Identifying capacity building priorities for implementation science in suicide prevention: a concept mapping study in Australia

Abstract

Background

Capacity building in the science and practice of implementation is critical for driving and evaluating meaningful change. However, capacity needs of stakeholders to effectively apply implementation science in suicide prevention remain unclear.

Purpose

To address this gap in Australia, we conducted a concept mapping (CM) study to identify and prioritize stakeholder capacity-building needs in this area.

Methods

Diverse stakeholders: practitioners, lived experience advocates, researchers, public servants, and clinicians working within the suicide prevention sector in Australia participated in the study. Using CM, participants individually brainstormed statements representing capacity-building needs for implementation science for suicide prevention. These statements were then organized in clusters and rated on importance and feasibility. Data were collected, analyzed, and visually represented using the groupwisdom™ platform.

Results

In total, 52 stakeholders participated in at least one stage of the study (i.e. brainstorming, sorting and/or rating). Brainstorming generated 59 statements, which were synthesized into 53 distinctive statements and supplemented with 13 statements from the literature. A multidimensional scaling and hierarchical cluster analysis generated many cluster solutions, out of which an eight-cluster solution was chosen as the most meaningful. Knowledge dissemination and lived experience integration clusters were rated highest on importance and feasibility.

Conclusions

This study provides a snapshot of implementation capacity-building needs and priorities within the suicide prevention sector in Australia. The identified priorities can help inform future sector capacity-building initiatives for implementation.

Teaser

What does it take to build implementation capacity in suicide prevention? Our study shows that training alone may not be enough; highlighting the need for coordinated, system-level initiatives to meet diverse stakeholder needs and drive meaningful impact.

Implications.

Practice: Programs to strengthen capacity for implementation in suicide prevention must be conceptualized and delivered across multiple system levels. Such programs must move beyond individual training and extend their reach beyond academic and research audiences.

Policy: Concerted efforts toward incorporating diverse perspectives of stakeholders in the sector are needed to foster partnerships, collaboration, and enable policy environments that strengthen sector-wide evidence translation.

Research: The study highlights gaps in stakeholder-specific capacity needs and tailored implementation strategies to effectively deliver these targeted initiatives.

Introduction

Suicide is a complex, multifaceted issue, making its prevention particularly challenging [1]. Recently, there is an emphasis on a public health approach to prevention, recognizing that risk factors are rooted in broader social determinants [2]. However, despite growing investment, there remains limited translational evidence to guide policy and practice [3]. Significant methodological challenges hinder understanding of what works, for whom, and in which contexts [4].

Implementation science is “the study of methods to promote the systematic uptake of evidence-based interventions into practice and policy to improve health” [5] (p. 1). The field has two key focus areas: implementation research and implementation practice. Implementation research seeks to understand and evaluate approaches used to translate evidence to the real world [6]. Implementation practice seeks to apply and adapt these approaches in different contexts and settings to achieve positive outcomes [6]. Despite its growth and integration across multiple disciplines, implementation science continues to face limitations in its broader applicability [7].

The use of implementation research and practice in suicide prevention remains inconsistent and under-prioritized [8, 9]. Contributing factors for this challenge arise from both suicide prevention and implementation science. These include methodological challenges, limited translational examples, inadequate stakeholder collaboration [3, 10, 11]; disciplinary silos [12]; limited guidance on co-production and stakeholder engagement [13, 14]; and rigid adherence to evidence-related paradigms [7]. Beyond a paradigmatic shift, strengthening implementation science in suicide prevention may require building capacity across the spectrum, from general awareness to advanced application and training of implementation science [7, 15, 16].

Capacity building for implementation science has been described as activities (e.g. training, identification of alternative resources, building internal assets) that build durable resources and enable the recipient setting or community to continue the delivery of an evidence-based intervention [17]. Six strategies for capacity building have been identified: training, tools, technical assistance, assessment and feedback, peer networking, and incentives. Capacity for implementation research and practice is specialized and multidimensional, requiring not only knowledge of implementation theory and methods but also competencies in stakeholder engagement [18]. Identified competencies span setting-specific knowledge, research methodology, management, communication, and collaboration [6, 18, 19]. Although stakeholder needs for capacity building have been explored through interviews, surveys, and concept mapping (CM) [19–21], gaps remain in understanding practical skill development and the “art and craft” of implementation [22].

