Research on characteristics of effective collaborative relationships and their effects on community-based programming come from diverse fields. Extrapolating from the motivational interviewing literature, collaborative relationships may be most effective when TA providers use encouragement, support and sensitivity to empower communities, and respond to the unique needs of particular groups working to implement various programs (Fixsen et al., 2005). Furthermore, within corporate settings, TA is often provided through executive coaching, in which success is maximized through the development of supportive, collaborative interactions (Kilburg, 1996). The field of prevention science has just begun to explore these topics, and findings thus far are inconclusive. One study found that the collaborative relationship between the TA provider and lead implementer/team related positively to future coalition internal processes (Chilenski et al., 2016). In addition, Becker and colleagues (2013) found small but significant links between coaching-based technical support activities and improved program implementation within school settings. However, a similar analysis of the same project found that the quality of the collaborative working relationship between the TA provider and lead implementer (i.e., teacher) was unrelated to the quality of evidence-based program implementation, though the study was slightly under powered at the level of the teacher (Domitrovich et al., 2015).

Collaborative working relationships have a history of measurement in the field of Prevention Science. The quality of the working relationship between the TA provider and lead implementer(s) at the level of a program is sometimes described broadly as an alliance. More specifically, in studies of school-based EBI, the quality of the working relationship between the TA provider and teacher has been described as the teacher-coach alliance, and measured by teachers (Wehby, Maggin, Moore Partin, & Robertson, 2012). Another study found four sub-dimensions within the teacher-coach alliance when measured from the teacher’s perspective: quality of the working relationship, descriptions of the coaching process, perceived investment in the coaching experience, and perceived benefits of the coaching experience. Ratings from coaches and teachers on the same dimensions have been found to correlate moderately, suggesting some overlap but also some unique understandings from the two differing perspectives (Johnson, Pas & Bradshaw, 2016).

The quality of working relationships among TA providers and coalition leaders/members have been operationalized in a variety of ways. Most often, indicators of successful, collaborative, productive working relationships are measured among coalition members without considering the quality of the working relationship with the technical assistance provider. Measures typically have included member participation, meeting productivity, effective leadership, communication, group cohesion, shared decision-making, and supportive environments (Butterfoss, Goodman, & Wandersman, 1996; Wolff, 2001).

Other than our prior work (Chilenski et al., 2016), the quality of the working relationship between the TA provider and coalition leader or members has only just started to gain more attention as a contributor to positive coalition outcomes. Most often, it is measured qualitatively with an emphasis on developing a partnership (Cheadle, et al., 2016; Le, Anthony, Bronheim, Holland, & Perry, 2016; Katz & Wandersman, 2016).

The activities of coalitions theoretically progress through a number of phases. As others, we consider three broad phases of coalition functioning that can be cyclical: organization, implementation, and sustainability (Butterfoss & Kegler, 2002; Livet & Wandersman, 2005). Each of these phases is associated with different challenges and tasks. The first phase, organization, typically involves team formation activities and planning activities. In this “startup” phase, a group of stakeholders comes together to focus around a particular set of goals, sets its operational procedures, and receives training. The second phase focuses on implementation. In this phase, the coalition begins its operations, actively planning, implementing EBIs, and monitoring implementation quality. The third phase, sustainability (and maintenance), is where activities become routine; longer-term funding sources are pursued.

Given that the nature of TA will vary by phase of coalition development (Butterfoss & Kegler, 2002), the collaborative working relationship between the TA provider and lead implementers would also likely change over time. In an earlier study, we found that the quality of the TA provider-community team working relationship at the implementation phase predicted the team’s functioning at the sustainability phase (Chilenski et al., 2016).

From the start, readiness factors identifiable prior to coalition formation may influence the coalition’s subsequent success. For example, Feinberg, Greenberg, and Osgood (2004) found that a readiness index—including community leaders’ ability and willingness to collaborate, the community’s history of successfully addressing problems, and the extent to which community members were engaged in their communities— predicted later ratings of coalition effectiveness. Furthermore, internal team functioning, involving the relationships among team members and their ability to work together effectively, mediated this association. Greenberg and colleagues partially replicated this finding later with a new set of coalitions and longitudinal design (Greenberg et al., 2007).

