Social Work with Families in Special Distress: Collaborative Practices

: Collaborative practices have emerged as an e ﬀ ective approach for conducting social work interventions with families in special distress. This study aimed to ascertain the perspective of the social workers located in basic community social services, in relation to the development of a collaborative approach with families in special distress. The main objective was to ﬁnd out the level of importance and the level of implementation that participants (N = 121) gave to the di ﬀ erent intervention criteria included in an Inventory of Collaborative Practices. The results indicate that criteria related to basic issues in social case work (active listening, respect, and empathy) as well as the management and bureaucracy of the speciﬁc case are the most valued and performed by social workers. The least valued and performed criteria have to do with issues that involve reﬂective processes in the helping relationship, both with the family and with the rest of the professionals. Implications for practice and quality enhancement are discussed, as they are key aspects in the development of collaborative interventions in social work.


Introduction
Families in special distress were firstly defined by Minuchin in the nineteen sixties as families with particular characteristics of disorganization and pathology, who were also permanently involved with institutions, agencies, and services (Reder 1986). These families have been known as excluded families (Tierney 1976), multi-problem families (Selig 1976), multi-agency families (Reder 1986), and multi-stressed families (Escudero 2009;Madsen 2007). Families in special distress (FED) experience different problems simultaneously, affecting two or more members and different domains Tausendfreund et al. 2016). Moreover, the typical profile of these families includes poverty, inadequate living conditions, deteriorated functioning as partners and as parents, substance abuse, antisocial activities, and a lack of support systems (Bodden and Dekovic 2016;Chagas 2014;Imber-Black 2000;Shamai et al. 2003).
FED are often involved with social workers and other professionals of the helping process. Sharlin and Shamai (2000) highlight the importance of using different approaches with these families (strategic, solution-centered, and narrative practice, among others), to better address their objectives. FED usually develop relationships with formal systems that last for years, even for generations (Imber-Black 2000). This chronicity generates feelings of despair and hopelessness. A coalition of despair arises, which affects not only the family members but also the professionals that are intervening with them (Sharlin and Shamai 2000). Pannebakker et al. (2018) state that these families are often multi-users of psychosocial and health care services. The multiple services that are working with a FED tend to work in a fragmented way (Lipchik 2004;Sousa and Costa 2010;Sousa and Eusébio 2005). In fact, agencies are normally organised by areas of intervention, focusing on specific aspects, such as drug addiction, parental functioning, or mental health. Each professional often works in a different setting, and each service has its own mandates and rules. This pre-defined role constrains the possibilities of the intervention (Steens et al. 2018). This may result in the family receiving contradictory messages or the duplicity of interventions. Imber-Black (2000) put forward a set of questions so that social workers and families are able to reflect on the intervention and assess the relationships with larger systems. These questions address subjects such as the former and current larger systems involved, the problem defined by each agent, their myth and belief systems, and previously applied solutions. Elaborating an ecomap with the family is a technique that contributes to a conjoint reflection on the relationships with all the professionals and services involved, but it is also a way of pinpointing informal support and other significant people in the family's life. Moreover, finding out which agents (formal and informal) are related to the family becomes essential for developing a collaborative relationship (Imber-Black 2000).
Collaborative practices have emerged as an effective approach to developing social work interventions with families in special distress. Saar-Heiman et al. (2017) support the idea that FED can be engaged in a process of change, even if they are in extreme poverty conditions. In this line, Bachler et al. (2017) claims that FED have higher drop-out rates than other families. Collaborative practices arise as a way of actively involving both the members of the family system and the professionals that make up the helping macrosystem. Thus, social workers are able to develop these practices by honoring the expertise of the family system, respecting their wishes and preferences (Drisko 2017), and accompanying them in the process of change (Madsen 2007;White and Epston 1993). It is also fundamental to define, conjointly, the objectives and tasks, and with FED, "it is no easy task to establish a positive relation ( . . . ) (and) having a positive outlook to the future" (Bachler et al. 2017, p.143). Likewise, it is essential to develop a cooperative and collaborative system with the other professionals involved. In this way, the duplication