Development and psychometric evaluation of the Ethical Behavior for Thai Nurses Scale

Pisamai Orathai, Wonnapha Prapaipanich, Manee Arpanantikul, Supanee Senadisai

Division of Fundamental Nursing, Ramathibodi School of Nursing, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Rachathewi, Bangkok 10400, Thailand

Abstract: Background: Ethical problems in the nursing profession are mostly found in the undesirable ethical behavior (EBE) of nurses. If there is no concern about these problems, it may negatively affect the quality of nursing care, patients’ safety, and trust in the nursing profession. In assessing nurses’ EBE, it is important to develop an instrument that reflects the actual EBE of nurses.Objectives: To develop and evaluate the psychometric properties of the Ethical Behavior for Thai Nurses Scale (EBTNS).Methods: The construct definition, 5 sub-constructs, 43 items, and the hypothesized model were synthesized and generated from a comprehensive literature review and existing research instruments by using document analysis. The EBTNS was verified for content validity by 4 experts. Of the 43 items, 37 items were retained for psychometric testing. The construct validity was tested by using second-order confirmatory factor analysis (CFA). The EBTNS was tested on 1500 professional nurses working at tertiary, secondary, and primary hospitals in 6 regions (the central, northern, northeast, eastern, western, and southern) of Thailand. Proportional quota sampling was applied to recruit participants.Results: The hypothesized model fitted the empirical data. All 37 first-order indicators and 5 second-order sub-constructs showed significantly standardized factor loadings and effect sizes. The standardized factor loadings of 37 first-order indicators ranged from 0.58 to 0.89, and the construct reliabilities were moderate to substantial (R2 = 0.34-0.79). The 5 second-order sub-constructs displayed the standardized effect sizes, ranging from 0.81 to 0.96, and accounted for 65%-96% of the total explained variance. In addition, the construct validity was supported by convergent validity and discriminant validity. It also satisfied Cronbach’s alpha coefficients of 5 sub-constructs from 0.87 to 0.94, and a full scale was 0.97.Conclusions: The psychometric properties of the EBTNS are acceptable. This instrument has the potential to be used for evaluating and monitoring to prevent ethical problems of Thai nurses.

Keywords: construct validity ? convergent validity ? discriminant validity ? ethics ? nurses ? scale development

1. Introduction

Ethical behavior (EBE) is an essential element of nursing practice and is expected of all nurses. Nurses must maintain ethical responsibilities in taking care of individuals of all ages, families, groups, and communities, sick or well, in all settings, including promotion of health, prevention of illness, care of clients, and health rehabilitation.1To date, nurses are faced with complex situations where they are expected to provide safe and high-quality care that should be led in a direction that can promote health integrity in holistic care, including physical, mental, emotional, social, and spiritual dimensions.2

The demand for optimum care is a complicated issue that results in ethical problems for nurses.3Moreover, contextual factors such as advancement in medicine and technology, new diseases, insufficient resource allocation, nursing shortage, rising health care costs, increases in the aging population, patients’ rights, and work overload may affect nurses’ EBE and lead to ethical problems.2,3

In 2018, Gallup Poll found that 84% of Americans rated the honesty and ethical standards of nurses as high or very high. Nurses earned the top ranking among a diverse group of professions for the 17th consecutive year.4In Thailand, the Thailand Nursing and Midwifery Council conducted a survey regarding the ethical problems of Thai nurses according to professional ethics in 4 regions of Thailand (central, northern, southern, and northeastern). The research team collected data in 1 province of each region by using a brainstorming strategy on the issues regarding ethical problems of Thai nurses. There were 20-25 people in a group for each province. Also, a research team visited to observe the EBE of nurses at health centers, community hospitals, general hospitals, center hospitals, and nursing colleges. The results showed that most nurses had positive EBEs such as good relationships, respect for others, responsibility, sacrifice, honesty, kindness, and propriety. However, some nurses had negative EBEs such as inappropriate responses to patients, less sacrifice, more commercial behavior, or providing unclear or delayed information to patients.5Recently, Arpanantikul, Senadisai, Orathai, & Prapaipanich have studied ethical problems in the nursing profession found in Thailand and its impact on clients. They collected data by interviewing 30 nurse administrators from professional nursing organizations. The results of content analysis revealed that ethical problems in the nursing profession found in Thailand were malpractice in nursing care, violating standard of practice, inefficient communication, disclosure of patients’ information, inappropriate work behavior, dishonesty, and personal ethical problems. Subsequently, these problems resulted in non-standard practice for patients, complications and adverse reactions under nurses’ care, prolonged stays in hospital, or even death of patients.6