Although recognition of implementation science in suicide prevention is increasing [23], its inconsistent application [8] suggests a significant capacity gap within the sector. This issue has attracted growing attention from policymakers and sector leaders, alongside broader efforts to strengthen the translation, implementation, and sustainability of suicide prevention initiatives. Our previous work has highlighted the limited and variable use of implementation science approaches in suicide prevention and identified a need to build implementation capacity among stakeholders [8]. Our international Delphi study with stakeholders across the sector further identified capacity building as a key strategy for advancing the use of implementation science in suicide prevention, highlighting the need to better understand stakeholders’ ability to apply these approaches and address the evidence–practice gap [24]. Despite growing interest in this area, there remains limited research examining implementation science capacity-building needs specific to suicide prevention. This study therefore builds on this emerging body of work by identifying and examining stakeholder capacity-building needs and priorities in Australia to support the effective use of implementation science in suicide prevention.

Methods

The study adopted a CM approach, a systematic participatory approach that integrates qualitative data collection with quantitative analytical techniques [25]. Unlike surveys and interviews, CM moves beyond the identification and examination of perspectives by enabling participants to generate, organize, and prioritize ideas [26]. Moreover, rather than encouraging convergence towards consensus, as is the aim of Delphi methods, CM accommodates diverse perspectives and provides a structured means of understanding both common priorities and differing perspectives across stakeholder groups [27]. This is particularly important in suicide prevention, a multidisciplinary field characterized by a diverse range of stakeholders whose differing roles, experiences, and priorities may influence perceptions of implementation science capacity-building needs.

Data collection comprises two primary activities: (i) brainstorming responses that characterize the issue of interest, which are based on a focus prompt (“To build capacity for implementation research and practice in suicide prevention, we need…”); and (ii) sorting the statements generated during brainstorming into groups of conceptually related ideas, followed by rating them on relevant scales (e.g. importance and feasibility). Although the literature recommends a minimum sample size of 10 participants each for sorting and rating, many studies comprise samples of 20 or more participants [28]. Data analysis involves multidimensional scaling (MDS) and hierarchical cluster analysis to position each statement as a point that can be visually displayed across a series of conceptual maps. These visualizations illustrate the relationships among statements [27].

Recruitment of participants

Participants were recruited using purposive and snowball sampling to ensure the inclusion of individuals with relevant expertise and experience in suicide prevention and implementation science. Consistent with CM methodology, the aim was to capture a breadth of informed perspectives on the topic rather than achieve population representativeness [29]. Participants were recruited using the Australian Institute for Suicide Research and Prevention’s email distribution list, social media links, and via a newsletter from the LIFEWAYS project. Additionally, members of the research team shared a study flyer on social media. Potential participants were provided with an overview of the study with links to an online participant information sheet and consent form. They were encouraged to share this invitation with others who may be interested in the study. After participants completed the consent form, they were directed to the Concept Systems groupwisdom™ data collection platform to create an account. Participants completed the five socio-demographic questions: the role(s) they identified with (researcher, practitioner, clinician, lived experience, public servant); the primary role they identified with; their primary work setting (university, research institute, not-for-profit organization, health authority, government, private sector, other); age; and gender.

Data were collected from October to December 2025, with a three-week gap between the brainstorming and sorting tasks. Member checking was not undertaken, considering the participatory nature of the CM process with participants contributing directly to the generation, structuring, and interpretation of ideas. All the data collection and analyses were conducted using the groupwisdom™ Group Concept Systems platform. Griffith University Human Research Ethics Committee approved all procedures for this study (2025/764).

Brainstorming

The following definitions were used to prepare participants for the brainstorming task:

• Implementation: Involves efforts to put evidence-based programs or practices into use by applying effective strategies for change [30].

• Implementation research: Studies how to move evidence into real-world use to achieve desired outcomes [6].

• Implementation practice: Focuses on applying and adapting these evidence-based approaches in different settings to achieve desired outcomes [6].

• Capacity building: Involves activities that strengthen skills, resources, and support to enable an individual, community, or organization to continue delivering an evidence-based program [17].

Participants were directed to consider these definitions and provide responses to the focus prompt: “To build capacity for implementation research and practice in suicide prevention, we need…”. There was no limit to the number of responses/statements that could be provided. This task was open for three weeks with two reminder emails sent to participants about completing brainstorming.

Members of the research team (S.K., L.T.B., L.R., K.K.) reviewed and synthesized the resulting statements over two meetings until these were distinct and clear. The original wording of statements was retained as much as possible. Complementary statements (n = 13) were added from research literature on implementation science-related capacity-building initiatives to ensure that important aspects identified in prior research were represented in the study.