of actions may be obviated, and we may contribute to the implementation of a holistic, integral intervention (Imber-Black 2000; Sousa and Costa 2010). Collaborative practices draw on different theoretical approaches, such as the narrative approach, solution-focused intervention, or strategic brief intervention. The common factors that explain change in socio-relational intervention are also fundamental in the development of a collaborative approach (Duncan 2002;Lambert 1992;Rosenzweig 1936). Building a helping alliance, engaging the members of the family system in the intervention process, generating positive expectations, and getting systematic feedback become the backbone for successful collaborative development. Jakob (2018, p. 29) highlights the "often extreme isolation of parents in multi-stressed families", and Goh (2017) explains the importance of establishing a working alliance with vulnerable families prior to the start of the intervention. Visscher et al. (2018) analyzed the type of interventions developed with FED in the Netherlands, and detailed eight categories: the assessment of problems, planning and evaluation, working on change, learning parenting skills, helping with concrete needs, activating the social network, and maintaining the practitioner-client collaboration. This last category includes collaborative practices, with interventions such as "talking about expectations" or "working on the quality of the relationship". Reupert and Maybery (2014) explain that professionals recognize the benefits of collaborative practices. However, they experience difficulties when networking with other professionals. The high rotation of staff within social services and the distrust of the family members regarding the exchange of information between institutions are significant challenges that social workers have to cope with in their daily practice. Other barriers are related to time constraints. Leigh and Miller (2004) state that a considerable caseload may prevent social workers from having enough time and space to reflect on each case. Moreover, the pressure those professionals face in order to prove the quality of their work (mostly defined by quantitative indicators) may compromise the alliance and the relationship with the family. Finally, we should bear in mind that working with FED often implies a combination of assistance and coercive interventions (Ng 2017). In fact, in social services settings in Mallorca (Spain), it is common that the social worker develops interventions in both contexts at the same time; they usually support the family and attend their economic needs, while they sometimes have to simultaneously assess their parental capacity (mandatory if the Child Protection System is involved). If both interventions are conducted by the same social worker, the alliance with the family is jeopardized, since those different relational agreements (assistance and coercive) generate confusion. To prevent this, Cirillo et al. (2018) and Sharlin and Shamai (2000) suggest the co-intervention of two professionals, one focused on the coercive interventions, while the other professional handles other concerns, such as economic, emotional, or relational issues. Sousa and Rodrigues (2012) and Pannebakker et al. (2018) enhance the role of the case manager to ensure the adequate coordination of all the agents in the intervention.
Supervision aims to help therapists and social workers to serve clients in a useful way from their specific services. Supervision, when working with FED, becomes a way to empower the professional, to help him/her take new paths to increase the efficacy of the intervention (Sharlin and Shamai 2000). During supervision, social workers may work to prevent the coalition of despair, finding ways to manage their own feelings towards the cases (Escudero 2013). In this line, Lavee and Strier (2018) suggest supervision and training in order to help social workers to manage the emotions that arise in their daily work with FED.
The present investigation is descriptive. It arises from the need to know the perspective of the social workers located in basic community social services across Mallorca (Spain), in relation to the development of a collaborative approach with families in special distress. To this end, an Inventory for the Development of Collaborative Practices in Social Work with Families in Special Distress has been compiled. Our main questions were "Do professionals value collaborative practices when working with FED?" and "Are professionals implementing those collaborative practices?". The main objective is, therefore, finding out the level of importance and the level of implementation that participants (N = 121) give to the different intervention criteria included in the inventory. Moreover, we aim to find out whether there are correlations between those intervention criteria and some organizational aspects in the different settings where the study is conducted. Our results may contribute to determining if professionals are implementing collaborative practices with FED, and to what extent. Besides, our results may contribute to determining if there is a need to implement changes at organizational and institutional levels. These results give an insight into the kind of interventions that should be promoted, in order to improve the performance of social workers in their daily practice with FED.

