Collaboration between community social services and healthcare institutions: The use of a collaborative individual plan.

BACKGROUND: Well-functioning care of people with substance use and psychiatric disorders presumes collaboration between different parties such as psychiatric care and substance use treatment centres, as well as social services. According to Swedish law, a collaborative individual plan, i.e., a written action plan to support structured inter-organisational collaboration, should be established. However, there are indications that such action plans are not used to a satisfactory extent. AIM: To explore current inter-organisational collaboration and use of collaborative individual plans among healthcare units and social services in Stockholm County.
DESIGN: The study uses a cross-sectional design. Participants (N = 797) in a course specifically aimed at improving the knowledge and use of collaborative individual plans were invited to take part in the study prior to attending the course. A total of 705 participants accepted. Data were collected through an electronic questionnaire sent to each participant's workplace. Non-respondents were offered a paper version to fill out.
RESULTS: Respondents reported participating in one to two collaborative individual plans per month and about 70% reported using a particular template. Respondents perceived mainly positive consequences of establishing a collaborative individual plan, for instance that it clarifies what measures are to be performed and who is responsible.
CONCLUSIONS: Although respondents were generally positive about establishing a collaborative individual plan and the consequences thereof, they reported low use of such action plans.

The de-institutionalisation of Swedish psychiatric care during the second half of the 20th century presupposed well-functioning horizontal integration by collaboration between psychiatric care, substance use treatment centres and municipal social services, to support individuals with substance use and psychiatric disorders. A call for both inter-organisational and inter-professional collaboration between parties was made. A simultaneous development was the implementation of a purchaser–provider system in most Swedish county councils. Purchasers govern the healthcare system by paying for certain performances from providers. In Stockholm, the purchaser organisation has responsibility for primary healthcare, psychiatry and geriatrics (Bergman, 1998). Together with the implementation of the healthcare choice reform (i.e., individual choice of healthcare provider) this has led to quasi-market flaws because of the competing family doctor system and integrated local healthcare in Stockholm (Ahlgren & Axelsson, 2011). Furthermore, since the year 2010, collaboration between healthcare and social services has been established in Swedish law (3 f § HSL, 2 kap. 7 § SoL). Professionals, patients and their families should all be part of the collaboration, to make patients more influential in their own care, as this may have positive consequences (Piippo & Aaltonen, 2004). Such integration and influence is not achieved within the top-down governed purchaser–provider model. The belief that collaboration between psychiatric healthcare and addiction healthcare supports not only positive outcomes among patients (Morisano, Babor, & Robaina, 2014), but also prevention, is obvious even in the federal health reform in the US (Affordable Care Act; see Abrams et al., 2015).

Different strategies have been explored to facilitate collaboration (Cameron & Lart, 2003; Cameron, Lart, Bostock, & Coomber, 2014). One strategy is to have pooled budgets, i.e., common budgets for different agencies. However, experiences from England and Sweden show that a pooled budget is not enough to make front-line professionals collaborate more efficiently (Hultberg, Glendinning, Allebeck, & Lonnroth, 2005). However, a model with both a common budget and common organisation in a public company, the Norrtaelje model (Andersson & Calltorp, 2015), has been successful in a Swedish context. The model was shown to have lower costs for delivery of services to the elderly over a six-year period, compared with other Swedish municipalities of equal size. Similarly, a Swedish study of Coordination Associations, characterised by both co-financing and the formation of multi-professional teams, concluded that the use of local strategies is also needed for efficient collaboration and to maintain collaboration over time (Ståhl, Svensson, Petersson, & Ekberg, 2011). Forming multi-professional teams is a commonly used model for joint work (Cameron et al., 2014). For instance, Strategic Collaboration Councils are composed of personal ombudsmen, case managers, and healthcare personnel. However, Liljegren (2013) found that the main activity in those councils was at the lowest level of integration, i.e., information exchange, and only to a small degree focused on problem identification and decisions about allocation of responsibility for measures. Moreover, Mossberg (2014) found indications of an overall larger interest in the collaboration model per se rather than in the client’s problems. However, it was shown that healthcare conducted by an inter-disciplinary team consisting of a physician, a nurse and a social worker was more successful than care conducted solely by a physician, at least as regards hospitalisations, readmissions and social activities (Sommers, Marton, Barbaccia, & Randolph, 2000).