Statement sorting and rating

The sorting task asked participants to “Sort the statements into piles in a way that makes sense to you. Group the statements on how similar in meaning or closely related they are to one another.” The participants were presented with the statements in random order and were instructed to create between 5 and 15 groups of statements, assigning descriptive names to each group. They were advised to avoid grouping statements by priority or placing unrelated statements into a generic “other” category.

Participants were then asked to rate each statement in terms of importance and the feasibility of accomplishing the activities. Specifically, participants were asked to rate, on a five-point scale, the importance of each statement as an activity/strategy for capacity building in implementation research and practice in suicide prevention (1 = less important; 5 = more important), and the feasibility of addressing each activity/strategy (1 = less feasible; 5 = more feasible).

Data analysis

After all activities were completed, the data were reviewed to confirm usability. Participants with incomplete submissions (e.g. completion of only one of the two rating tasks) were prompted to finish the remaining activities through personalized emails. Participants who sorted fewer than 75% of the statements into meaningful categories were excluded from analysis. The 75% completion threshold represents a pragmatic data quality criterion commonly used in CM practice. To minimize the influence of uniform response patterns (i.e. the tendency to assign similar ratings across all statements), participants were informed that the rating scales should be interpreted as relative rather than absolute measures and were encouraged to use the full range of response options available. Participants’ rating data were excluded if their responses did not span at least three of the five available scale points (i.e. 1–3, 2–4, 3–5) [31]. This was applied to ensure sufficient discrimination between statements and meaningful contribution to the prioritization analyses [26].

Statistical analyses began with the construction of a similarity matrix, which served as input for nonmetric MDS to generate a two-dimensional point map. The MDS procedure produced a stress value (0–1) to evaluate the goodness of fit between the point map and the original sorting data. Hierarchical cluster analysis using Ward’s method was then applied to the point map to derive a cluster solution consisting of nonoverlapping clusters. Because there is no established mathematical criterion for determining the optimal number of clusters in CM, we followed recommended practice by first identifying the maximum number of clusters deemed interpretable [32]. The final cluster solution was selected through an iterative review of neighboring cluster solutions. This process involved consideration of quantitative indicators (e.g. cluster bridging values) alongside conceptual coherence, interpretability, and alignment with the study aims. While some degree of researcher judgment is inherent in this process, decisions were made collaboratively by the research team and guided by established CM principles (Supplementary Material 1).

This process began with 10 clusters, which was informed by starting with three clusters above the mean and mode of the number of piles generated by participants during the sorting task. Cluster solutions were then examined iteratively by progressively reducing the number of clusters, with decisions at each level guided by judgments of conceptual coherence and practical interpretability [32]. By reviewing the 10- to 5-cluster map solutions, we found the cluster solution that included distinct conceptual differences useful for interpretation, whilst merging clusters that were less distinct and therefore a better conceptual fit in the context of this study [31]. Finally, the eight-cluster solution was retained because it provided the most meaningful balance between parsimony and conceptual distinction.

We then reviewed each statement on the periphery of clusters to determine whether any statements were a better conceptual fit in an adjacent cluster, which would allow for the redrawing of cluster boundaries. Decisions regarding the redrawing of cluster boundaries followed a structured, collaborative review process among the research team. Specifically, modifications were only considered when: (i) a statement appeared to be a poor conceptual fit within its assigned cluster; (ii) represented a stronger conceptual fit within a neighboring cluster; (iii) did not create overlapping cluster boundaries; and (iv) were supported by similarity matrix and spanning information available within groupwisdom™.

First, the similarity matrix generated by the groupwisdom™ platform was examined to determine the frequency with which pairs of statements were sorted together across participants. Next, spanning information was reviewed to evaluate how often statements considered for reassignment were grouped with statements in neighboring clusters relative to their original cluster. These indicators were used to guide boundary adjustments in support of coherent conceptual interpretation. Finally, cluster labels were discussed by the research team. These were assigned based on the thematic content of the statements within each cluster and the cluster names proposed by participants.

Mean ratings were calculated for each statement and used to construct a two-dimensional Go-Zone map. The Go-Zone graph displays the mean importance ratings on one axis and the mean feasibility ratings on the other, thereby visually representing the relative positioning of statements across both dimensions. To examine the relationship between perceived importance and feasibility, a Pearson’s correlation coefficient was calculated between the mean importance and mean feasibility ratings for each statement (n = 66 statements).

Results

Participants

In total, 52 participants engaged in at least one activity. Consistent with CM methodology, participation varied across brainstorming (n = 25), sorting (n = 44), importance rating (n = 36), and feasibility rating (n = 35) activities, as participants were not required to complete every phase.