Design and Procedure
We made an initial panel of 113 items, based on literature of social work practice with families, including aspects such as collaborative and narrative models, common factors that explain changes in psychosocial interventions, and the interprofessional collaboration with larger systems. A two-step process revision of the initial panel was carried out, and eventually, 67 items were chosen. The first stage was completed by two social workers with more than 15 years' experience in community basic social services. The second stage of the revision was implemented by four full time professors at the University of Illes Balears (Spain), all with previous experience in social work with FED.

Instruments
Two instruments were applied: an Inventory of Collaborative Practices in Social Work with FED, and a sociodemographic questionnaire specifically designed for this study.
The 67 intervention criteria of this Inventory refer to the different skills, attitudes, and relational stances that the literature conveys as fundamental for carrying out collaborative practices in social work (basic social services), not only with families in special distress but also with larger systems involved with them. Thus, among others, there are items related to the construction of a helping relationship, the importance of a relational stance as an appreciative ally, reflection with the family about the intervention of different professionals at the same time, or the co-construction of objectives and tasks with the clients. Participants were asked to assess the level of importance of each item and the level of implementation in their current practice. A five-point frequency scale was used (5 = extremely important/always, 4 = very important/very often, 3 = important/often, 2 = not very important/sometimes, and 1 = not important/hardly ever).
Participants also filled out a socio-demographic questionnaire, including questions on gender, age, and issues related to their specific workplace (such as supervision or spaces for interprofessional coordination).
This survey was conducted between June and October 2018. A list of 146 social workers in public basic social services in Mallorca (Spain) provided the sampling frame for this study. The participants worked in different municipalities across Mallorca. Public basic social services provide a range of services for individuals, families, and groups. Act 4/2009 (Illes Balears, Spain) establishes the basic social services to be offered by the regional administration. Thus, social workers attend to different kinds of requests, and they provide information, counseling, and support, together with resource management guidance. These professionals attend FED and work with them at different levels: (a) individual, including with economic difficulties, social isolation, and professional insertion; (b) family-relational, including with the management of the household economy, parenting abilities, the improvement of relationships, relationships with larger systems involved in their lives, and coordination to ensure an integral intervention; and (c) socio-cultural, including with their community connections, their values, and culture.
The principal investigator PI of the study (T.C.) personally contacted each manager of the different services, in order to explain the aim of the present research and to ask for their collaboration in the recruitment of participants. Regarding data collection, participants completed the self-administered questionnaire in their workplaces (during working hours) by prior appointment. The PI was always in the same room, in order to respond any inquiries and to guarantee that there were no interruptions while the questionnaire was being completed. The Ethics Committee of the University of Illes Balears approved the study (70-CER-18).

Sample
The sample was made up of 121 social workers (mean age = 42 years old; SD = 8.30, range = 23-64; 92.6% females), who carried out their work in public basic social services in Mallorca (Spain). The sample universe was made up of 146 social workers; therefore, the sample constituted 82.88% of the total population under study. The sample was significant, with a 95% confidence level and a 4% margin of error.
The inclusion criteria were: (1) having a degree in Social Work; (2) developing their work in basic social services in (location), with at least 2 months in the current service; and (3) currently conducting practice with families in special distress (known as "multi-problem" families).

Data Analysis
The Z Kolmogorov-Smirnov normality test was conducted to decide whether parametric or non-parametric tests should be run. This test was conducted for each item of the questionnaire (using as variables the result of subtracting the levels of implementation from the levels of importance). The scores were 0.000 for all the items. This test was also conducted regarding the sociodemographic questionnaire, and as the scores did not exceed 0.05, non-parametric tests were conducted. Univariate descriptive analysis, frequency tables, and the Wilcoxon signed-rank test were applied to the answers of participants regarding the Inventory items (both the levels of importance and implementation). Bivariate analyses (Kruskal-Wallis and Mann-Whitney U tests) were applied to analyze the relationships between the questions and answers in the Inventory related to the workplace organization. Statistical analyses were carried out using SPSS 20.0.