A number of different barriers have been identified with regard to joint work and many are related to the fact that personnel from different disciplines need to collaborate (Cameron & Lart, 2003; Cameron et al., 2014; Maslin-Prothero & Bennion, 2010). For instance, there may be issues with regard to establishing common aims and objectives (Drennan et al., 2005), difficulties in communication and status inequalities (Cameron, Macdonald, Turner, & Lloyd, 2007; Holtom, 2001), and problems with understanding the roles and responsibilities of individual professionals and parties (Glasby, Martin, & Regen, 2008). Moreover, in a study by Magnusson and Lützén (2009), it was concluded that different ideologies, experiences and goals for care and service influence the effectiveness of collaboration at an inter-professional level. Similar barriers were identified by Widmark, Sandahl, Piuva, and Bergman (2011), who conducted a study of collaboration between healthcare, social services and schools in Stockholm. It was concluded that these parties had difficulties collaborating due to a perceived lack of clarity in the allocation of responsibilities from the organisation, different approaches to client needs, and a mistrust of the other professionals’ competence.

One strategy to overcome these barriers and increase the patient’s influence over his or her own care is to implement a collaborative individual plan (CIP), which can be considered a written action plan embodying a policy on how organisations should collaborate in a structured way to benefit a patient. In such collaboration, more services believed to lead to a better prognosis can be offered to the client. In fact, in accordance with Swedish law, inter-organisational collaboration between healthcare institutions and social services should involve a CIP. Thus, within Stockholm County Council and municipal social services, there are a certain set of mandatory items to be included in a CIP, such as what the needs of the client are, what measures need to be performed and which party is responsible, aims and goals of the measures, who has the overall responsibility for the action plan, and when to follow up the action plan. To facilitate this work, there is a particular template available for establishing a CIP within the Stockholm County Council patient records system “Take Care”, to be used among healthcare and substance use treatment professionals. Brief training courses on CIPs have been developed and offered to healthcare professionals and personnel within social services. However, there are indications that these action plans are not used to a satisfactory extent. Although training courses and information dissemination are often used when implementing a method, Fixsen, Naoom, Blasé, Friedman, and Wallace (2005) found that such measures are ineffective by themselves. Instead, Fixsen et al. suggest a multilevel model for implementation consisting of core components (e.g., selection of personnel, pre-service and in-service training, coaching), organisational components (facilitative administrative structures, training and support), and influential factors (social, economic, political) that should be applied in order for implementation to be effective. Stockholm County Council and the Stockholm County Association of Local Authorities (KSL) therefore jointly developed a three-day CIP course intentionally allotting eligible professionals to be trained together. The aim of the present study was to investigate the current inter-organisational collaboration and use of CIPs among professionals before taking part in the course. Specific research questions were: How often are CIPs established? How often and why have respondents abstained from establishing a CIP although there was reason to establish one? How many know about the existing CIP template and the contents thereof? What are the perceived consequences of establishing a CIP?

Methods

Design and procedure

During the period November 2014 to March 2015, personnel working in municipal social services, psychiatric care and substance use treatment services within Stockholm County participated in a three-day course (one full day and two half days) focusing on establishing a CIP. The overall aim of the course was to improve the inter-organisational collaboration and care of people with psychiatric problems and dependency disorders by using an action plan. The course was run on ten different occasions at different locations to reach as many participants as possible. The head of each participating unit decided which personnel would have the chance to attend the course. About two weeks before each course date, an invitation to participate in the study was sent out via email to each course participant. The invitation contained a link to a web-based questionnaire. In order to increase the response rate, up to three reminders were sent and non-respondents were also offered the chance to respond to a paper version of the questionnaire in connection to the start of each course. Thus, data were collected in ten waves. This article reports the result of the baseline assessment, i.e., before participation in the course. A follow-up to assess possible changes in the collaborative work is planned one year after each course date, i.e., during November 2015 to March 2016. The results from the follow-up assessment will be reported in a future article.

