Translation and Validation of the Family-Focused Mental Health Practice Questionnaire-Japanese Version

: Children of parents with mental illness are more likely to experience mental health concerns, compared to other children in the community. Thus, it is important mental health professionals identify and intervene appropriately in these families. To develop professional development activities, practice guidelines and organizational supports, it is important to first identify those practices that professionals employ and do not employ. Accordingly, a Japanese version of the Family-Focused Mental Health Practice Questionnaire was developed. In the first instance, the Questionnaire was translated into Japanese and disseminated to mental health workers in various Japanese settings. Based on 212 participants’ data, exploratory factor analysis was conducted to examine the structure of the Family-Focused Mental Health Practice Questionnaire Japanese version. Questionnaire results from professionals with and without previous family-focused/child-focused practice training was also compared. The Kaiser-Meyer-Olkin (KMO = 0.85) index and Bartlett’s test (p < 0.001) showed acceptable results. After Promax rotation, 13 factors with eigenvalues greater than 1 remained. Furthermore, the scree plot suggested a 13-factor solution as the best fit for the data. The factors accounted for 66.7% of the variance and most of the factor loadings were higher than 0.30, and most of the variables were uniquely loaded in one factor. Thus, thirteen subscales emerged from exploratory factor analysis. Comparing scores of each 13-subscale between two groups, there were significant differences on four (family-focused practice) and eight (child-focused practice) subscales. The Family-Focused Mental Health Practice Questionnaire Japanese version has good to acceptable construct validity. Further longitudinal design studies are required to explore the utility of the Family-Focused Mental Health Practice Questionnaire Japanese version in clinical settings. These results demonstrate the cross-cultural validation and psychometric properties of the Japanese version of the Family-Focused Mental Health Practice Questionnaire (FFMHPQ-JV). The original items for the FFMHPQ were based on a systematic review of the literature [21] and developed with input from a highly specialized workforce FaPMI (Families where a Parent has a Mental Illness) coordinators in Victoria, Australia [16]. In this study, the FFMHPQ-JV was shown to have sufficient face and content validity. Moreover, subscales of the FFMHPQ-JV’s


Introduction
Children of parents with mental illness are amongst the most vulnerable in the community. Reupert et al. [1] found that these children are at a higher risk of developing behavioral, developmental, and emotional difficulties, compared to their same-age peers. Likewise, a meta-analysis found that children of parents with severe mental illness are at increased risk for a range of psychiatric disorders, and one-third of them may develop a severe mental illness by early adulthood [2]. Qualitative research has shown that some children do not understand what is happening with their parent, and as a result are often confused and frightened [3].
Due to the stigma associated with parental mental illness they may become socially isolated and withdrawn [3]. For some, the caring responsibilities they assume for their parent and/or siblings may be onerous and adversely impact educational outcomes [3].
The Kawasaki City survey indicated that approximately 25 percent of people with mental illness had at least one child [4]. Another report found that the numbers of pregnant women with mental illness has been increasing [5]. According to the Patient Survey of the Ministry of Health, Labour and Welfare in Japan, the total number of persons with mental illness was approximately four million, and approximately 50% of them were 25 to 54 years old and thus of parenting age [6]. In Australia, prevalence rates suggest that 21-23% of children have a parent with a mental illness [7] and a recent systematic review showed that between 12 to 45% of adult mental health clients are parents caring for dependent children [8]. The prevalence of mental illness among parents of children attending child psychiatry clinics has been estimated to be 36% for mothers and 33% for fathers [9].
Given the high prevalence and needs of this cohort, it is important that healthcare professionals in Japan provide support to parents with mental illnesses and their dependent children. A national study in Japan found that almost all of the professionals (i.e., 581 of 584 [99.5%]) who have clients who are parents with mental illnesses recognized that it was necessary to support their clients' children [10]. Moreover, 75% of the professionals who supported parents with mental illnesses and/or their dependent children had undergone consultations with the parents about their dependent children's care and had supported their parenting [10]. Similarly, 78% of the professionals had assessed the parenting circumstances of the parents who were their clients [10]. Though these results might indicate that Japanese professionals support these families as part of usual practice, 82% of the professionals indicated that they had limited experience in this area and needed more support and training [10]. In order to develop appropriate professional development programs and systems to support Japanese health professionals, it is important to identify what Japanese professionals do (and do not do) when they work with parents and children, and whether these practices differ depending on discipline and workforce setting, as others have shown [11,12].
Moreover, family focused practice has been shown to vary according to workforce setting, clinician background and prior training. For example, Irish community-based nurses were found to be more family oriented than nurses in inpatient psychiatry settings [13] and social workers, compared to psychiatric nurses and psychologists, have been found to be more family focused [14]. In Thailand, nurses with training in family and/or child focused areas were found to be more family focused than their untrained colleagues [15]. Tungpunkom et al. [15] found that those with family/child-oriented training were more confident and skilled and more likely to provide support to carers and children and promote family connectedness.
Maybery et al. [16] developed the Family-Focused Mental Health Practice Questionnaire (FFMHPQ). The Questionnaire aims to measure all aspects of family-focused practice, including workplace, worker, and patient and family, that might influence family-focused practices [16]. In recognition of the need to acknowledge different language, culture and policy differences across countries, the FFMHPQ has been modified for use in Ireland, including psychometric evaluation by exploratory factor analysis [13,17]. The FFMHPQ was also translated into several other languages in Thailand [15] and in Norway [18].The aim of this study is to develop the Japanese version of the FFMHPQ, which can be used to identify the family focused practices of the Japanese health workforce. Results from the FFMHPQ can be used to inform and evaluate professional development efforts, practice benchmarks and organizational improvement.

