Psychometric properties of the Persian version of the Second Victim Experience and Support Instrument

Abstract Aim This study was designed to characterize the psychometric properties of the Persian version of the Second Victim Experience and Support Instrument (P‐SVEST). Design This study was a methodological and cross‐sectional study. Methods The SVEST was back‐translated into Persian and 10 experts assessed its content validity. Construct validity was determined through exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) with a total of 754 critical care and emergency nurses. Results The results of exploratory factor analysis showed that the P‐SVEST had four factors. These four factors accounted for 51.67% of the total variance. Also, these factors were confirmed by confirmatory factor analysis (root mean square error of approximation = (90%. confidence interval) = 0.058 [0.045, 0.071], goodness‐of‐fit index = 0.932, comparative fit index = 0.956, non‐normal fit index = 0.918, incremental fit index = 0.957 and Tucker–Lewis index = 0.944). Coefficients of Cronbach's alpha, McDonald's omega, composite reliability and maximum reliability for all of the factors were >0.7, demonstrating satisfied internal consistency.


| INTRODUC TI ON
Health care professionals face stressful situations and a high workload all the time. These working conditions can be frustrating for staff and increase medical errors (Griffiths et al., 2020;Sturm et al., 2019).
Medical errors and adverse events are inevitable in any health care organization. These issues often have catastrophic effects on the patient, health care providers and the organization. Although most health care professionals experience medical errors (Khammarnia et al., 2021;Mosadeghrad et al., 2020), nurses are more prone to medical errors due to their high workload (Di Muzio et al., 2019).
Outcomes of medical errors and the patient have adverse effects on health care professionals (Cabilan & Kynoch, 2017;Finney et al., 2021). Following a medical error, the patients and their families are considered the first victims, and the health care professionals who are injured after the incident are the second victims (Busch et al., 2020). The feeling of health and well-being is affected in the second victim after the safety events. The victims experience psychological and physical distress and endure a high level of stress (Chan et al., 2017;Ozeke et al., 2019;Miller et al., 2019;Van Gerven, Deweer, et al., 2016;Van Gerven, Vander Elst, et al., 2016). After safety incidents, adverse outcomes in health care professionals include physiological disorders, sleep disorders, occupational dysfunction, burnout, powerlessness, decreased job satisfaction, decreased job confidence, guilt, anger and shame, anxiety about punishment, job loss and litigation (Bari et al., 2016;Garrouste-Orgeas et al., 2015;Joesten et al., 2015;Lee et al., 2019;Miller et al., 2019;Mohsenpour et al., 2017;Tawfik et al., 2018;Van Gerven, Deweer, et al., 2016;Van Gerven, Vander Elst, et al., 2016). Health care professionals who feel like a second victims need support (Schrøder et al., 2019).
However, organizations' support resources are often insufficient to prevent and reduce work-related injuries. Lack of organizational support leads to reduced patient safety and consequently, patient injury and injury to health care providers (Farokhzadian et al., 2018;Rinaldi et al., 2016). The second victim's mental, physical and occupational distress reflects a culture of punitive safety and a lack of organizational support (Burlison et al., 2017;Quillivan et al., 2016).
It is essential to examine the experiences of people involved in safety incidents with a valid and reliable instrument. Using a valid and reliable instrument can help implement programmes to increase support for second victims and increase patient safety. The Second Victim Experience and Support Instrument (SVEST) was originally developed by Burlison et al. (2017). This instrument measures health care professionals' distress after an error occurs and assesses the quality of support resources and the outcomes for staff. The instrument has been validated in different countries (Brunelli et al., 2021;Chen et al., 2019;Kim et al., 2020;Knudsen et al., 2021;Koca et al., 2022;Mohd Kamaruzaman et al., 2022;Santana-Domínguez et al., 2021Scarpis et al., 2021). Ajoudani et al. also evaluated the psychometric properties of this instrument in Urmia, Iran, on 298 nurses working in general wards. In the study of Ajoudani et al. (2021), exploratory factor analysis (EFA) was not performed and only confirmatory factor analysis (CFA) was performed. Nurses in intensive care units (ICUs) and emergency departments (EDs) make more errors than nurses in general wards, so they experience more of the second victim experience. On the other hand, it seems necessary to perform EFA to find latent variables in this instrument based on Iranian culture. Therefore, this study was designed to assess the psychometric properties of the Persian version of the Second Victim Experience and Support Instrument (P-SVEST).

| Research design
This methodological cross-sectional study was conducted on the critical care and emergency nurses from September to October 2021 to evaluate validity and reliability of the P-SVEST. The inclusion criteria of the participants in this study were willingness to participate in the study, employment in critical care units and emergency departments and having 2 years of work experience and medical error experience.