Levels of community poverty and community norms may also influence the relationship between the TA provider and the coalition. For example, community-level poverty has been identified as a factor contributing to reduced effectiveness of community coalitions, possibly because of scarce resources and relatively weak or unstable infrastructures such as schools in low-income communities (Brown, Feinberg, & Greenberg, 2010; Greenberg et al., 2007). Community-level norms, beliefs, and attitudes have been shown to impact the functioning of coalitions (Brown et al., 2010; Feinberg et al., 2007). For example, Feinberg et al. (2007) found that team members’ perceptions of community norms favorable to youth substance use were associated with better team leadership and work efficiency, and were negatively related to tensions between team members. Although this seems initially counterintuitive, it is possible that this perception galvanizes team members to prepare for an “uphill battle” and motivates them to engage more collaboratively with the TA provider.

PROSPER intervention community teams participated in a series of formal and informal TA activities as part of a proactive TA model; see Chilenski et al. (2016) for more details. Each community team partners with a specific TA provider called a Prevention Coordinator who has expertise in coalition development and implementation of evidence-based programs. The Prevention Coordinator “links” the community-level team to knowledge and resources that are typically available at the state/university-level (Spoth et al., 2004). Each Prevention Coordinator designated 25% FTE per team. During the timeframe of the current study, the Prevention Coordinator undertook multiple responsibilities, from attending their teams’ monthly meetings, to biweekly phone calls with the community team leader, to conducting additional, as-needed communications with team leaders or members. The PROSPER model also included an annual statewide meeting to promote professional development, and learning communities began 2.5 years into the project. Prevention Coordinators received supervision and support from University/state-level prevention researchers and Extension faculty at least weekly.

Data for this project were collected from: 1) Prevention Coordinators, 2) community prevention team members, 3) directors of local service organizations, and the 4) U.S. Census. See Table 1 (online) for an overview of the intervention and assessment schedule timeline.

TA Collaboration fluctuates over time (grand mean = 0.00; SD = 0.72; Min = −2.67; Max = 0.93). The communities have a relatively low percentage of families that are living below the federal poverty threshold (M = 6.95; SD = 2.49) and report slightly favorable perceptions of readiness (M = 2.79; SD = 0.28). Levels of community norms of adolescent substance use indicate that individuals slightly disagreed that adolescent substance use was a normal in their communities (M = 2.29; SD = 0.24).

Table 2 displays the results for our first hypothesis, the model fit statistics for the baseline, Random Intercept-Only (i.e., Empty) Model, and the four next best-fitting models. The best-fitting model is the piecewise model that allows for different slopes to be independently estimated across each phase of team functioning. The fit statistics of the piecewise model that allows a random slope during the implementation phase are the lowest across all models (also see Figure 1, available online). Indeed, ratings of TA-team collaboration were relatively stable across all teams during organization. Then, levels of collaboration generally decreased during implementation and the degree and direction of change differed by community. Levels of collaboration then increased equivalently for all teams during sustainability.

Table 3 summarizes the results of hypotheses 2a, 2b, and 2c. The only community characteristic that predicted levels of TA Collaboration was readiness rated by Agency Directors (B = 1.49, p = .007). Readiness predicted initial levels of TA Collaboration. The interaction between AD-rated readiness and the slope of TA Collaboration (B = −1.27, p = .02) during the implementation phase significantly predicted variation in community differences in how levels of TA Collaboration changed over the implementation phase (see Figure 2). Specifically, communities with high ratings of readiness had the highest levels of team-PC collaboration during the organization phase, but had the greatest decrease during the implementation phase. Teams with average levels of readiness had moderate levels of team-PC collaboration during organization and demonstrated a small decrease during implementation. However, communities with the lowest ratings of readiness had the lowest ratings of collaboration during the organization phase but increased their levels of collaboration during the implementation phase.

We hypothesized that a model that accounted for the distinct phases of coalition development would be the best way to describe the data; this hypothesis was supported. This finding adds to prior evidence of measureable differences across phases in community coalition prevention or public health efforts (Butterfoss & Kegler, 2002; Welsh et al, 2016). Specifically, TA collaboration is relatively high during organization and decreases during implementation, then increases again during sustainability. Based largely on conversations with Prevention Coordinators and Team Leaders over the duration of the PROSPER project, we speculate that this ebb and flow may relate to the different responsibilities and challenges faced by Team Leaders in different phases. At the time that PROSPER began, most Team Leaders were largely unfamiliar with prevention science principles or research protocols. While organization involved many meetings and discussions that were not particularly demanding, implementation involved responsibility for a wide array of new and time-consuming tasks, from recruiting families to attend programming to supervising program facilitators, monitoring implementation, and continuing to coordinate the community team. During this phase, many Team Leaders expressed feeling overwhelmed and indicated that PROSPER had too many “moving parts.” Thus, the overall decline in collaboration observed during the implementation phase was related to the stress associated with these experiences.