Measures

A questionnaire was developed and pilot-tested by three employees at dependency units within Stockholm County Council who were not taking part in the course. A self-constructed questionnaire was developed, which contained a total of 33 questions and took approximately ten minutes to complete. The questionnaire included six background questions, followed by a question about whether or not the respondent met clients who might be in need of a CIP. This was followed by 19 questions about the respondent’s current and past work with CIPs (in a 1-, 6- and 12-month perspective), such as, for instance, the presence of a template and the number of action plans established, reasons for not establishing plans, and the content of the current template. Those who responded “no” to the question “Have you ever participated in the establishment of a CIP?” were asked to skip items about the plan and template. Finally, the respondents answered six questions about their perceived working climate in relation to current inter-organisational collaborations.

Sample

The course was aimed at front-line personnel working in psychiatric care, substance use treatment services, and municipal social services within Stockholm County. The types of psychiatric care and substance use treatment services primarily included outpatient services comprising both primary and special healthcare centres (e.g., units focusing on youth and adult psychiatry, substance use treatment, rehabilitation, eating disorders), while the social services, in addition to general social services, included, for instance, units focusing on substance use problems, social psychiatry, housing support, and financial assistance. Operation managers and unit heads were informed about the course and asked to register their personnel. In total, the course had the capacity for 1000 individuals. However, for administrative reasons, 797 employees enrolled in the course and were invited to respond to the questionnaire. A total of 642 individuals responded to the web-based questionnaire and 63 individuals filled out the paper questionnaire, resulting in 705 respondents (88.4% response rate).

Statistics

Descriptive statistics, Pearson’s χ2, Student’s t-test, and comparisons between groups using one-way ANOVA were analysed using SPSS 22.0. Results yielding a p < .05 were considered statistically significant.

Ethical statement

The Regional Ethical Review Board in Stockholm (2015/483–31/5) approved this study.

Results

Self-reported demographic data are presented in Table 1. The majority of the participants were women, mirroring the gender distribution among employees at this level within social services and the county council. Twenty per cent of the participants were men and they were significantly older than the women [t(697) = 2.27, p = .02]. The level of education was high, as more than 90% of both men and women reported university or college as their highest education level. More than half of the respondents reported that they currently worked in municipal social services, about one fourth reported that they worked in psychiatric care, and one fifth worked in substance use treatment services. Only a small proportion of the course participants (about 4%) worked in other areas of the healthcare system (Table 1). The distribution of gender was approximately similar across current work areas [χ2(3, N = 703) = 1.40, p = .706]. Both men and women reported an experience of their current work of somewhat more than six years.

Table 1.

Demographic information about the respondents. Means (M), standard deviations (SD) and proportions in per cent.

	Women (N = 563)	Men (N = 140)	Total (N = 703)
Age, M (SD)	44.3 (11.7)	46.7 (10.9)	44.7 (11.5)
Highest education, % (N)
Primary education	–	0.7 (1)	0.1 (1)
Secondary education	5.3 (30)	4.3 (6)	5.1 (36)
Other post-primary education	0.4 (2)	0.7 (1)	0.5 (3)
College or university	92.2 (518)	90.0 (126)	91.6 (644)
Practical job training	2.1 (12)	4.3 (6)	2.6 (18)
Current work area, % (N)
Social services	81.5 (295)	18.5 (67)	51.6 (362)
Psychiatric care	77.7 (136)	22.3 (39)	25.0 (175)
Substance use treatment services	78.5 (106)	21.5 (29)	19.3 (135)
Other healthcare areas	82.8 (24)	17.2 (5)	4.1 (29)
Years at current workplace, M (SD)	6.1 (6.8)	6.3 (6.9)	6.1 (6.8)