It can be seen that ethical problems in the nursing profession are mostly found in the undesirable EBE of nurses. Even though the ethical problems occur once in a while in such a small number of nurses, if there is no concern about these problems, they may negatively affect the quality of nursing care, patients’ safety, and trust in the nursing profession.6Thus, nurses are required to have nursing competency and ethical-based practice as standards to provide safe and high-quality nursing care.

In assessing nurses’ EBE, it is important to develop an instrument that reflects the actual EBE of nurses. A literature review of instruments measuring the EBE of Thai nurses found that there are several instruments measuring EBE in the nursing profession, and each instrument measures specific dimensions of EBE.7-12In order to have an instrument that measures EBE covering every dimension in the nursing profession, the research team was interested in developing an instrument measuring EBE for Thai nurses, and this research had the main objectives to develop and evaluate the construct validity and reliability tests of the developed instrument.

2. Literature review

EBE is essential for a nurse to function properly.13Nurses who behave unethically will negatively affect the quality of nursing care, patients’ safety, and other people’s confidence in the nursing profession. Therefore, unethical behavior should be reduced and protected for all nurses.

Based on a literature review, there are different definitions of EBE in the nursing profession. Most of them defined the specific dimension of EBE.7-12In order to have a clearly operationalized definition of EBE covering every dimension in the nursing profession, the research team reviewed the related literature and existing research instruments from the Thailand Nursing and Midwifery Council by using document analysis.14-18Document analysis is an effective method of gathering data because documents are manageable, practical, easily accessible, with reliable sources of data. Moreover, documents are stable and had non-reactive data sources. The research team read and reviewed the data multiple times. Also, the documents were not changed by the research team or research process.19,20All the related literature and instruments were granted access by the Thailand Nursing and Midwifery Council.

In this research, the definition of EBE, sub-constructs of EBE, and indicators of each sub-construct were derived from comprehensive literature review and existing research instruments under the following the Nursing and Midwifery Professional Ethics Regulations, B. E. 2550.16Also, they were used for developing the Ethical Behavior for Thai Nurses Scale (EBTNS). “Ethical behavior” means behaving according to the moral principles and standards determined by the professional organization in nursing. Among the sub-constructs of EBE are (1) the profession of nursing (PFN) (2) the practice of nursing (PAN), (3) the treatment for patients or clients (TPC), (4) the treatment for colleagues both in and out of the nursing profession (TCB), and (5) the practice of human subjects research (PHR). “The profession of nursing” represents the behaviors of nurses that promote the reputation of the nursing profession. “The practice of nursing” reflects the behaviors of nurses that include providing nursing care and midwifery to patients or clients according to professional standards. “The treatment for patients or clients” means the behaviors of nurses that support the ethical and legal PAN and midwifery for patients or clients. “The treatment for colleagues both in and out of the nursing profession” constitutes the behaviors of nurses that demonstrate respect for dignity among nurses and professionals in the health care team and provide collaboration among colleagues in the health care team for the benefit of patients or clients. “The practice of human subjects research” acts on the behaviors of nurses that follow the ethical guidelines of human subjects research and the code of conduct for researchers. These 5 sub-constructs comprise 6, 13, 7, 6, and 5 indicators, respectively. Later, the hypothesized model of EBE was constructed, as displayed in Figure 1.

Figure 1. Hypothesized model.

3. Methods

The principal objectives of this study were to develop and evaluate the psychometric properties of the EBTNS.

