Statistical Validation of the Self-Harm Antipathy Scale-Japanese Version (SHAS-J)

Aim: This study aimed to determine the reliability and validity of the Self-Harm Antipathy Scale-Japanese Version (SHAS-J). Methods: A self-administered questionnaire was distributed to 764 nurses working in 32 emergency departments across Japan and 302 (39.5%) of them responded effectively. The questionnaires collected demographic data about the nurses and the SHAS-J. Data were evaluated with Cronbach’s alpha along with exploratory factor, confirmatory factor, and correlation analyses. Results: Factor analysis of the SHAS-J resulted in extraction of four factors. The four factors comprised “low empathic practice competence”, “care futility”, “lack of active understanding” and “ignorance about rights and responsibilities”. Cronbach’s alpha for the four factors were 0.83-0.54. Conclusion: The reliability and validity of the SHAS-J were approximately verified.


Background
Suicide rates in Japan are high compared with other countries [1] and have become a serious social problem. Patients who attempt suicide require care-based sympathy from a nurse [2]. However, caring for patients who attempt suicide or self-harm is difficult for nurses working in an emergency setting, and this situation is not beneficial to patients [3,4]. In Japan, no measures are currently in place to evaluate negative feelings such as antipathy in patients who engage in selfinjury. Therefore, to help evaluate the difficulties associated with such patients encountered by nurses, a reliable Japanese version of the Self-Harm Antipathy Scale (SHAS) developed in the U.K. is needed.

Aim
This study aimed to determine the reliability and validity of the SHAS-Japanese Version (SHAS-J).

Method
An anonymous self-report questionnaire survey was conducted on 764 nurses working in emergency departments in Japan. The questionnaires collected demographic data about the nurses and the SHAS-J. To assess the concept validity of the SHAS-J, exploratory factor analysis was used to verify the factor structure, and covariance structure analysis using confirmatory factor analysis was used to verify the factor structure between the original SHAS and the SHAS-J and confirm the conformance. Data were analyzed using SPSS (Ver. 22 for Windows) and Amos (Ver. 22 for Windows). In translation to Japanese of the original SHAS, the original author Patterson was already retired and the authorship Whittington received license and translation permission. The original version of the SHAS was translated into Japanese by a certified translator. This study was approved by the ethics committee of Hamamatsu University School of Medicine (E 14-240) [5].

Results
The questionnaire was distributed to 764 nurses working in 32 emergency departments across Japan; 302 (39.5%) responses were received. The mean age of the respondents was 34.8 years (standard deviation=7.4 years). As shown in Table 1, the largest number of respondents were members of the High care units (73 persons; 24.2%), followed by the Intensive care units (65 persons; 21.5%).

Validation of reliability
Cronbach's α coefficient for each SHAS-J factor ranged from 0.83 to 0.54 ( Table 2).

Validation of validity
Regarding the 30 items on the original SHAS, the ceiling and floor effects, Item-Total correlation, and inter-item correlation were confirmed, but it was decided that the content of the item was not excluded considering the contents of the item ( Table 2). In the same way as the original SHAS, a factor analysis was performed using the varimax rotation, and four factors were extracted. After excluding six items with a factor loading of 0.35 or less were excluded (Items 3, 13, 18, 19, 20 and 25), the resulting scale was named the SHAS-J ( Table 3). The first and third factors were reverse-scored items. The first factor consisted of seven items, such as "I find it rewarding to care for selfharm patients" and "I try to help self-harm patients feel positive about them", so it was named "Low Empathic practice competence". The second factor consisted of 11 items, such as "A self-harm patient is someone who is only trying to get attention" and "A self-harm patient is a complete waste of time", so it was named "Care futility". The third factor consisted of three items, such as "Acts of self-harm are a form of communication about their situation" and "For some individuals, selfharm can be a way of relieving tension", so it was named "Lack of active understanding". The fourth factor consisted of three items, such as "People should be allowed to engage in self-harm in a safe environment" and "An individual has the right to engage in self-harm", so it was named "Rights and responsibilities". Then, to examine the validity of the constitutive concept, the factor structure of the SHAS-J was analyzed. In terms of compatibility, for the four-factor model of the SHAS-J, the GFI=0.85, AGFI=0.82, CFI=0.37 and RMSEA=0.06. In terms of the compatibility of the six-factor model, which was assumed to have the same factor structure as the original SHAS, the GFI=0.85, AGFI=0.81, CFI=0.21 and RMSEA=0.07.

S. No.
Ceiling effect Floor effect Item-total correlation Inter-item correlation  The result of principal factor method Varimax rotation was shown The enclosure of numbers shows the factor loading with the highest factor Table 3: Resurt of factor analysis of "SHAH-J".

Discussion
Cronbach's α coefficient for each factor of the SHAS-J ranged from 0.83 to 0.54, which is the same as that for the original SHAS, thereby confirming the internal consistency of the SHAS-J. The reliability of the SHAS-J was also confirmed using Cronbach's α. The results of the exploratory factor analysis showed that the SHAS-J had a four-factor structure and was not completely consistent with the six-factor structure of the original SHAS. The reason for the difference in the factor structure between the original SHAS and the SHAS-J was thought to be the cultural differences between Japan and the U.K. in terms of religion and individual rights and responsibilities. However, although the SHAS-J has a different factor structure than does the original SHAS, the measurable content of both scales are considered similar. These results suggest that the SHAS-J has good reliability and validity, but back-translation should be carried out the in future.
It is important that nurses empathize with their patients, but it is difficult for nurses do it [2,4]. Rogers reported that to sense the client's private world as if it were your own, but without ever losing the "as if " quality -this is empathy and this seems essential to therapy. Item of "Lack of active understanding" is indicated "empathy" [6] and considered that it is significant attitudes in understanding self-harm patients.