Almost all (93.6%, N = 658) of the respondents reported that they, in their current work, met clients for whom a CIP should be established. There was no statistically significant difference between professionals within psychiatric care, substance use treatment services, or social services. Furthermore, 494 respondents (70.1%) reported having been involved in work involving a CIP on at least one occasion. Whether or not the participants established such action plans was assessed and Table 2 shows how many times the participants had collaborated in establishing an action plan. They reported having participated in establishment of a CIP between one and two times during the previous month and in establishment of between five and eight action plans during the previous six months. There were significant differences between work areas with regard to the number of action plans respondents had participated in during the previous 30 days [F(2, 406) = 6.98, p < .001] and six months [F(2, 403) = 6.66, p = .001]. Pairwise comparisons post hoc according to Scheffé’s method showed that community social services reported participation in significantly fewer action plans than either psychiatric care or substance use treatment services, for both time intervals (p < .03). The effect size when comparing social services and psychiatric care over the previous 30 days was d = .362, and when comparing social services and substance use treatment services, it was d = .335. The corresponding effect sizes for six months were d = .400 and d = .363, respectively, which are small effects according to Cohen (1988). There were also significant differences between work areas in the number of self-initiated CIPs during both the previous 30 days [F(2, 303) = 5.61, p = .004] and the previous six months [F(2, 304) = 6.48, p < .002]. Pairwise comparisons post hoc showed that psychiatric care initiated significantly more action plans than municipal social services at both the 30-day and the six-month intervals (p < .004). The largest effect size (d = .411) was between psychiatric care and social services, which is, however, still a small effect according to Cohen (1988). Table 2 also reveals that participants self-reported that they personally initiated most of the CIPs they participated in and that personnel within psychiatric care as a group reported somewhat more self-initiated action plans than the total number of action plans they participated in. Of those respondents who reported having participated in work involving an action plan (N = 494), 14.0% reported that they, during the previous six months, had abstained from establishing an action plan although there was a reason to establish one. The most frequently reported reason for abstaining from establishing a plan was perceived difficulties in collaborating with other parties (47.8%), followed by lack of time (34.8%), and lack of client consent (23.2%).

Table 2.

The mean (M) and median number of collaborative individual plans (CIPs) established within the social services, psychiatric care and substance use treatment (SUT) services; standard deviations (SD) within brackets.

	M (SD)	Median
Number of CIPs participated in during the last 30 days
Social services (N = 209)	1.2 (1.7)	1.0
Psychiatric care (N = 117)	1.9 (2.3)	1.0
SUT services (N = 83)	1.8 (2.0)	1.0
Number of self-initiated CIPs during the last 30 days
Social services (N = 147)	1.0 (1.6)	1.0
Psychiatric care (N = 87)	2.1 (3.6)	1.0
SUT services (N = 72)	1.3 (1.8)	1.0
Number of CIPs participated in during the last six months
Social services (N = 207)	5.0 (6.3)	3.0
Psychiatric care (N = 117)	8.1 (9.8)	4.0
SUT services (N = 82)	7.5 (8.2)	5.0
Number of self-initiated CIPs during the last six months
Social services (N = 148)	3.5 (5.4)	2.0
Psychiatric care (N = 88)	8.2 (17.8)	3.0
SUT services (N = 71)	3.8 (4.2)	2.0