Sociodemographic Data
Participants were located in different basic social services in Mallorca (Spain). A proportion of 47.1% were working in Palma, the biggest city in Mallorca, with almost half a million inhabitants. A proportion of 52.9% were working in other municipalities (the biggest one had 50,000 inhabitants). Therefore, the sample was balanced, with participants from urban and rural areas.
Half of the respondents had more than 15 years' experience working in social services, while 9.3% had less than five years' experience. In the current workplace, the mean was almost 11 years' experience (SD = 9.06). Beside this, 89.1% of the participants were working full-time (35-40 h per week), while 9.1% were working part-time (20-30 h per week).
Most participants (76.9%) had received specific training in social intervention with families in the previous five years.

Caseload, Supervision, and Interprofessional Coordination
A proportion of 45% of the respondents devoted between 41% and 60% of their time to direct client attention, while 40% of the respondents devoted between 20% and 40% of their time to working directly with clients. Bureaucracy and coordination with other institutions took up between 36% and 59% of the time of 48.3% of the respondents. Beside this, 38.3% of the participants devoted between 8% and 35% of their time to paperwork. A proportion of 62% of the respondents attended between 7 and 20 clients weekly, and 16.1% attended more than 20 clients each week. The range of new cases each week varied from one to two for the 24.8% of the participants, while 60.3% attended three or four new cases each week. As regards the closure of cases, 42.1% of the respondents closed one or two cases each month, while 42.1% closed three or four cases per month.
Supervision was not an established routine in most cases. Thus, 47.1% of the respondents stated that they received no supervision in their workplaces, while 28.1% of the participants had supervision once a month, and 28.1% of the respondents had supervision twice a month.
Interprofessional coordination implies having time and space. A proportion of 52.9% of the respondents had a weekly meeting with the professionals assigned to the same department. Nevertheless, meetings with professionals from other services were scarcer. In fact, 42.1% of the respondents stated that they had fewer meetings than were necessary.

Inventory of Collaborative Practices in Social Work with FED
A principal component analysis was conducted previously (Casado and Cardona), and six factors explained 46.88% of the total cumulative variance. The reliability was calculated through Cronbach's alpha, and the six factors or dimensions had adequate values, between 0.70 and 0.90. The results in this manuscript are organized around those six factors. Table 1 shows descriptive statistics of each factor.
The assessment of the agreement with the family on objectives and tasks was the most fulfilled factor for both levels (importance and implementation). The assessment of the stance of an appreciative ally was the factor with the least differences between the importance rate and the level of implementation (24.69); it was also the most implemented factor, with 61.28% of the respondents answering they implemented those items "very often" or "always". The assessment of the interprofessional collaboration, on the other hand, was the factor with more differences between the levels of importance and implementation (40.28). Finally, the assessment of the relationships between family and larger systems was the factor least fulfilled for both levels; thus, 63.38% of the respondents considered those items "very" or "extremely important", while 34.59% of the respondents answered that they implemented those items "very often" or "always".
One of the aims of this study was to determine which items were considered the most and the least important by participants in their current practice, and which items were the most and the least implemented. Because of this, the six tables in this section show: (a) the median and interquartile range of each item (both levels: importance and implementation); (b) Wilcoxon signed-rank tests (p value); (c) the percentages of answers "4" and "5" for both levels, importance and implementation ("very" or "extremely important" for the level of importance, and "very often" and "always" for the level of implementation); and (d) the differences between those percentages.

Assessment of Relationships between Family and Larger Systems
Factor I contains 16 items (see Table 2); thus, the possible score ranges from 16 to 80 points. For the level of importance, this factor had a minimum score of 43 and a maximum of 80 (M = 60.17; SD = 8.78). For the level of implementation, the minimum score was 27, and the maximum was 70 (M = 47.83; SD = 9.65).
A proportion of 81.7% the of respondents considered items 41, 42, and 67 the most important, while item 15 was considered the least important (28.9% of the respondents considered this item "very" or "extremely important"). Regarding the level of implementation, 68.6% of the respondents answered that they implemented item 14 "very often" or "always" (the most implemented item in this factor), while less than 20% of the respondents implemented items 38 and 35 "very often" or "always" in their current practice (the least implemented items in this factor).
The results show that, for all of the 16 items contained in this factor 1, the respondents scored the level of importance significantly more highly than the level of implementation (p < 0.001). The least differences between the scores for importance and implementation were for item 15, while the most differences were for item 35, which referred to elaborating an eco-map together with the family to jointly reflect about the relationships between the family and all the professionals and other significant people involved in their lives.