Data Collection
Ethical approval to conduct the present study was provided by the first researcher's university committee. The recruitment method involved distributing explanatory statements about the study to the managers of psychiatric hospitals, public health centers, child guidance centers, public mental health and welfare centers. In addition, convenient and the snowball sampling methods were employed, by asking those that had completed the questionnaire to forward the link to others that may be interested. The explanation sheet included an URL and a QR (quick response) code to access the online questionnaire, through SurveyMonkey.

Participants
To determine the sample size, we conducted a power calculation based on the average standard deviation in an Australian sample with 90% power to detect a difference in groups with a level of significance of 0.05 for a two-sample test on mean scores [13,16]. A sample size of N = 200 was found to be sufficient to achieve the aims of the study and provide adequate power to determine FFMHPQ scores.
In total, 394 mental health workers responded to our questionnaire on SurveyMonkey between June and November 2019. Due to missing data, we ultimately used the data from 212 participants. There were 162 (76.4%) females and 49 (23.1%) males (with one not specifying gender). Participants had an average age of 41.63 years (ranging from 23 to 63 years). Most participants were public health nurses (n = 68, 32.1%) followed by psychiatric nurses (n = 63, 29.7%), social welfare professionals (n = 52, 25.5%) and psychologists (n = 18, 8.5%). The remaining participants were doctors, a nurse, other types of welfare officers, occupational therapists, or dieticians. Most participants had either a degree (n = 128, 60.4%) or a graduate diploma/associate degree (n = 82, 38.7%). The average length of time that participants had been working in their current role was 8.52 years (ranging from 0 to 36 years). Those who had worked as an adult clinician recorded an average length of time working of 8.72 years (0-35 years), with children 5.1 years (0-25 years), and as a child/adolescent clinician for 5.27 years (0-25 years). Seventy-seven participants (36.3%) had previously attended family-focused training, while 92 participants (43.4%) had attended child-focused training. In piloting questionnaire, we had asked participants to indicate the type of previous training. Previous family-focused training included family psychoeducation, clinical supervision, group art therapy, group behavioral therapy, the signs of safety approach, parent training, and family therapy. Previous child-focused training included play therapy, behavioral therapy, educational therapy, and clinical supervision as well.