A total of 52.6% (N = 346) of the respondents who met clients in need of a CIP reported that they knew that there existed a template within their organisation, while 11.4% (N = 75) reported that there was no existing template available and 10.5% (N = 69) that they were not aware of a template. A significant difference between work areas was revealed [F(2, 470) = 10.35, p < .001]. Pairwise comparisons post hoc showed that in comparison to respondents working in the social services, a statistically significant higher proportion of respondents working in psychiatric care and substance use treatment services reported being aware of a template (62.2% in social services versus 77.7% in psychiatric care and 85.7% in substance use treatment services, p < .03). As can be seen in Table 3, about 90% or more of the respondents agreed that a certain set of items was included in their template, for instance, what measures have to be performed and who is responsible for each measure, what kind of support the client needs, when to follow up the action plan, and who has the overall coordination responsibility for the client. Almost 80% reported that their template contained an item about what measures the client wished for, while almost 70% and about 60% reported that their template contained items about long-term and short-term goals, respectively. Not even half of the respondents reported that the template contained items about the client’s housing situation or occupation, or how the follow-up should be conducted.

Table 3.

The content (%) of currently used collaborative individual plan (CIP) templates as reported by the respondents.

	Yes % (N)	No % (N)	Don’t know % (N)
Client personal data (N = 342)	95.3 (326)	0	4.7 (16)
Date the CIP was established (N = 341)	95.0 (324)	0.6 (2)	4.4 (15)
Who participated at the meeting about CIP (N = 341)	94.1 (321)	0.9 (3)	5.0 (17)
What measures have to be performed (N = 340)	93.8 (319)	1.5 (5)	4.7 (16)
Which party has the responsibility for each measure (N = 340)	92.9 (316)	1.2 (4)	5.9 (20)
What kind of support does the client need (N = 341)	92.1 (314)	1.5 (5)	6.5 (22)
Time-point for the CIP follow-up (N = 341)	89.1 (304)	3.5 (12)	7.3 (25)
Who has the overall coordination responsibility (N = 340)	86.2 (293)	4.1 (14)	9.7 (33)
What measures the client wishes for (N = 339)	77.0 (261)	12.1 (41)	10.9 (37)
Long-term goals (N = 340)	68.8 (324)	16.8 (57)	14.4 (49)
Short-term goals (N = 340)	63.2 (215)	20.9 (71)	15.9 (54)
The client’s housing situation (N = 340)	45.0 (153)	39.1 (133)	15.9 (54)
How the follow-up should be conducted (N = 340)	42.9 (146)	33.8 (115)	23.2 (79)
The client’s occupation (N = 339)	39.8 (135)	41.9 (142)	18.3 (62)

Table 4 shows the participants’ perceived positive consequences of collaborative work resulting from the establishment of a CIP. Clarified responsibilities, improved client influence on his or her own care and the designation of patients to adequate care are examples of consequences that personnel reported as results from establishing an action plan.

Table 4.

Respondents’ perception of positive consequences (%) following establishment of a collaborative individual plan.

	Yes % (N)	No % (N)
The responsibilities are clarified (N = 467)	92.7 (433)	7.3 (34)
Increased client influence and participation in planning of own healthcare (N = 469)	86.1 (404)	13.9 (65)
The client gets adequate support (N = 470)	83.2 (391)	16.8 (79)
Improved coordination of the client’s care (N = 468)	83.1 (389)	16.9 (79)
Increased initiation of interventions and support (N = 470)	82.3 (387)	17.7 (83)
Easier for the client to receive adequate care (N = 468)	81.6 (382)	18.4 (86)
Increased initiation of investigations (N = 468)	70.9 (332)	29.1 (136)

A question was asked about what other organisations the respondents had collaborated with using a CIP during the previous 12 months (Table 5). Results revealed that respondents working in substance use treatment services, psychiatric care and social services, respectively, most often reported having engaged in collaboration with social services, psychiatric care, and substance use treatment services. In addition, other commonly reported collaborations were with the Swedish Social Insurance Agency (e.g., 29.8% of those working in psychiatric care had collaborated with this organisation), employment services (e.g., 26.0% among psychiatric care personnel), and correctional treatment (e.g., 34.0% among personnel in substance use treatment services). Primary healthcare was the organisation least commonly reported among collaboration partners over the previous 12 months.