The mean age of participants was 50.57 years. Most participants were women (n = 38; 73.08%). Participants reported their primary role as practitioner (program/project staff, manager) (n = 20; 38.46%), lived experience (n = 11; 21.15%), researcher (n = 8; 15.38%), public servant (i.e. working in government) (n = 7; 13.46%), or clinician (n = 5; 9.62%); one participant reported not applicable. A range of different work settings were reported, with health authority (including hospital, community) (n = 11; 21.15%), not-for-profit organization (n = 11; 21.15%), and government (n = 10; 19.23%) being the most common, followed by university (n = 6; 11.54%), other (n = 5; 9.62%), and the private sector (n = 4; 7.69%). Total numbers of participants who engaged in each activity are presented in Table 1, with a flowchart showing the number of participants who started and completed each activity in Supplementary Material 2.

Table 1.

Description of participants per activity.

	Brainstorming a	Sorting	Rating (importance)	Rating (feasibility)
Primary role (total)	25	28	35	31
Lived experience	3	3	7	7
Practitioner (program/project staff, manager)	8	16	16	16
Public servant (i.e. working in government)	3	3	3	1
Clinician	3	2	3	2
Researcher	6	4	6	5
Primary work setting
Not-for-profit organization	2	9	10	8
University	5	2	4	4
Health authority (including hospital, community)	3	9	9	8
Government	7	3	4	4
Private sector	1	2	3	3
Research institute (within a health authority or university)	2	2	2	2
Other	3	1	3	2
Gender
Male	5	5	6	5
Female	17	22	28	25
Other	1	1	1	1
Age
Below 45 years	5	8	9	6
45 to <60 years	13	12	16	16
60 years and over	5	8	10	9

^aTwo participants from the brainstorming activity did not respond to participant questions. Therefore, the brainstorming column includes data from 23 of the 25 participants.

Brainstorming and sorting

Participants (n = 25) generated 59 statements that were synthesized into 53 distinctive statements, which were supplemented with 13 statements from the literature (Table 2). Adding statements from the literature is a suggested practice for generating a rich and varied set of statements for CM [32]. The added statements were identified from seminal research publications on capacity building initiatives conducted in implementation science. Statements were particularly chosen from studies in which diverse stakeholder perspectives on these initiatives were reported. The final list of 66 statements was sorted by participants (n = 28) into 3–17 groups (mean = 6.89, median = 7.00, mode = 7.00).

Table 2.

Capacity building for implementation and research in suicide prevention (synthesized statements) grouped by cluster, including bridging value, mean importance and feasibility, and Go-Zone quadrant