Assessment of the Interprofessional Collaboration
Factor II contains 14 items (see Table 3); thus, the possible score ranges from 14 to 70 points. For the level of importance, this factor had a minimum score of 45 and a maximum of 70 (M = 61.17; SD = 5.72). For the level of implementation, the minimum score was 22 and the maximum was 68 (M = 48.04; SD = 8.24).
This factor had high scores for the level of importance. In fact, the least valued was item 58, with 79.3% of respondents answering "very" or "extremely important", while the rest of the items in this factor scored above 80%. On the other hand, for the level of implementation, items 62 and 55 were the most implemented, while items 59 and 64 were the least implemented.  Jointly draft objectives and tasks with the individual/family in such a way that they are easily understandable, achievable, and limited enough to assess their degree of fulfillment. In case it is necessary to modify the context of a professional intervention (for example, in situations of unprotected minors), give an explanation to the individual/family of the professional's legal and ethical responsibilities, maintaining an honest position with them, so that action alternatives are provided that do not jeopardize continuity in the helping relationship. Maintain telephone or email contact between professionals to share information and the progress of the family, and to adjust the different interventions.
62 5 (4-5) 4 (4-5) <0.001 98.3 78.5 19.8 In a collaborative framework, jointly decide which professional will lead the coordination of the created macro-helping system. In the event of a relational conflict, suggest a meeting between concerned professionals and the individual/family so that they can reflect on the helping process and jointly make the necessary decisions. The results show that, for all of the 14 items contained in this factor II, the respondents scored the level of importance significantly more highly than the level of implementation (p < 0.001). The greatest differences between the scores for importance and implementation were for item 59 (61.7%), while item 62 had the lowest differences (19.8%).

Assessment of the Agreement with the Family on Objectives and Tasks
The factor III contains 10 items (see Table 4); thus, the possible score ranges from 10 to 50 points. For the level of importance, this factor had a minimum score of 30 and a maximum of 50 (M = 44.22; SD = 4.15). For the level of implementation, the minimum score was 21 and the maximum was 49 (M = 36.73; SD = 6.09).
For this factor, the results show the highest levels of importance for all of the items. In fact, at least 89% of the respondents scored "very" or "extremely important" for all of the items. Regarding the level of implementation, item 33 was the most implemented (80.2%), whereas items 31 and 51 were implemented "very often" or "always" by less than 45% of the respondents.
The results show that, for all the 10 items contained in factor III, the respondents scored the level of importance significantly more highly than the level of implementation (p < 0.001). The smallest differences between the scores for importance and implementation were for item 33 (12.4%), while the largest differences were for item 51 (49.1%).

Assessment of the Capacity-Centered Approach
Factor IV contains 10 items (see Table 5); thus, the possible score ranges from 10 to 50 points. For the level of importance, this factor had a minimum score of 29 and a maximum of 50 (M = 43.50; SD = 4.27). For the level of implementation, the minimum score was 22 and the maximum was 46 (M = 35.70; SD = 5.38).
At least 73% of the respondents considered all these items "very" or "extremely important". Regarding the level of implementation, items 65 and 66 were the least implemented, with less than 35% of the respondents answering "very often" and "always". The most implemented items were 3 and 18 (rates over 75%). Beside this, item 18 presented the lowest differences between the rates of importance and implementation (17.4%), while items 48 and 49 had the highest differences between both levels.
The results show that, for all of the 10 items contained in this factor IV, respondents scored the level of importance significantly more highly than the level of implementation (p < 0.001).