| Measures
There were two sections in the instrument survey. The first section consisted of items regarding participants' profiles, such as age, gender and workplace unit. In the second section, the 29-item SVEST was used to measure the feeling of the second victim and the support received by the participants (Burlison et al., 2017). Items were measured on five-point Likert scales, with scores ranging from 1 "strongly disagree"to 5 "strongly agree". Also, seven support options were included in this instrument, and the desirability of support options was measured by items anchored on a five-point Likert scale ranging from 1 "strongly do not desire"to 5 "strongly desire".

| Procedure
Regarding the validity of translation process, the SVEST was translated based on the standards recommended in the guidelines (Beaton et al., 2002). Initially, written permission for the SVEST was obtained from the developer of the scale, Dr. James M. Hoffman, via email.
Subsequently, based on the forward-backward translation technique, two English-Persian translators were asked to translate the SVEST into Persian. The two translators translated the SVEST into the Persian version independently. Afterwards, these two Persian versions of the SVEST were reviewed and commented on by a group of experts, including some authors of this article and another two professional translators, to form a single P-SVEST. Finally, the single P-SVEST was back-translated to English by a Persian-English translator, and the translation accuracy was confirmed by a group of experts.

| Content validity
The face validity was obtained by providing the instrument to 20 nurses and asking them to identify the sentences and phrases that are vague. Content validity was performed by 10 experts (six faculty members from the department of intensive care nursing and emergency nursing and four nursing managers) who used the four-point method from 1 being irrelevant to 4 being highly relevant to score each item of the P-SVEST. These experts evaluated the content of the P-SVEST. Afterwards, the item-and scale-level content validity indexes (I-CVI and S-CVI, respectively) were calculated for the P-SVEST instrument. The acceptable ranges for I-CVI and S-CVI are >0.80 and >0.92, respectively (Polit & Yang, 2016). Content validity index correction of random chance agreement was made (Pa) using  Table 1).

| Construct validity and reliability
This study conducted both EFA and CFA to confirm the factorial structure and the construct validity and reliability of the P-SVEST.
The sample size should be at least 200 cases for factor analysis (Beaton et al., 2002). A total of 754 nurses were recruited via online data gathering. Online data gathering was performed for this section where the online questionnaire was created via Porsline and its URL link was sent by social networking applications such as Telegram channel or WhatsApp groups. Data were then extracted into an Excel file from the Porsline.
For data analysis, the dataset (n = 754) was randomly divided into two parts. The data were split randomly using: Data -> Select Cases and "Random sample of cases" was chosen. The first dataset (n = 377) was used to conduct EFA using SPSS version 27, and the second one (n = 377) was used to conduct CFA using AMOS version 27. This study employed maximum-likelihood EFA with Promax rotation, the Kaiser-Meyer-Olkin (KMO) > 0.8 and Bartlett's test of sphericity (p < 0.05) to assess the relevance and appropriateness of the data for conducting the factor analysis. The EFA was run by Pearson matrices.
The convergent and divergent validity of the P-SVEST were estimated using Fornell and Larcker's approach (Fornell & Larcker, 1981). For convergent validity, composite reliability (CR) should be higher than 0.7, and average variance extracted (AVE) should be >0.5 (Sharif Nia et al., 2015). This study also assessed the construct reliability over its internal consistency (Cronbach's alpha

| Multivariate normality and outliers
Both univariate and multivariate normality of the data was evaluated in this study. The univariate distributions were tested for outliers, skewness and kurtosis. Also, the multivariate normality was assessed using Mardia's coefficient of multivariate kurtosis and Mardia's coefficient. Mardia's coefficient of multivariate kurtosis <8 can be considered indicative of departure from multivariate normality (Henseler & Fassott, 2010). Moreover, the outliers of the multivariate were detected using Mahalanobis distance (p < 0.001; Leys et al., 2018).

| Ethical considerations
This study was conducted after obtaining the code of the research ethics committee (IR.ZUMS.REC.1399.335) from Zanjan University of Medical Sciences. On the first page of the electronic instrument, the study's objectives were explained to the participants. If the participants were willing to participate in the survey, they would sign the electronic consent form and complete the instrument.

| Participants' profiles
In total, 754 Iranian nurses participated in this study, including 203 males and 552 females. Most of the participants were in the age group of 31-40 years with 6-10 years of work experience (Table 2). TA B L E 1 Content validity analysis by item.  Next, maximum-likelihood CFA (n = 377) was conducted to validate the factorial structure obtained from EFA. As shown in Figure 1, to improve the model, three pairs of measurement error were allowed to co-vary freely (i.e. e 3 to e 4 , e 5 to e 6 and e 11 to e 12 ). MaxR for all of the factors were >0.7, indicating the satisfied internal consistency and construct reliability. Moreover, the AVE for three factors was less than the required threshold of 0.5 or MSV, and AVE is a strict measurement for convergent validity. CR more than 0.7 can be used to assess convergent validity in psychological studies.