No	Statement	Bridging value	Importance mean score (1–5)	Feasibility mean score (1–5)	Go-Zone quadrant
CL 1	Resources and leadership commitment	0.66	4.30	3.20
1	To ensure funding opportunities prioritize implementation measurement in suicide prevention activities	0.85	4.33	3.55	2
2	Supportive employers who can allow for emphasis on implementation as a key focus in their research/practice agendas	0.64	4.25	3.42	2
3	Supportive government who can allow for emphasis on implementation as a key focus in their research/practice agendas	0.65	4.31	2.79	2
4	Supportive funding bodies who can allow for emphasis on implementation as a key focus in their research/practice agendas	0.73	4.47	3.28	2
10	To prioritize allocation of implementation resources (time, infrastructure, expertise, guidance materials)	0.60	4.25	3.48	2
14	Enhancements that strengthen outcomes without overtaxing existing systems	0.56	4.00	3.18	2
34	Organizational commitment to building capacity in implementation	0.56	4.53	3.30	2
CL 2	Knowledge dissemination	0.48	4.25	3.88
29	Clear communication strategies that help to translate evidence into practice	0.36	4.19	4.12	1
48	To make sure that outcomes of research are disseminated for use in real-world settings	0.54	3.89	3.85	1
54a	To present research results in a simple and intuitively understandable way	0.42	4.11	4.03	1
58a	To build effective relationships and partnerships with community partners, practitioners, and policymakers	0.58	4.44	3.78	1
CL 3	Lived experience integration	0.83	4.14	3.89
9	To value the vital contribution of people with lived experience in implementation	1.00	4.50	4.36	1
21	Jargon-free language	0.85	4.00	4.03	1
23	Peer and/or expert support given to implement real-world projects	0.57	4.17	3.70	1
31	Lived experience insights	0.83	4.42	4.24	1
53	Upskill and involve non-academic lived experience researchers	0.85	4.29	3.73	1
65a	Communities of practice for sharing experiences, resources, and lessons learned	0.95	4.03	4.41	1
38	To build on current strengths	0.65	3.92	3.70	3
50	Individuals serving as bridges between theory and practice	0.98	3.86	3.39	4
CL 4	Ground-up informed initiatives	0.77	4.01	3.61
43	Ground-up insights driving intervention design and delivery	0.67	4.14	3.91	1
44	Ground-up insights driving research design and delivery	0.73	4.20	3.76	1
52	Co-design and co-lead research implementation projects	0.77	4.11	3.76	1
20	Involvement of frontline clinicians in study design	0.79	4.08	3.45	2
30	Qualitative data collection methods	0.88	3.72	3.91	3
61a	To build and manage interdisciplinary research teams to partner across disciplines and perspectives	0.76	3.89	3.30	4
CL 5	Strengthening the evidence base	0.57	3.87	3.61
11	To identify implementation gaps and opportunities to guide strategic action	0.55	4.17	3.79	1
15	Indicators of success to guide resources to where they are most needed	0.72	4.00	3.63	2
18	To identify current practices	0.54	3.43	3.97	3
19	Evaluate current practices	0.45	3.94	3.82	3
41	To identify best evidence base for implementation	0.63	3.94	3.91	3
64a	To use evidence to tailor and adapt implementation strategies	0.47	3.92	4.06	3
17	Knowledge about what is working elsewhere	0.51	3.92	4.03	3
5	Additional funding to hire expert implementation staff, even on short-term or casual basis	0.86	3.86	2.73	4
8	To integrate and harness implementation thinking in upstream and population-level prevention activities	0.43	3.89	3.45	4
16	Best practice	0.50	3.81	3.53	4
49	To prioritize implementation practice	0.59	3.89	3.47	4
CL 6	Hands-on implementation support	0.44	4.00	3.65
25	Practical steps to apply to projects/outcomes	0.35	4.06	3.91	1
36	Understanding of implementation strategies in real-world contexts (i.e. how it intersects with project planning, resourcing etc.)	0.59	4.36	3.88	1
37	Understanding of implementation science in real-world contexts (i.e. how it intersects with project planning, resourcing, etc.)	0.57	4.26	3.82	1
28	Implementation support	0.42	4.03	3.58	2
22	Hands-on project implementation	0.58	3.94	3.79	3
12	Expert models to guide and educate	0.25	3.56	3.64	4
32	To understand change management for improved systems thinking	0.35	3.83	3.45	4
CL 7	Education	0.26	3.83	3.65
51	To clearly explain implementation research in plain, accessible language	0.40	4.00	4.18	1
60a	To learn strategies to increase readiness of practitioners and organizational leaders	0.20	4.08	3.70	1
57a	To train community stakeholders to identify needs, select and adopt evidence-based practices	0.24	4.00	3.64	2
39	To identify competencies	0.24	3.42	3.82	3
40	Skills in stakeholder engagement	0.18	3.83	3.76	3
55a	Mixed methods training—better integration of qualitative and quantitative	0.21	3.71	3.79	3
56a	To instill in trainees the relevance and usefulness of research from the start	0.39	3.85	4.03	3
59a	To learn about tools for translating research into action	0.19	3.97	3.85	3
7	Ready on-call or on-demand access to expert implementation guidance	0.23	3.78	3.00	4
13	Expert mentors to guide and educate	0.34	3.82	3.38	4
CL 8	Training and skill development	0.10	3.73	3.57
26	Implementation training	0.06	4.08	4.06	1
35	Skills to incorporate implementation into evaluation and research studies	0.22	4.06	3.85	1
63a	To develop skills in understanding contextual factors in implementation (setting characteristics, culture, capacity, and readiness)	0.04	4.33	3.76	1
6	Education and training via easy-to-access online modules and/or interactive webinars	0.08	3.80	3.94	3
27	Implementation education	0.06	3.89	3.94	3
66a	To receive training in project management	0.03	3.03	3.81	3
24	Guided program where learning is provided	0.06	3.57	3.66	4
33	To receive training in behavior change strategies	0.00	3.77	3.64	4
42	Skill in developing program logic	0.14	3.49	3.64	4
45	Acceptable implementation science training programs	0.03	3.57	3.55	4
46	Sustainability of implementation science training programs	0.31	3.57	2.97	4
47	Implementation science training that reaches all parties involved	0.24	3.57	3.21	4
62a	Tailored specialist training materials for diverse audiences	0.10	3.89	3.36	4

^aDenotes statement added from literature. Statements highlighted in bold were rated above the mean for both importance and feasibility (Quadrant 1 of the Go-Zone map).