The Family-Focused Mental Health Practice Questionnaire
The FFMHPQ was developed by Maybery et al. [16] and consists of 53 items, answered on a sevenpoint Likert scale (ranging from strongly disagree (1 point) to strongly agree (7 point)). The FFMHPQ has excellent face, content, and construct validity and good reliability, with Cronbach's alpha coefficients ranging between 0.70-0.90, while Maybery et al. noted that 3 of 14 subscales of the FFMHPQ had inadequate reliabilities, and care should be taken when using those subscales [15,16]. The FFMHPQ has 18 subscales; a low score in a particular subscale suggests reduced family-focused practices, while a high score indicates an increased family focus [13].
Following guidelines established by Beaton et al. [19], we translated the FFMHPQ into Japanese and developed a FFMHPQ-Japanese version (FFMHPQ-JV). First, two translators with Japanese as their first language independently translated the FFMHPQ from English to Japanese and then reached consensus on a combined translation. Next, two translators with English as their first language independently backtranslated the questionnaire. We consulted the developers of the original FFMHPQ, and we eventually reached a consensus on the pre-final version. Before data were collected from a larger sample, a pilot study of the pre-final version was conducted with 30 Japanese mental health professionals (not included in the present study). We interviewed all 30 of the participants to ensure that the pre-final version was appropriate. Their feedback was incorporated after which we consulted with the developers again, and reached a consensus on the final version of the FFMHPQ-JV.

Data analysis
Exploratory factor analysis (EFA) was conducted to examine the structure of the FFMHPQ-JV. Independent t-tests were undertaken to examine the differences between the group (i.e., with and without previous family-focused practice training/with and without child-focused practice training) means. Effect size was also calculated to show the sizes of differences between the groups. In addition, we calculated Cronbach's alpha coefficients of each subscale for reliability and compared each subscale's score between participants with and without previous training for determining known-groups validity.

Results
The structure of the FFMHPQ-JV was initially explored by a maximum likelihood factor analysis with Promax rotation and Kaiser normalization. The Kaiser-Meyer-Olkin (KMO = 0.85) index and Bartlett's test (p < 0.001) showed acceptable results. After Promax rotation, 13 factors with eigenvalues greater than 1 remained. Furthermore, the scree plot suggested a 13-factor solution as the best fit for the data. Thirteen factors accounted for 66.7% of the variance. Most of the factor loadings were higher than 0.30, and most of the variables were uniquely loaded in one factor. The factor loadings of Q4 (Government policy regarding family focused practice is very clear) was lower than 0.3; hence, we determined that they should be used independently (Tab. 1). Referring to the original FFMHPQ subscales' names, we named 13 subscales after items contributing to each subscale as follows: Factor 1: "Worker confidence with skill and knowledge"; Factor 2: "Support for profession from workplace and coworkers"; Factor 3: "Training"; Factor 4: "Referrals to other programs/services"; Factor 5: "Parenting and child-focused support"; Factor 6: "Assessing parents' awareness of child connectedness and impact"; Factor 7: "Team-working"; Factor 8: "Engagement with family members"; Factor 9: "Family support"; Factor 10: "Time and workload"; Factor 11: "Support to carers and/or children"; Factor 12: "Support to consumers and their families"; and Factor 13: "Parenting and mental illness." We therefore used the subscales that emerged from EFA, and Q4 (Government policy regarding family focused practice is very clear) to make subsequent calculations. Tab. 2 shows the descriptive statistics, significant differences, and effect sizes for those who did/did not receive previous family-or child-focused training. According to Sawilowsky [20], Cohen's d can be expounded as very small (0.01), small (0.20), medium (0.50), large (0.80), very large (1.20) or huge (2.0). In family-focused practice training, the participants who had received previous training scored higher on the following FFMHPQ-JV subscales than those who did not: "Worker confidence with skills and knowledge," "Referrals to other programs/services," and "Support to consumers and their families." The participants without previous training showed higher scores than those with previous training on the subscale of "Parenting and mental illness." In "Parenting and mental illness," lower scores mean better family-focused practices. In child-focused practice training, the participants with previous training scored higher on the following FFMHPQ-JV subscales than did those without previous training: "Worker confidence with skills and knowledge," "Support for profession from workplace and coworkers," "Training," "Referrals to other programs/services," "Parenting and child-focused support," "Assessing parents' awareness of child connectedness and impact," and "Support to consumers and their families." On the other hand, as in family-focused practice training, the participants without previous training showed higher scourers than their counterparts on "Parenting and mental illness." As a whole, Japanese mental health professionals' family-focused practice is almost the same as that in Australia. The FFMHPQ-JV consists of 13 subscales, which is five less than the number of original FFMHPQ subscales. However, for example, one of the FFMHPQ-JV subscales, "Worker confidence with skill and knowledge," is a combination of "Worker confidence" and "Skill and knowledge," which are subscales of the original FFMHPQ. As such, it can be said that the components of the FFMHPQ-JV adopt the original FFMHPQ subscales. As mentioned above, the FFMHPQ has sufficient validity, and the FFMHPQ-JV has good validity. In addition, most of the subscales' internal consistencies are high enough (e.g., a Cronbach's alpha coefficient of 0.86) to be acceptable (e.g., higher than 0.60), although 4 of 13 subscales should carefully be interpreted (i.e., engagement with family members (Cronbach's alpha coefficient of 0.48), family support (0.47), support to carers and/or children (0.49), and support to consumers and their families (0.44)).