3.1. Sample and sampling procedure

For a study using confirmatory factor analysis (CFA), a large-sample technique, the issue is whether the sample size is adequate for achieving the desired power for factor loadings and effect sizes tests.21-23The rule of thumb for CFA is the ratio of cases to estimated parameters, and a ratio of 10:1 to 20:1 is a commonly suggested ratio for minimum recommendations.21,24,25In this study, a sample size was calculated using 15 participants:1 estimated parameter. There were 85 estimated parameters. Therefore, 85 estimated parameters multiplied by 15 equals 1275. However, planning to deal with the violation of multivariate normality requires a larger sample size to increase the robustness of the standard errors and parameter estimates and result in trustworthy results.21,24,25An additional 20% of participants were added, and 1500 participants were recruited in the study.

Proportional quota sampling stratified by regions (central, northern, northeast, eastern, western, and southern) and level of hospital (tertiary, secondary, and primary hospitals) was applied to recruit participants. Based on the proportion of hospital level equals 0.22:0.62:0.16. Of the 1500 professional nurses who were willing to participate voluntarily in the study, there were 338, 923, and 239 participants from tertiary, secondary, and primary hospitals, respectively.

3.2. Scale development and psychometric evaluation

The research process comprised 2 steps. The first step was scale development, which began to develop instruments by identifying construct definitions and sub-constructs, synthesizing and generating items, and constructing hypothesized measurement models. The second step was conducted to evaluate the psychometric properties of the developed instrument.

Step 1:The scale development step comprised 2 sub-steps: (1) identifying construct definition, constructing sub-constructs, synthesizing and generating items, and constructing hypothesized measurement model, and (2) examining content validity and internal consistency reliability.

(1) ldentifying construct definition, constructing sub-constructs, synthesizing and generating items, and constructing hypothesized measurement model:A comprehensive literature review resulted in construct definition.1,7-12,14-17Document analysis was performed for constructing sub-constructs and generating items for each sub-construct. There were 5 subconstructs and 43 items synthesized and generated in the first draft of EBTNS. Subsequently, the hypothesized measurement model was proposed with 5 subconstructs and 43 indicators.

(2) Examining content validity and internal consistency reliability:Draft 1 of the EBTNS was verified for content validity by a panel of 4 experts with consideration as to whether the items together adequately provided the theoretical soundness of the construct. The panel composed of 3 nurse instructors specializing in nursing ethics and a nurse instructor specializing in tool development. Of the 43 items, 37 items that had item-level content validity indexes (I-CVIs) ＞0.80 were retained and 6 items were deleted. The final draft of the EBTNS comprised 37 items with scalelevel content validity index (average) (S-CVI/Ave) of 0.96 and scale-level content validity index (universal agreement) (S-CVI/UA) of 0.86.26,27A pilot study with 30 participants was conducted for internal consistency reliability testing. It satisfied Cronbach’s alpha coefficients of 5 sub-constructs from 0.87 to 0.94, and a total scale was 0.97.

Step 2:The psychometric evaluation step included construct validity, convergent validity, discriminant validity, and construct reliability tests.

Prior to evaluating the psychometric properties of the final draft of the EBTNS, the characteristics of the participants were analyzed by descriptive statistics, in terms of frequency, percentage, mean, and standard deviation, by using the SPSS program. Later, the construct validity of the EBTNS was tested by second-order CFA using the Linear Structural Relationships (LISREL) program. Before conducting the analysis, the assumptions underlying CFA were tested, which included multivariate normality, linearity, and multicollinearity. The assumptions of linearity and multicollinearity were achieved, but multivariate normality was violated. Nevertheless, multivariate analysis is rather robust to mild to moderate violation of normality, especially as the sample size increases.28A very large sample size is needed to get robust standard errors and reliable parameter estimates by using the maximum likelihood with robust standard errors (MLR) method.21In this study, the total number of participants was 1500, which was enough for the analysis.