Concept map

An eight-cluster map was chosen to represent the data by the research team members as it preserved meaningful distinctions between domains that seemed theoretically and practically different, allowing each cluster to represent a coherent, yet distinct concept (Fig. 1). Clusters were labeled as: (i) Resources and leadership commitment, (ii) Knowledge dissemination, (iii) Lived experience integration, (iv) Ground-up informed initiatives, (v) Strengthening the evidence base, (vi) Hands-on implementation support, (vii) Education, and (viii) Training and skill development (see Table 2 for the statements within each cluster). When a statement was considered a better conceptual fit in a neighboring cluster, and this judgment was supported by the spanning data (i.e. frequency with which participants grouped statements together), cluster boundaries were adjusted accordingly. We redrew cluster boundaries to allow for statement 17 (“Knowledge about what is working elsewhere”) to be included in cluster 5 (“Strengthening the evidence base”); and statements 6 (“Education and training via easy to access online modules and/or interactive webinars”) and 62 (“Tailored specialist training materials for diverse audiences”) to be positioned within cluster 8 (“Training and skill development”). The stress value of the final map was 0.308. Previous syntheses of CM studies have reported mean stress values of ∼0.28 (SD = 0.04), with most studies falling within accepted ranges for CM applications [33]. Similarly, Donnelly (2017) [34] reported stress values ranging from 0.13 to 0.36 across 96 concept maps and concluded that values in the range of 0.29–0.33 remain within the typically accepted range.

Figure 1.

Eight-cluster map.

Rating

Statement ratings are reported for importance (n = 35) and feasibility (n = 31) in Table 2. The clusters regarded as the most important activity/strategy for building implementation research and practice capacity in suicide prevention (based on mean scores) were Resources and leadership commitment (4.30) and Knowledge dissemination (4.25). The least important cluster was Training and skill development (3.73). The clusters regarded as the most feasible capacity building activity/strategy were Lived experience integration (3.89) and Knowledge dissemination (3.88), with Resources and leadership commitment (3.20) deemed the least feasible.

Go-Zone map

Figure 2 presents the Go-Zone map, with four quadrants: the top-right (“Go-Zone”) includes statements rated above the mean for both importance and feasibility (Quadrant 1); the bottom-right includes statements high in importance but low in feasibility (Quadrant 2); the top-left includes statements high in feasibility but low in importance (Quadrant 3); and the bottom-left includes statements rated low on both importance and feasibility (Quadrant 4).

Figure 2.

Go-Zone map.

There was a weak positive correlation (r = 0.107; 95% CI [–.138, .341], p = .391) between the perceived importance of activities/strategies for capacity building implementation research and practice in suicide prevention and the perceived feasibility of addressing them. The top right-hand quadrant is the “Go-Zone,” which includes statements with the highest (above the grand mean scores) relative importance and feasibility ratings (see Fig. 2 for cluster mean ratings for importance and feasibility). A total of 22 statements were in the Go-Zone. All four statements from the Knowledge dissemination cluster, and most (6/8; 75%) of the statements from the Lived experience integration cluster were in the Go-Zone. Half (3/6) of the statements from the Ground-up informed initiatives cluster were included in the Go-Zone, with 42.9% (3/7) of statements from the Hands-on implementation support cluster also included. Both the Training and skill development cluster (3/13) and Education cluster (2/10) had at least 20% of statements included in the Go-Zone. One statement from the Strengthening the evidence base cluster and no statement from the Resources and leadership commitment cluster were in the Go-Zone.

Discussion

This study aimed to identify the capacity building needs and priorities of various stakeholders in Australia to effectively utilize implementation science for suicide prevention. Using CM, our study identified an eight-cluster concept map highlighting capacity building needs across eight areas/clusters: (i) resources and leadership commitment; (ii) knowledge dissemination; (iii) lived experience integration; (iv) ground-up informed initiatives; (v) strengthening the evidence base; (vi) hands-on implementation support; (vii) education; and (viii) training and skill development. A variety of stakeholders including researchers, practitioners, lived experience advocates, public servants, and clinicians contributed their ideas to different phases of the study. In addition to identifying needs, the CM process also highlights priority areas by highlighting statements rated above the mean for both importance and feasibility by participant stakeholders. These can be used to inform strategies to broaden the reach, relevance, and impact of capacity building initiatives.