Assessment of the Stance of an Appreciative Ally
Factor V contains nine items (see Table 6); thus, the possible score ranges from 9 to 50 points. For the level of importance, this factor had a minimum score of 26 and a maximum of 45 (M = 38.93; SD = 3.85). For the level of implementation, the minimum score was 21 and the maximum was 45 (M = 33.65; SD = 4.94).
This dimension had high scores for all the items for the level of importance. Thus, the least rated was item 8 (79.2% of respondents considered it "very" or "extremely important"). On the other hand, the most implemented was item 2 (81%), while item 10 was the least implemented (40.5% of respondents implemented item 10 "very often" or "always").
The results show that, for all the nine items contained in this factor V, the respondents scored the level of importance significantly more highly than the level of implementation (p < 0.001). The item with the greatest differences between these levels was item 9 (39.2%), while item 1 showed the lowest differences (10.9%).

Assessment of Family Expectations
Factor VI contains eight items (see Table 7); thus, the possible score ranges from 8 to 40 points. For the level of importance, this factor had a minimum score of 12 and a maximum of 40 (M = 34.32; SD = 3.85). For the level of implementation, the minimum score was 11 and the maximum was 39 (M = 28.94; SD = 5.18). When explaining how we will work, agree with the individual/family on cross-cutting issues related to the helping process: accompaniment of the professional in the decision-making and leadership of the process by the family, the confidentiality of the process, and the connection with other professionals involved, and teamwork. Based on the feedback received from the individual/family session by session, both of you will partake in the reflections and changes that are considered appropriate.  Consider that the individual/family has the capacities and resources necessary for change, and express this idea explicitly.
18 5 (4-5) 4 (4-5) <0.001 96.7 79.3 17.4 Listen to the life story that the individual/family tells, without losing sight of the fact that this narrative is inscribed in a dominant cultural system and values, typical of a broader socio-cultural context.

4 (4-5) 4 (3-4)
<0.001 91.7 67.8 23.9 When the understanding of the situation is considered sufficient, arrange a meeting to build a co-diagnosis with the individual/family. When faced with obstacles that may occur in the helping process, understand that the difficulty lies in the interaction, rather than in the family or in the professional.  Transmit the need to understand what they tell us, from a stance of "not knowing", out of genuine curiosity.  Table 7. Factor VI: assessment of family expectations.

Item
Import. 1 Median (IQR) Implem. 1 Median (IQR) Wilcoxon p Value Import. 1 (% 4 + 5) * Implem. 1 (% 4 + 5) * Difference (%) ** Transmit to the individual/family that they are the protagonist of the process, and encourage them to take active roles in decision-making regarding the intervention process. Make an initial work agreement with the family: clarify the initial objectives (that justify the beginning of the helping relationship), that is, what things we agree to work on together.
26 5 (4-5) 4 (3.5-5) <0.001 95.9 78.5 18.2 The item least important and least implemented for the respondents was item 25 (71.7% of respondents rated it as "very" or "extremely important", and 40.8% of the respondents implemented this item "very often" or "always"). The other items in this factor scored above 80% for the level of importance. For the level of implementation, items 7 and 26 were the most implemented and showed the smallest differences between importance and implementation. Beside this, items 19 and 21 had the greatest differences between importance and implementation.
The results show that, for all of the items contained in this factor VI, the respondents scored the level of importance significantly more highly than the level of implementation (p < 0.001).

Relationships between Inventory and Sociodemographic Questionnaire
The Kruskal-Wallis and Mann-Whitney U tests were applied to determine whether the levels of importance or implementation of the Inventory factors had significant differences according to sociodemographic data and/or organizational aspects. No differences were found related to gender, previous experience, caseload, developing supervision, or having protocols about interprofessional collaboration. Below, we detail those aspects that showed differences between groups.

Age of Participants
Those participants aged between 23 and 35 years scored all the factors more highly than the older participants for both the importance and implementation levels. Nevertheless, only three factors showed statistically significant differences for the level of importance (Kruskal-Wallis p value): factors II (p = 0.006), III (p = 0.012), and V (p = 0.009).

Training of Participants
Although the participants who had received more hours of training in social intervention with families scored almost all the factors more highly (both in importance and implementation), none of the factors showed significant differences (Kruskal-Wallis test).

Location of Workplace
Those participants who were located in small municipalities (with fewer than 10,000 inhabitants) or in municipalities with fewer than 15,000 inhabitants gave higher scores than respondents working in bigger municipalities. The Kruskal-Wallis p value was significant only for the level of importance, in factors I (p = 0.027) and V (p = 0.014).