Activities became more routine 2.5 years into the project. At that point teams had two years of family program implementation and one year of school-based program implementation completed; no new programs were added. Hence, this time period began the transition into sustainability. Teams had also been working with their assigned Prevention Coordinators for a length of time. During the transition to sustainability, collaborative learning communities began. The learning communities brought together community team leaders within each state, Prevention Coordinators, state-level prevention experts, and at times other experts in specialized topics relevant to the challenges of sustaining youth programming. Organization and planning of these learning communities initially was conducted by the Prevention Coordinators and state-level prevention experts, but this changed over time to being driven by community team leaders (Mincemoyer, Perkins, & Santiago, 2008). A culture of honest, supportive and constructive discussion was created. The TA process became more reflective and oriented towards professional development, rather than task-specific. Perhaps the collaborative working relationship may have improved because the Team Leaders and members now recognized the value of TA at helping them to hone the new and relevant skills to this phase.

We hypothesized that community characteristics would be related to the quality of the collaborative coalition-TA provider working relationship. Specifically, prior research led us to anticipate that community-level poverty would pose a barrier to effective collaboration, but higher levels of readiness and social norms endorsing youth substance use would lead to an improved relationship. This hypothesis was largely unsupported. Neither community level poverty nor social norms were related to collaboration. Unlike community team members, Prevention Coordinators were typically “outsiders” who were familiar with, but did not live in, the PROSPER communities. Perhaps, the internal team dynamics are more sensitive to community-specific factors (e.g., poverty and social norms) than the team-TA relationship.

Our data suggest that associations between community readiness and TA collaboration were interesting and complex. There is substantial community-level variability in how much and in what direction TA collaboration changed over time. Examining these changes based on initial levels of readiness, the sites rated highest in readiness at the start of the project also started out high in collaboration, but experienced relatively steep declines during the implementation phase, and ended up having the lowest levels of collaboration during the sustainability phase. The lowest readiness sites were the lowest in collaboration initially, but steadily improved over time, ending up with the highest levels of collaboration during the sustainability period. There are several possible interpretations of these findings. First, perhaps leaders of low readiness teams may have had more difficulty getting their initiative off the ground, and this rough beginning was reflected in their poorer initial ratings of collaboration with their Prevention Coordinators. However, as time progressed, these Team Leaders were able to form good collaborative relationships around program implementation and sustainability. For these teams, a low level of confidence and readiness at the outset may have created fertile ground for a supportive and helpful TA provider that then endured over time. Yet, the decline in collaboration observed in the highest-readiness teams may have occurred because these Team Leaders were more confident at the start and did not perceive the need for or value TA. During the implementation phase, then, they may have perceived TA as bothersome during a busy time (Feinberg et al., 2004).

Based on these findings, we conclude that the collaborative relationship between a TA provider and community prevention teams may be a critical component for the success of coalitions over time, supporting their positive impact on public health. Thus, we recommend:

There are a few limitations in this study. This is a relatively small sample of community prevention teams who were all focused on reducing and delaying early adolescent substance use with very similar strategies. This focused sample may have reduced potential variability in the data which could weaken statistical power. These findings are most confidently generalized to community prevention teams in similar contexts. It is also possible that there were some unmeasured differences in implementation that affected the quality of the collaborative relationship. We were limited to the number of variables that we could investigate simultaneously; as a result, we constrained our analyses to investigate one community characteristic at a time. We also do not have team leader or team member ratings of the PC-team collaborative relationship to validate the ratings with a different reporter. Future research with a larger number of community prevention teams is desirable. Longitudinal data and different reporters for the independent and dependent variables in this study contribute to its strength.

Our study examined community characteristics and their association to TA collaboration, as well as the way in which this collaboration ebbed and flowed over time. Future studies should extend our research by exploring the relations between collaboration and phase-specific outcomes. Additionally, research on TA to community coalitions could examine the role of specific TA strategies and styles for community prevention team functioning and outcomes of their initiatives. Additional knowledge in these areas will help evidence-based interventions achieve their full potential on improving public health.