The goodness of fit of a hypothesized model of EBE on the second-order measurement model of the EBTNS was evaluated following guidelines for goodness of fit indices, including (1) comparative fit index (CFI) ＞0.90, (2) relative fit index (RFI) ＞0.90, (3) nonnormed fit index (NNFI) or Tucker-Lewis index (TLI) ＞0.90, (4) incremental fit index (IFI) ＞0.90, and (5) standardized root mean square residual (SRMR) and root mean square error of approximation (RMSEA) ＜0.08.25The statistical significance of first-order indicators and second-order effect sizes were evaluated byt-values that exceeded the critical values of ± 2.58 at the 0.01 significant levels. The squared multiple correlation (R2) for structural equations was used to assess the total explained variance of the secondorder sub-constructs with the recommendation as follows:R2= 0.25, 0.50, and 0.75 were levels of weak, moderate, and substantial.29Next, the construct reliability was assessed. The squared multiple correlation (R2) for observed variables was used to assess construct reliability values of the first-order indicators with the acceptable recommendation at the value of 0.50 or higher.21,25Last, convergent validity and discriminant validity were assessed. Convergent validity was assessed by values of average variance extracted (AVE) and composite reliability (CR) of each sub-construct. AVE measures the level of variance captured by a sub-construct versus the level due to measurement error, with the level of 0.50 being acceptable, and 0.70 and above being good. CR is a less biased estimate of reliability than internal consistency reliability. The acceptable value of CR is 0.70 and above,30whereas discriminant validity was assessed by comparing the amount of variance captured by a sub-construct with the shared variance with other sub-constructs.30Otherwise, the levels of the AVE for each sub-construct should be greater than the squared correlation involving the sub-constructs.30,31

4. Results

4.1. Participant characteristics

Most of the study participants were female. The average age and work experience of the participants were 36.70 (S.D. = 8.20) and 12.50 (S.D. = 8.60) years, respectively. Ninety-two percent of the participants earned bachelor’s degrees. More than half of them (64.20%) reported being specialist nurses. The participants were from primary hospitals, secondary hospitals, and tertiary hospitals in the following percentages: 16, 61.50, and 22.50, respectively. Among them, 26%, 19.50%, and 15.70% were from northeast, west, and central regions, and the rest of 38.80% were from northern, eastern, and southern regions. Interestingly, twenty-second point 5 of the participants reported that they were not educated on ethics in the nursing profession.