Table 5.

Respondents working in social services, psychiatric care, and substance use treatment (SUT) services reporting collaboration with other organisations (%)*.

	Social services % (N)	Psychiatric care % (N)	SUT services % (N)	Social Insurance Agency % (N)	Employment services % (N)	Correctional treatment % (N)	Primary healthcare % (N)
Social services (N = 363)	–	71.7 (170)	50.2 (119)	13.1 (31)	11.8 (28)	13.1 (31)	8.0 (19)
Psychiatric care (N = 175)	75.6 (99)	–	38.9 (51)	29.8 (39)	26.0 (34)	5.3 (7)	16.0 (21)
SUT services (N = 135)	78.7 (74)	58.5 (55)	–	18.1 (17)	18.1 (17)	34.0 (32)	6.4 (6)

*Respondents working in other areas (N = 29) are not presented.

A question was also added about which organisations had been the easiest to collaborate with during the previous 12 months. Here, respondents working in psychiatric care and substance use treatment services reported that they believed it was easiest to work with social services (70.8% and 83.0%, respectively), while respondents from social services reported that it was easiest to work with psychiatric care (49.6%). On the other hand, respondents working at social services also reported that they thought it was most difficult to collaborate with psychiatric care (47.5%). Furthermore, respondents working in substance use treatment services reported that it was most difficult to collaborate with psychiatric care (46.5%), while about one in four of those working in psychiatric care reported that it was most difficult to collaborate with social services (26.7%) and substance use treatment services (23.3%). A multivariable regression analysis showed that a model with gender, age, workplace, and work experience explained only 3% of the variance with regard to the number of CIPs participated in during the previous six months. The only significant variable was “working in social services” (p = .001).

Discussion

It has been shown that both integration and user influence in healthcare lead to more successful care (Sommers et al., 2000) and lower costs (Andersson & Calltorp, 2015) than a top-down governed purchaser–provider model. In an attempt to investigate the current inter-organisational collaboration and use of CIPs among professionals within Stockholm County, we have conducted a survey among professionals before taking part in a CIP course. This study evaluates collaboration among personnel at social services, psychiatric care, and substance use treatment services, and the use of CIPs among these professionals. Our results reveal that the level of established CIPs is low and that respondents report that they to some degree abstain from establishing CIPs, for instance due to difficulties in collaborating with others. On the other hand, the majority of respondents are aware of the concept of CIPs and that there is an existing CIP template, and most respondents identify several positive consequences of implementing CIPs.

Because the concept of CIPs is well established and a template can be found in the patient records system of Stockholm County Council, the majority of the respondents were aware of the CIP template. For instance, about 70% of the respondents reported that they had been involved in work with such an action plan at least once and that there was a template available within their current organisation. This means that the top-down governed purchaser–provider organisation exists at the same time as horizontal collaboration is supported. About nine out of ten respondents reported that they, in their current work, met clients for whom a CIP would be useful. This suggests that collaboration to some extent involves establishing an action plan. However, results presented here indicate that these action plans are currently used at a relatively low level, as the number of CIPs that the respondents reported that they had participated in during the previous 30 days was between one and two, which probably only covers a small part of the potential need. The same figure for a six-month perspective was between five and eight action plans, with respondents working in social services reporting the lowest participation rate (5.0 action plans during the previous six months) relative to psychiatric care (8.1 plans) and substance use treatment services (7.5 plans). The observed low degree of implementation is probably due to the implementation methods per se, mainly consisting of training and information dissemination, which, according to Fixsen et al. (2005), are not sufficient to improve collaborative work. To let the eligible collaborators be trained together and have discussions about authentic vignettes was apparently not enough. The low level of CIPs established in Stockholm may be due to dual policies, making personnel confused. Since a common template exists within psychiatric care and substance use treatment services, as a part of the patient data system (Take Care), but not within municipal social services, one conclusion would be that in order to improve collaboration using a CIP, the template should also be made available to social service personnel. Further, the respondents claimed that they themselves initiated the main part of the action plans they participated in. This indicates that collaboration currently is not at an optimal level, since they should be invited by other initiators too. Furthermore, due to reasons such as difficulties in collaborating with other parties and lack of time, almost 15% of the respondents reported that during the previous six months they had abstained from establishing a CIP, even though there was reason to establish one.