Priorities

Priorities for strengthening capacity in implementation research and practice can be drawn from those statements rated as both highly important as well as feasible. Overall, the knowledge dissemination and lived experience integration clusters included statements rated above the mean for importance and feasibility. These statements highlight priority areas for action. The knowledge dissemination cluster comprised the smallest number of statements yet remained critical. It emphasized the role of effectively communicating and translating research evidence to strengthen implementation capacity. All four statements in the cluster were rated above the mean for both feasibility and importance, highlighting priorities such as developing clear communication strategies (statement 29), prioritizing dissemination of research findings (statement 48), presenting research in accessible formats (statement 54), and fostering effective partnerships across diverse stakeholder groups (statement 58). In a similar CM study in the United States on implementation science training needs in health research, “communicate research findings” (related to dissemination strategies) was rated as highly important but the most difficult to implement, although no reasons were provided for this rating [19]. Dissemination has been defined as an active approach of spreading evidence or evidence-based interventions to audiences via determined channels using planned strategies [17]. Knowledge dissemination is widely recognized as essential for evidence use. But the key challenges are (i) how to do it effectively; (ii) how to build capacity for it; and (iii) how to ensure that dissemination itself can foster greater engagement with evidence in practice and support practice-informed research. Brownson et al [35] suggest strengthening dissemination by firstly, applying audience segmentation and user-centered design to tailor messages to specific stakeholder groups. Secondly, using strategic framing to translate scientific evidence into meaningful, audience-relevant narratives.

The lived experience integration cluster highlighted the value of incorporating lived experience insights to deepen understanding of how programs operate in practice. Six out of eight priority statements rated above the mean as important and feasible emphasized the use of jargon-free language (statement 21), greater involvement of non-academic lived experience researchers (statement 53), and the relevance of communities of practice (statement 65). The prioritization of the lived experience integration cluster was unsurprising given the strong representation of lived experience advocates in the sample and leadership in Australia on this issue. Beyond existing guidance on lived experience engagement in research [36] and program development [37], this cluster emphasized the need for evidence to reflect real-world experiences, be communicated in accessible language, and foster broader engagement. Thereby, highlighting the need to strengthen capacity for evidence generation and utilization among lived experience advocates.

Priority statements also emerged from the ground-up informed initiatives cluster, tied closely with the lived experience integration cluster. In this cluster, incorporating on-the-ground insights into intervention design and delivery, as well as supporting co-led research, were endorsed. Within the hands-on implementation support cluster, participants identified a need for practical steps that can be applied to specific projects/outcomes (statement 25) and a better understanding of implementation strategies and science in real-world contexts (statements 36, 37) as priorities. In the Tabak et al [19] study, the clusters “make research more relevant” and “consider and enhance fit” were both rated high on importance and moderate on difficulty. These reflect broader concepts such as practice-based evidence [38] and designing for dissemination [39]. These approaches call for an integration of the experiential knowledge of consumers, practitioners, and researchers to ensure interventions align with real-world needs and contexts. However, gaps remain in effectively designing for dissemination. The current academic system, shaped by peer review norms, journal impact metrics, and promotion criteria, does not adequately reward prioritizing practice-based evidence [40]. While researchers tend to prioritize knowledge generation, practitioners and policymakers emphasize its use. Bridging these differing incentives presents an opportunity to promote mutual learning and strengthen shared capacity building.

Two closely related clusters: education and training and skill development, also contained some priority statements. The education cluster reflected broader philosophical considerations regarding how programs should be conceptualized and designed. Here, participants endorsed making implementation research accessible in plain language (statement 51) and building the readiness of practitioners and organizational leaders (statement 60) as important and feasible. In the training and skill development cluster, more specific competencies, such as skills in conducting evaluation, were prioritized as both important and achievable. In strengthening the evidence base, identifying implementation gaps and opportunities to inform strategic action (statement 11) was the only statement prioritized. No statements from the resources and leadership commitment cluster were prioritized as both important and feasible.

Areas of divergence

Interestingly, there was a weak positive correlation (r = 0.10) between the perceived importance of activities/strategies for capacity building implementation research and practice in suicide prevention and the perceived feasibility of addressing them. This suggests that participant stakeholders in the study differentiated between what is important and what is realistically achievable. This distinction is conceptually meaningful because it indicates that structural, socio-political, and resource constraints may limit action even when the course of action is clear.