Sub-constructs/measurement items Parameter estimates B b S.E. (b) t R2 1. PFN(1) To live in a society with justice 0.67 0.39 ＜ - -> ＜ - -> 0.44(2) Respect the law of the country 0.72 0.39 0.02 24.28＊＊ 0.52(3) Profession with good intentions 0.81 0.41 0.02 21.96＊＊ 0.66(4) Profession without discrimination 0.76 0.39 0.02 20.86＊＊ 0.58(5) Do not behave or do anything that may cause the reputation of the profession 0.79 0.41 0.02 21.97＊＊ 0.62(6) Profession without asking any compensation other than the normal service charge 0.66 0.41 0.02 18.40＊＊ 0.43 2. PAN(7) Provide nursing practice according to professional standards 0.70 0.35 ＜ - -> ＜ - -> 0.49(8) Provide nursing care with regard to patient/client safety 0.69 0.30 0.01 20.83＊＊ 0.47(9) Provide nursing care with regard to patient/client saving 0.58 0.36 0.02 17.65＊＊ 0.34(10) Provide nursing care using empirical evidence 0.66 0.39 0.02 19.96＊＊ 0.43(11) Do not give false opinions on any subject with regard to nursing practice and midwifery 0.63 0.35 0.02 19.21＊＊ 0.40(12) Allow clients to make their own decisions without compulsion when they were explained about the treatments 0.76 0.40 0.02 23.12＊＊ 0.58(13) Provide nursing practice with the consent of the clients 0.76 0.38 0.02 23.11＊＊ 0.58(14) Decide on providing nursing care for patients/clients based on the severity of symptoms under appropriate discretion 0.79 0.38 0.02 24.04＊＊ 0.63(15) Tell the truth to patients/clients and their relatives with regard to suitability 0.79 0.41 0.02 23.81＊＊ 0.62(16) Accept mistakes when providing a missed nursing practice 0.83 0.42 0.02 25.04＊＊ 0.68(17) Provide nursing care based on patient/client needs 0.73 0.43 0.02 22.30＊＊ 0.54(18) Care for the patients/clients 0.77 0.37 0.02 22.23＊＊ 0.59(19) Provide nursing practice with regard to patient/client privacy 0.75 0.39 0.02 22.78＊＊ 0.56 3. TPC(20) Do not motivate or persuade anyone to use the nursing service or midwifery for their own benefit 0.66 0.40 ＜ - -> ＜ - -> 0.43(21) Treat patients or clients politely without coercion 0.70 0.39 0.02 19.49＊＊ 0.49(22) Do not deceive patients or clients to mislead interests for their own benefit 0.71 0.37 0.01 26.03＊＊ 0.50(23) Do not disclose the confidentiality of the patients/clients to the unrelated person unless the patient’s consent is granted or to follow the law or responsibility 0.74 0.35 0.02 20.42＊＊ 0.55(24) Do not reject to help those who are in danger from illness when requested and in a role to help 0.76 0.38 0.02 20.90＊＊ 0.58(25) Do not engage in public ways or places unless in an emergency case of first aid 0.77 0.45 0.02 21.31＊＊ 0.60(26) Do not practice/support the illegal PAN and midwifery, practice of medicine, or public health or arts 0.77 0.36 0.02 21.05＊＊ 0.59 4. TCB(27) Honor and respect for dignity among nurses 0.80 0.40 ＜ - -> ＜ - -> 0.65(28) Do not accuse or bully among nurses 0.84 0.44 0.01 33.46＊＊ 0.70(29) Do not persuade patients/clients of other nurses to be their own 0.77 0.44 0.02 27.50＊＊ 0.60(30) To honor and respect the dignity of colleagues in the health care team out of the nursing profession 0.87 0.45 0.01 33.00＊＊ 0.76(31) Do not accuse or bully among colleagues in the health care team out of the nursing profession(32) Provide collaboration among professionals in the health care team for the benefit of patients/clients 5. PHR(33) The consent of the research participants must be required when conducting a human subjects research 0.89 0.47 0.02 29.26＊＊ 0.79 0.83 0.41 0.01 29.80＊＊ 0.68 0.73 0.35 ＜ - -> ＜ - -> 0.54(34) Ready to protect the research participants from harm that might be caused by the research 0.83 0.47 0.02 25.75＊＊ 0.69(Continued)

4.2. Psychometric evaluation

The results of the second-order CFA revealed that the hypothesized measurement model of EBE fitted the empirical data (CFI = 0.98; RFI = 0.98; NNFI/TLI = 0.98; IFI = 0.98; SRMR = 0.05; RMSEA = 0.07; 90% CI for RMSEA = 0.07-0.08). All 37 first-order indicators showed significantly standardized factor loadings atP＜ 0.01. The values of standardized factor loadings ranged from 0.58 to 0.89. Besides, the construct reliabilities of those indicators were moderate to significant, with the values ofR2ranging from 0.34 to 0.79 (Table 1). Furthermore, the 5 second-order sub-constructs showed significantly standardized effect sizes atP＜ 0.01. The standardized effect sizes of those sub-constructs were substantial. The values of standardized effect sizes ranged from 0.81 to 0.96 and accounted for 65%-96% of the total explained variance (Table 2). The results showed that EBE could be measured by 5 sub-constructs and 37 items. These support the construct validity of the EBTNS.

Table 1. Parameter estimates of 5 sub-constructs and their measurement items.

In view of the AVE and CR of EBE, the levels of the AVE and the CR were very good (AVE = 0.97; CR = 0.98). In addition, the levels of the AVE and the CR of 5 sub-constructs were substantial. The AVE of 5 subconstructs ranged from 0.79 to 0.91, and the values of CR ranged from 0.88 to 0.94 (Tables 3 and 4). The levels of the AVE and the CR of each sub-construct were greater than the value of 0.70. Therefore, the convergent validity of the EBTNS was confirmed. Besides, the levels of the AVE of each sub-construct are greater than the squared correlation involving the sub-constructs (Table 5). This result confirmed the discriminant validity of the EBTNS.

Table 2. Results of the second-order CFA on EBE and 5 subconstructs.