To a large extent, respondents agreed about the content of their respective CIP templates; for instance what kind of support the client needs, what measures have to be performed and who is responsible for each measure. Nonetheless, it appears that the templates currently used by healthcare and social services are lacking some information such as current work and housing status and how follow-up should be conducted. Thus, the results presented here suggest that the current templates could be improved. However, the fact that respondents, to a relatively high degree, agree about much of the content of their templates indicates a higher degree of collaboration when utilising structured action plan templates than was found by Liljegren when investigating the work within Strategic Collaboration Councils (Liljegren, 2013). Participants in this study also perceived positive consequences of collaboration using a CIP. Similar to what was found by Ståhl and co-workers (2011) when investigating Coordination Associations, the present study found that personnel perceived collaboration as having positive consequences. These positive consequences should work as reinforcement for more collaboration; however, it appears that they do not. Some respondents indicated that this was due to difficulties in collaborating and a perceived lack of time. We have, however, no indication of the reason collaboration was at such a low level despite the perceived positive consequences. One probable explanation that has been observed in other studies (Hudson, 2002; Widmark et al., 2011) is that the low level of collaboration is due to obstacles at the individual level. Individual level factors may also explain what appear to be contradictory results; for instance half of the respondents in social services reported that they think it is both easy and difficult to collaborate with psychiatric care. Thus, it is probable that respondents do not perceive an organisation as one entity, but rather as many individuals who can be either easy or difficult to collaborate with. Nonetheless, it may be recommended that collaboration training should be an element, and perhaps even synchronised, in the course syllabus for university-level healthcare and social work education. Moreover, when designing a CIP training course, it is advisable to gather professionals from the relevant organisations supposed to engage in collaboration. Further, the training should be related to the participants’ actual practice, which could be facilitated by training exercises and coaching, as suggested by Fixsen et al. (2005). The successful implementation of CIPs also depends on organisational factors, which could be facilitated via monetary incentives for establishing CIPs and also by the management who should encourage participation in CIP training and use within their organisations.

There are some limitations to this study. The main concern applies to the representativeness of the respondents, as course participants (and hence study participants) were selected by their managers. This calls the generalisability of the results into question. However, it should be noted that the response rate was high, with almost 90% of the course participants responding to the questionnaire. This indicates that the respondents represent the course participants to a large degree. Another possible drawback is the fact that a self-constructed questionnaire was used, asking about specific items covered in the subsequent training course. It is possible that collaboration on other items is more frequent, which leads to the question of whether or not a CIP is a good measure of collaboration.

This study also has some strengths which are attributed to the quantitative approach used. Despite the concerns about representativeness, this study is based on a relatively large sample and one could argue that this makes it possible to extrapolate the results to the whole population of eligible collaborators. In addition, this study covers a larger span of information compared with previous qualitative studies published on this topic, as we have used a quantitative method.

Conclusions

The majority of respondents in this study reported having used a CIP in their previous work and that there was a particular template available in their current organisation. Furthermore, respondents perceived positive consequences of collaborative work following establishment of an action plan, which could work as reinforcement for more collaboration. Nonetheless, results reveal that these action plans are currently used at a relatively low level. Thus, it appears as though the level of collaboration among professionals in psychiatric care, substance use treatment services and social services is low. In particular, results reveal that the level of collaboration appears to be the lowest with regard to collaborating with substance use treatment services. This study has not investigated why this is the case. However, these obstacles to collaboration may be explained by individual-level factors as well as other factors. This forms the basis for further studies.