Discussion
The results show that participants considered most of the intervention criteria as "very" or "extremely important". Thus, 54 criteria obtained high scores in more than 75% of the cases. The two criteria that obtained the highest rating for the level of importance are related to the management of resources (mail and telephone contacts with professionals, and the use of local resources). Other criteria that were among the most highly valued were those related to active listening, as well as showing respect to the family and developing empathy.
The levels of implementation manifested by the social workers are significantly lower than the levels of importance. Thus, professionals may increase their levels of implementation for all the items, considering that all the items proved to be more valued than implemented. This would contribute to the development of collaborative practices with FED in a more efficient way.
The two most implemented criteria were related to active listening and the identification of the services involved in the case. The least valued and implemented criteria were those that are either novel because they are presented in narrative language or refer to reflection on the relationships between the family and larger systems. Thus, having a conversation with the family about their future relationships with larger systems was the least implemented (factor I, item 38), and elaborating an ecomap with the family was the second least implemented (factor I, item 35). Interprofessional coordination becomes essential when working with FED. In fact, authors such as Imber-Black (2000), Lipchik (2004), and Sluzki (1996) claim the importance of reflecting with the family about their relationships with formal and informal support systems, and the eco-map is a useful visual tool when working with FED, since they often maintain a significant number of relationships with other professionals, implying a variety of objectives and tasks. Pannebakker et al. (2018) explain that having a coordinator that guarantees a holistic intervention may be a way to better attend to these families. Therefore, we suggest that regional administrations should introduce the eco-map as a mandatory tool, in order to guarantee that social workers and families become aware of all those relationships. A regular use of the eco-map would avoid duplicities and/or contradictory messages, and would generate possibilities for collaborative work between the family and the multiple agents involved (Imber-Black 2000).
On the other hand, the items related with the management of resources and contacts with professionals (factor II, items 62 and 55) were the most implemented. Therefore, we can state that criteria that involve reflection processes-with professionals or with family members-seem to remain in the background, privileging, instead, aspects related to resource management. These results are aligned with other studies that claim the need for social workers to have spaces for reflection about their perfomance with FED (Imber-Black 2000 ;Lipchik 2004;White and Epston 1993).
Collaborative practices privilege the voices of the clients. Accordingly, feedback becomes essential for establishing a relationship based on respect and equity (Duncan 2002;Escudero 2013;Madsen 2007). Our results showed that asking for feedback at every meeting with the family was the third least implemented item (factor II, item 59). Thus, our main findings show that social workers do not implement feedback as a systematic routine with their clients or ask for their expectations about the service and the professional. Client expectations play an important role in the helping relationship and are one of the common factors that explain change in psychosocial interventions (Duncan 2002;Lambert 1992;Rosenzweig 1936). If the client trusts the social worker and considers that he/she is competent, then this client will be more engaged in the objectives and tasks (Bedi et al. 2005;Lipchik 2004). In the same line, Ayala-Nunes et al. (2014) state that when the professional asks the family about their expectations in relation to the service, the power asymmetry between them diminishes.
In post-modernism, professionals ensure that the family is engaged in the process (White and Epston 1993). In fact, there are numerous studies that have proven the multiple benefits of asking the clients for their opinion on the ongoing process. Anker et al. (2011) stated that clients value the opportunity for giving feedback to the professional. Several feedback systems have been put forward, such as the Outcome Questionnaire (OQ45.2) (Lambert 1992) and the Working Alliance Inventory (Horvath and Luborsky 1993). More recently, the Partners for Change Outcome Management System (PCOMS) has integrated both elements: outcome and alliance (Duncan 2014;Duncan and Reese 2015;Sparks and Duncan 2018). With the PCOMS, at each session, the client gives feedback about outcomes (with the Outcome Rating Scale) and about the therapeutic alliance (with the Session Rating Scale). Despite the fact that in Spain, these instruments are not yet widely used in social service settings, systematic feedback may increase the effectiveness of the interventions developed with the families.
We should bear in mind that this sample reported having a high caseload, and this hinders their capacity to find the time and space for reflecting on the work developed. Other studies have also reported these high caseloads (e.g., Leigh and Miller 2004;Marthinsen et al. 2020). Moreover, as we highlight in the results (Section 3.2), supervision is not included as a mandatory activity for most social workers in Mallorca (Spain), and this may contribute to not finding reflection spaces. Different authors consider that social workers are in between two paradigms: modernism and post-modernism (Madsen 2007;Sousa and Eusébio 2005;Sousa et al. 2007). The first paradigm implies a perspective based on deficits (the main question is "what do you need"?), and, accordingly, the professional suggests changes to help the individual or family to achieve standard levels of social functioning. The second paradigm implies that the professional and client have a more balanced relationship; social workers honor the expertise of the family members (the main question is "where do you want to go?"). The professional adopts a not-knowing stance (Anderson 1999) and asks the family for permission to co-investigate the situation, thinking together about other new paths and possibilities (Madsen 2007). Our results may indicate that social workers are still in between those paradigms.