Discussion
These results demonstrate the cross-cultural validation and psychometric properties of the Japanese version of the Family-Focused Mental Health Practice Questionnaire (FFMHPQ-JV). The original items for the FFMHPQ were based on a systematic review of the literature [21] and developed with input from a highly specialized workforce FaPMI (Families where a Parent has a Mental Illness) coordinators in Victoria, Australia [16]. In this study, the FFMHPQ-JV was shown to have sufficient face and content validity. Moreover, subscales of the FFMHPQ-JV's were slightly different from the original FFMHPQ subscales. We found 13 subscales, while the original FFMHPQ consisted of 18 subscales. For example, one of the FFMHPQ-JV subscales was "Worker confidence, skill and knowledge," which consisted of seven items, and is a combination of two original subscales (i.e., "Worker confidence," which consists of three items, and "Skill and knowledge," which consists of five items). Overall, the FFMHPQ-JV shows good to acceptable construct validity and accordingly might be used to inform and evaluate different professional development activities, benchmarking services and organizational initiatives.
On the other hand, care needs to be undertaken regarding the reliability of the FFMHPQ-JV. Leonard et al. [17] found that since the development of the original scale, there have been several attempts to adapt the FFMHPQ. Some studies have adapted the FFMHPQ without attempting to validate their changes [15,22], while those that tested internal reliabilities revealed many of these to be unacceptable [13,23]. van Doesum et al. [24] found that although the FFMHPQ had been proven to be valid and has a reasonable reliability, the reliability was poorer in all the subscales in their study, and recommend that in future research using the FFMHPQ is necessary to improve the reliability of the weaker items. In the current study, notably, 4 of 13 subscales had lower than Cronbach's alpha coefficient value of 0.60. In some literatures it is reported that reliability is "acceptable" if Cronbach's alpha is greater than 0.60 or 0.70 but it is recommended that care should be taken when using these subscales of the FFMHPQ-JV. Those subscales' interitem Pearson coefficients were as follows; engagement with family members (3 items; r = 0.213 to 0.272), family support (2 items, r = 0.304), support to carers and/or children (3 items, r = 0.180 to 0.305) and support to consumers and their families (3 items, r = 0.124 to 0.349). While some coefficients were below 0.20 generally, interitem correlation considered as an acceptable value when it in the range 0.20-0.40. This explains why some Cronbach's alpha coefficients were below 0.60.
In addition, Leonard et al. [17] conducted EFA twice and they eventually concluded the FFMHPQ consisted of 20 items and two factors. Their findings indicated that the FFMHPQ might be modified as such. However, an alternative explanation is that a two factor structure might represent higher order conceptual measurement and not the lower order structure that gives greater insight and detail regarding specific family focused practices in mental health services. As a consequence, it was thought important to Japanese version of the FFMHPQ-JV in order to build up and improve specific Japanese family-focused mental health practices. Thus, we would like to use the FFMHPQ-JV by respecting and using all the original items regardless of each item's psychometric property. In future, if our family-focused practice would have progress, we could modify the questionnaire to adapt Japanese professional needs and culture in more detail.