Table 3. Effect sizes of 5 sub-constructs, AVE, and the CR of EBE.

Table 4. Factor loadings of measurement items for each sub-construct, AVE, and the CR of 5 sub-constructs.

Table 5. Convergent validity and discriminant validity.

In summary, the construct validity of the EBTNS was supported by CFA, convergent validity, and discriminant validity.

5. Discussion

The researchers tested the hypothesized model of EBE to validate 5 sub-constructs and 37 indicators by using CFA. The fit indices of the CFA results confirmed that the overall model fit the data well. All 37 first-order indicators and 5 second-order sub-constructs showed significantly standardized factor loadings and effect sizes. The effect size values of those sub-constructs were substantial. In addition, the factor loading values of indicators of each sub-construct were moderate to significant. These confirm that 5 sub-constructs and 37 indicators can measure EBE. Furthermore, the developed instrument attained both convergent validity and discriminant validity. According to empirical findings, the EBTNS has strong construct validity.30,31The EBTNS is a well-developed instrument to measure EBE in Thai nurses. The researchers developed this instrument through a comprehensive literature review and document analysis. The EBTNS demonstrates that it can accurately delineate EBE’s constructed definition and measurement. Therefore, these provide a clear operational definition of EBE covering every dimension in the nursing profession and enhance the construct validity of the developed instrument.31

The EBTNS has good reliability. The values of internal consistency reliability and construct reliability are satisfied because the researchers synthesized the 5 sub-constructs and 37 indicators from existing research instruments following the Nursing and Midwifery Professional Ethics Regulations, B. E. 2550.16These reflect the crucial aspects of the EBE of Thai nurses. Moreover, each sub-construct item represents the essential attributes of EBE appropriate to use in the Thai context, resulting in enhanced instrument reliability.31

The effect sizes of 5 sub-constructs were substantial. Also, the factor loading values of the indicators of each sub-construct were moderate to significant. These reflect the importance of 5 sub-constructs and 37 indicators to measure EBE. The International Council for Nurses (ICN) developed the ICN code of ethics. There are 4 principal elements of ethical standards: nurses and people, nurses and practice, nurses and the profession, and nurses and co-workers.1The PFN and the PHR subconstructs of the EBTNS are congruent with nurses and the professional elements of the ICN ethical standards. They represent the behaviors of nurses that promote the reputation of the nursing profession and develop research-based knowledge to support evidence-based practice. The PAN sub-construct of the EBTNS is consistent with nurses and practice elements of the ICN ethical standards, which reflect the behaviors of nurses that provide nursing care and midwifery to patients or clients according to professional standards. The TPC sub-construct of the EBTNS complies with nurses and people elements of the ICN ethical standards, which constitutes the behaviors of nurses that support the ethical and legal PAN and midwifery for patients or clients. The TCB subconstruct of the EBTNS follows the ICN ethical standards for nurses and co-worker elements. It demonstrates the behaviors of nurses that sustain a collaborative and respectful relationship with co-workers in nursing and other fields for the benefit of patients or clients.

The results of this study revealed that the construct of the EBE of Thai nurses was proper for validating the EBTNS.

6. Conclusions

The EBTNS will be useful for measuring and evaluating the EBE of Thai nurses. Nurse administrators can use it to assess and monitor nurses’ EBE to prevent ethical problems. In addition, they can develop strategies to maintain the good EBE of Thai nurses. Finally, this scale may also have applicability to other Asian nurses who have similar cultures.

Limitations

The psychometric properties of the EBTNS in this research were investigated using empirical data in this study. It is necessary to further examine the psychometric properties of this scale after monitoring its use for a while. Future research should examine the psychometric properties of the EBTNS in terms of validity, reliability, objectivity, and feasibility. At last, the EBTNS should be tested and used with nurses in other Asian countries.

Ethics approval

This research was approved by the Institutional Review Board, Faculty of Medicine Ramathibodi Hospital, Mahidol University. The research objectives, procedures, potential risks and benefits, protection of confidentiality, and the right to withdraw from the study were declared individually to all participants before they signed the consent form.

Conflicts of interest

All contributing authors declare no conflicts of interest.