Conclusions
Social workers tend to privilege resource management over reflection on the ongoing process with the family and other professionals involved. That is, most social workers regard collaborative practices as highly important, though it is often clear that they do not always implement these practices in the same proportion. Several causes may be interrelated, but further research should be conducted in order to explore possible factors that could be influencing these results. First of all, the client ratio is quite high in Mallorca; thus, social workers usually meet five new families each week, while they close three to five cases per month. This implies a high caseload that may hinder their capacity for having space to reflect on their intervention and the alliance established. Nevertheless, the results showed no differences for those participants with higher caseloads, and this may imply other reasons for not developing collaborative practices as the literature recommends. Secondly, supervision is not routine for most social workers in Mallorca; therefore, they do not have the opportunity to reflect in a formal space about an ongoing case in order to decide on a course of action. Our results showed no difference for those participants who had supervision as a routine activity; nevertheless, further research should be carried out in order to deepen our knowledge of the way in which that supervision is undertaken. Beside this, asking the client for feedback is a great way to become an appreciative ally with the individual or family. Our results showed that participants do not ask for feedback or for the expectations of the family in relation to the intervention process. Therefore, the need arises for regional administrations to provide social workers with the framework and tools required so that they can develop a strong working alliance, not only with the family but also with other professionals involved (specific training in collaborative practices, more time to conduct interviews, spaces for interprofessional coordination, and monthly supervisions, among others).
This study has limitations. First, the fact that the Inventory is self-administered, first asking about the level of importance and then about the implementation of each intervention criterion, may have influenced the results. Secondly, coherence is observed in the responses provided for the criteria of the different areas in the inventory, although the criteria were not ordered by dimensions. Another limitation is that this study focuses on the perspective of the professionals, not the families, so the results and conclusions only address part of the helping macrosystem created ad hoc. Finally, we consider that future research should include qualitative methodologies. In this way, by having access to the narrative of the professionals themselves, we will be able to deeply discuss issues related to the opinions and considerations of the professionals regarding the interventions with these families.
This study highlights that FED often have specificities that social workers must bear in mind (distrust, hopelessness, low auto-efficacy, and poor communication abilities, among others). Thus, enhancing the implementation of techniques and attitudes that help these families to gain confidence and trust in the process of change remains a key challenge for many social workers. Privileging resource management over reflection spaces is, therefore, ineffective and counterproductive. Our Inventory may be an opportunity to reflect on how social workers are intervening with FED. It might also promote the performance of attitudes and techniques that enable professionals to develop a working alliance with their clients. Accordingly, the present Inventory might help professionals and clients to jointly reflect about their expectations, desires, and capacities. Our results have implications not only for the way in which social workers develop their interventions with FED, but also for re-thinking organizational aspects in community basic social services. Moreover, a lower caseload will provide social workers with more time and space to reflect on their intervention. In this line, allowing social workers the full exercise of interprofessional coordination when necessary is another issue that regional administrations should consider and promote. Finally, supervision also becomes an essential activity that should be mandatory in social services, especially when working with FED, as it is an essential tool for ensuring that professionals feel supported in their performance.