years and cared for more than 11 people who had required ventilator care (Table 1).

| Item selection and domain development
Among the 141 items collected for the Q-RPN-IMV, we classified each item into one of four domains (i.e., observation, assessment, practice and evaluation). Many of the assessment and evaluation items overlapped; therefore, we combined these items into the assessment domain (i.e., observation 26, assessment 67 and practice 48).

| Domain 1: IMV care related to observation
The factor analysis of Domain 1 yielded five factors (Table 2). Items 21 and 22 were related to multiple factors. However, we conducted an inspective factor analysis and arranged these items with factor 2. Twenty-six items that were divided among the five factors had eigenvalues >1.0. Domain 1 consisted of 26 items of five factors.

| Domain 2: IMV care related to assessment
The factor analysis of Domain 2 yielded six factors (Table 3). Item 36 loaded at <0.35 and was removed from the table. Items 12, 49, 50, 51, 53 and 57 were related to multiple factors. To assess these items, we created possible models and conducted an inspective factor analysis. The best-determined arrangement of the question items was as follows: factor 1 comprised items 51, 53 and 57; factor 3 comprised items 49 and 50; and factor 4 comprised item 12. Sixty-six items were arranged among the six factors. The six factors (66 items) included had eigenvalues >1.0. Domain 2 comprised 66 items of six factors.

| Domain 3: IMV care related to practice
The factor analysis of Domain 3 generated six factors (Table 4) (Table 5). Pearson's correlation coefficients were negative for all the items. Accordingly, clinical nurses should be educated on the Q-RPN-IMV items, which are infrequently practiced (r s = −0.37-−0.46; p < 0.001).

| Validity and reliability testing
For internal consistency, Cronbach's α was measured for all factors. Cronbach's α coefficients for all factors ranged from 0.52 to 0.93 (Table 6). For reproducibility, the intraclass correlation coefficient for each subscale, which was calculated using data from the 152 participants who returned their test-retest responses, ranged from 0.55 to 0.85 (Table 6).

| D ISCUSS I ON
To ensure that IMV care reflects education and supports the practice of high-quality IMV care, we classified respiratory care items during a nursing intervention process and developed a questionnaire that could confirm the frequency of respiratory care practices performed by clinical nurses. In addition, we classified the enforcement of six practice factors (44 items

| Reliability of the Q-RPN-IMV
With regard to the internal consistency, we calculated Cronbach's α coefficients and obtained estimates of 0.64-0.93, indicating a satisfactory consistency and suggesting the potential ability to confirm internal consistency for all the factors except factor 1 (Streiner & Norman, 2003 With regard to stable examinations, we calculated the intraclass correlation coefficients between every tested-retested factor to determine the stability of this questionnaire (r s = 0.55-0.85; p < 0.01).
Factors that indicated a firm intraclass correlation coefficient or regular performance in a standard hospital included "assessment of ventilation situation," "assessment of early detection of complications via radiography" and "practices for maintaining maximum breathing capacity." The category that frequently received the lowest ranking, "confirmation of positioning and emergency preparation," was considered to include the items that could be easily addressed.

| Validity of the Q-RPN-IMV
We observed the respiratory care provided by specialists to ensure that the Q-RPN-IMV covered all the items, including ideation.
We qualitatively classified these items with regard to the observation, assessment and practice of nursing processes. We confirmed the contents with a respiratory care specialist and subsequently assessed the validity of the instrument. We performed a pretest among the nurses who participated in the study and confirmed the time required to answer an item. The estimated time burden of this necessary measure on each respondent was approximately 30 min.
Although the extent of the burden on the respondents was considered, we did not modify the pretest because the responders did not indicate the difficulty of each answer and we deemed all the items indispensable for IMV care.

| Clinical implementation of the Q-RPN-IMV
The possible applications of this questionnaire are unlimited, depending on the method of use. It covers all the practice items necessary for respiratory care and the nursing process, including ideation.
Therefore, we determined the meanings and situations appropriately and validated the practice items to ensure that respiratory care practitioners could use this questionnaire during the nursing process. The respiratory care enforcement frequency increased during the retest evaluation because of increased awareness, which likely leads to an increase in the scores of the completed questionnaires.
This nurse practicing respiratory care list could be used as a tool for self-evaluating practice processes. In particular, it could trigger

| Educational implementation of the Q-RPN-IMV
Checklists and textbooks related to respiratory care education often list only the setting and management of machines and/or items related to care provision (Hong, Chen, & Na, 2012). Using our questionnaire, which addresses all the items related to respiratory care (including ideation), we can confirm the type of care provided by a respiratory care practitioner during a nursing intervention process. Using the questionnaire provided in this study, nurses will be able to safely administer IMV care with confidence, perform self-evaluations and improve their IMV care skills. The questionnaire can also be used to create an educational programme that supports high-quality IMV care. Furthermore, educational programme developers, who determine the instructions required for the nursing intervention process, could use this questionnaire for developing respiratory care education depending on the practitioner's needs.

| Study limitations
We extracted the items that were identified as representing crucial respiratory care knowledge into the categories of observation, assessment and practice. However, the nursing process comprises a series of steps and is not as simple as a set of divided processes. A nurse's background, which includes experience and departmental area, determines the required skill level.
We identified the associations between respiratory factors and the educational programme using a covariance structure analysis. Therefore, it is necessary to examine continuity and factor characteristics.

| CON CLUS IONS
The Q-RPN-IMV comprises many items and does not feature an ideal structure. Manuals related to IMV care have been gradually created in bundles. However, IMV care is complicated in the context of nursing practice. Despite good awareness of nurses' difficulties and the provision of care that is directly linked to people' lives, a system of educational programmes related to the nursing process has not yet been established. According to our survey, IMV care items could be covered in tandem with nursing practice processes. The findings of this survey represent a major step towards the future establishment of a ventilator-care education programme for nurses.
Respiratory care nurses can use the questionnaire developed in this study for self-evaluating practice processes. Educational programme developers can control nursing intervention process instructions and use this questionnaire for respiratory care education, depending on the needs of the practitioner. By analysing the structures between factors, education protocols can be created based on continuity and order.

CO N FLI C T O F I NTE R E S T
No conflict of interests have been declared by the authors.