Compared to professionals with no training in this area, Japanese professionals with previous familyfocused/child-focused practice training reported higher confidence in both family focused skill and knowledge and indicated that they provided more referrals of children and parents to other programs/services. Likewise, Tungpunkom et al. found that clinicians with previous training experiences in this area reported more family-focused practices than those without previous training [15]. It is difficult for mental health professionals to talk with and support parents who have a mental illness and training appears to be helpful in this regard. The difficulties in this work may be due in part because parents themselves may be reluctant to discuss their children. Ueno et al. found that Japanese with chronic mental illness felt remorseful about how their illness impacted upon children [25]. Other research has shown that because of the fear about child protection, parents are hesitant to talk about their children and to ask for help [26,27]. Hence professional development is important in this field. However, notwithstanding the importance of professional development in this area, training of itself is not sufficient in promoting ongoing, sustained family focused practice. Legislation, ongoing supervision, managerial support and appropriate infrastructure (e.g., identification of parenting status at intake) are other initiatives needed to embed family focused practice into services [28].
There is Japanese mental health related legislation (i.e., Act on Mental Health and Welfare for the Mentally Disabled) and child-focused legislation (e.g., the Child Welfare Act, Act on Child Abuse Prevention). However, there is no legislation concerned with working with families were parents have mental health issues. Thus, for Japanese professionals, Q4 (Government policy regarding family focused practice is very clear) might be difficult to answer. Conversely, other countries do have legislation in this area, for example, in Norway, clinicians are mandated to work with the children of the clients of mental health services [29]. Likewise, in Finland, a nationwide program for parents with mental illnesses and their children, the Effective Child & Family Programme, was launched in 2001 and was supported by the Ministry of Social Affairs and Health [30]. At the same time, legislation is not sufficient to prompt broad scale organisational change; Tchernegovski et al. [31] found that even though a Victorian (in Australia) policy stipulated the needs of children in these families, this alone was not enough to ensure the needs of these families were met. Thus an instrument such as the FFMHPQ-JV may play an important role in benchmarking practice, the monitoring and evaluating of services, and developing professional development programs and in this way inform government policy and organisational procedures.

Limitations
The FFMHPQ-JV had good validity, while care should be taken in some of the subscales due to lack of enough internal consistency. Although the participants' professional composition was similar to those in the previous Australian [16] and Thai studies [15], most of the participants were public health nurses (n = 68, 32.1%) or psychiatric nurses (n = 64, 30.2%); the remaining numbers of professional groups were smaller (e.g., 18 psychologists, 8.5%). Further research is needed to analyze potential differences in professional discipline that might influence family-focused practices. Further longitudinal design studies are required on the FFMHPQ-JV's utility in the clinical setting. Finally, the FFMHPQ is a self-report questionnaire from the perspective of professionals. How their managers and arguably more importantly, how clients and their families perceive these services, remains to be seen.

Conclusions
The Family-Focused Mental Health Practice Questionnaire Japanese Version (FFMHPQ-JV) was developed following cross cultural adaptation guidelines from Beaton et al. [19]. In the present study, acceptable validity and reliability of the FFMHPQ-JV was found, while some of the subscales' internal consistencies might have been unacceptable. Care should be taken when using these subscales and improving the reliability of the weaker items should be guaranteed in future. The validated questionnaire can now play an important role in benchmarking services, developing and evaluating professional development programs and providing guidance as to the types of supports the Japanese workforce requires.