| Validity and reliability
Therefore, the CR convergent validity achieved in this study was >0.7 for both factors (Table 4).

| DISCUSS ION
This study aimed to determine the psychometric properties of the P-SVEST instrument. Using EFA, four factors were extracted in the P-SVEST instrument, which explained 51.67% of the total variance.
Based on the Iranian culture, P-SVEST identified four factors: physical distress, psychological distress and institute support with four items each and professional self-efficacy factor with three items.
Eventually, the CFA results showed that the psychometric properties of P-SVEST had good validity and reliability. In

TA B L E 3
The result of EFA and internal consistency on the four factors of the P-SVEST (N = 377). of repeated errors and an increase in patient safety. In the present study, this factor was called hospital support with three items (Farokhzadian et al., 2018;Rinaldi et al., 2016;Schrøder et al., 2019).
The feeling of professional inefficiency after an error is associated with feelings such as occupational dysfunction, burnout, powerlessness, decreased job satisfaction and decreased job confidence which was conducted on 298 nurses in Iran, confirmed seven factors and two outcomes by conducting CFA. The participants in the study of these researchers were nurses working in general wards, also, EFA was not conducted to find latent variables (Ajoudani et al., 2021). In Tan et al.'s study,21 items in four factors were extracted using the maximum-likelihood and Promax rotation methods (Tan et al., 2020).
In the study of Strametz et al. (2021), 11 factors were extracted using F I G U R E 1 Factor structure of the P-SVEST with correlations among the five factors, standardized factor loadings, and error terms.  -Domínguez et al., 2022). In some psychometric studies of this instrument, EFA was not performed and only CFA was applied.
Also, some studies have tested the dimensions of the original version (Ajoudani et al., 2021;Knudsen et al., 2021;Pieretti et al., 2022). In our study, unlike the original version, physical distress was included in factor 1 and psychotic distress was included in factor 2. In addition, unlike the original version, in the study of Chen et al. (2019), factor 1 was allocated to institutional support and supervisor support, and factor 2 was allocated to self-efficacy and negative work outcomes occupational. This change in factors and items in studies reflects the cultural differences of different countries in the experience of feeling the second victim of staff.
According to this study results, most nurses who experience a patient's safety event in Iran reported physical and psychological distress. Consistent with the results of our study, in other studies, people experience these symptoms following a medical error (Zhang et al., 2019;Zheng et al., 2022). According to Iranian nurses, support resources after patient safety events are mostly related to the supervisor and the institution. Unlike previous studies, colleague support and non-work-related support have a negligible effect on improving the condition of feeling like a second victim (Chen et al., 2019;Quillivan et al., 2016;Santana-Domínguez et al., 2021). In Iranian culture, due to the lack of nurses, absenteeism is almost impossible. Therefore, contrary to previous studies, this factor was removed in our study (Brunelli et al., 2021;Burlison et al., 2017;Kim et al., 2020;Pieretti et al., 2022).
Since having a job is a basic need for Iranian citizens, nurses have to be present at work after these events, even in the worst physical, mental and professional self-efficacy conditions. Therefore, in these circumstances, the support of the supervisor and the institute creates more peace for the nurses than the support of colleagues and family members. Otherwise, they cannot ensure their job security.
In the present study, all factors with Cronbach's alpha >0.7 had good reliability. However, studying the psychometric of the Danish version of SVEST, Knudsen et al. (2021) showed that Cronbach's alpha did not have adequate reliability in the factors of colleague support (0.4) and institute support (0.68). Elsewhere, studying the original version (Burlison et al., 2017), the Korean SVEST version (Kim et al., 2020) and the Argentinian SVEST study (Brunelli et al., 2021) reported poor-to-questionable Cronbach's alpha coefficients for these two factors. In the present study, these two factors were used to remove the EFA phase.
The results of this study demonstrated that nurses in intensive care units and the emergency department endure psychological and physical distress after patient safety events, which is consistent with other studies (Quillivan et al., 2016;Zhang et al., 2019). However, the consequences of the second victim experience for Iranian nurses are different in some countries (Brunelli et al., 2021;Kim et al., 2020).
In this study, due to the removal of three factors and 14 items from the original instrument, the generalizability of the instrument is limited. On the other hand, because the data collection method was online, nurses from most provinces of Iran participated in this study.
Therefore, this instrument can be used to assess the experience of the second victim and nurses' support resources in critical care units and emergency departments of Iran. In addition, since the participants in this study only include the nursing community, the generalizability of the P-SVEST to other professional health care providers is limited.

| CON CLUS ION
The

ACK N O WLE D G E M ENTS
The researchers appreciate the financial support of the Vice Chancellor for Research of Zanjan University of Medical Sciences (A-11-148-26) and the participation of nurses in this study.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author.

E TH I C A L A PPROVA L
This study was conducted after obtaining the code of the re-