To illustrate, statements in the cluster resources and leadership commitment were rated as the most important. However, these statements were also rated as the least feasible to implement (see Table 2). Interestingly, training and skill development, which seems to be the most obvious step towards capacity building, was rated as the least important by different groups of stakeholders. It may indicate stakeholder fatigue with training-based initiatives and a need for alternative strategies. This could imply that capacity building is understood as a broader undertaking than simply engaging in skills-based training [41]. This may suggest the need to move beyond individual skill-building programs and consider systemic, structural approaches.

Implications

Although CM has been widely used in public health and healthcare settings, its application within suicide prevention remains relatively limited. Nevertheless, emerging studies have demonstrated its value for engaging diverse stakeholders and identifying priorities relating to suicide prevention systems and service delivery contexts [42, 43]. The findings of the present study extend this emerging body of work by specifically examining implementation science capacity-building needs within the suicide prevention sector.

We conducted this CM study because there is limited understanding of the sector’s capacity needs for the science and practice of implementation, essential for understanding what works, for whom, and why. Previous studies on the delivery, adoption, and sustainment of complex suicide prevention interventions identified stakeholder capacity at multiple levels as a critical determinant of implementation success [9, 44]. Capacity-related challenges were evident within organizations, intervention settings, and the broader community, often linked to insufficient resources. These challenges included limited ability to support interventions, apply evidence-based implementation strategies, generate and use practice-based evidence, integrate implementation science into research, and establish dedicated evaluation or training teams.

Over the past two decades, “capacity building” has become widely used in the development sector. It is quite indiscriminately and sometimes interchangeably used with training, which has diluted its meaning to an extent [45]. In the early 2000s, capacity building was described as strengthening the skills, abilities, and systems needed for organizations and communities to function effectively and adapt over time [46]. More recent work has attempted to clarify what constitutes capacity and how it can be measured. For example, Nordin et al [41] provide a broader definition of capacity as including infrastructural capacity (resources and systems), management capacity (planning and performance management), and dynamic capabilities (adapting to changing contexts). Capacity has also been conceptualized across system levels [47]. At the individual level, capacity could mean developing skills and competencies achieved through professional development, training programs, and educational opportunities. At the organizational level, it refers to institutional systems, structures and processes [48], which are emboldened through strategic planning, board development, program design and evaluation, and organizational culture change. At the system level, enabling the broader environment within which organizations function [49] could be achieved through building partnerships and collaborations, sector-wide practices and standards, and supportive policy environments. Our findings align with this broader, multi-level understanding of capacity. The needs identified extend beyond individual training to organizational and system-wide change, signaling the need to shift how implementation science is understood and utilized in the sector.

Brownson et al [39] describe capacity for implementation science as an essential “ingredient,” involving both a “push” and “pull.” The “push” refers to training researchers to generate evidence that meets stakeholder needs and supports adoption. The “pull” refers to strengthening the capacity of implementers, clinicians, and practice-based researchers to use this evidence. Historically, greater emphasis has been placed on the push rather than the pull. Advancing suicide prevention requires strengthening both, particularly given that examining adoption, delivery, and sustainment of interventions has received less attention than outcomes.

Although many implementation science capacity-building programs exist globally, there is a focus on research and academic audiences. Common gaps include emphasis on broad competencies, limited attention to understanding the craft of implementation, and restricted accessibility due to awareness, eligibility criteria, or cost [15]. As Beidas et al [7] argue, not all stakeholders need specialist expertise; rather, different levels of skills and experience are required. This has also been identified as a priority area for action [24]. Because this study adopted a broad, multi-stakeholder lens, its focus extended beyond implementation research capacity alone. Accordingly, any capacity-building initiatives informed by these findings must operate across multiple levels and move beyond individual training and education programs. A key strength of this study is its lived experience and practitioner representation, which underscores the need for more inclusive and diverse capacity-building efforts that are not targeted primarily for researchers. There is also scope to learn from existing capacity-building initiatives within the suicide prevention sector, particularly their implementation strategies [50]. Consistent with this, findings from a sector-wide translational needs survey emphasize the development of translational tools that draw on diverse knowledge-sharing approaches. These include educational sessions, webinars, eNewsletters, and website content, while also integrating lived experience knowledge and expertise [51].

CM is a novel approach within suicide research. It overcomes three limitations of a Delphi study, a more commonly used approach for developing and prioritizing research guidelines related to suicide [52]. First, it allows participants to generate statements rather than respond to a predefined list of statements. Therefore, the methodology offers space for participant experiences to be included in the study. Second, it captures complex phenomena (e.g. capacity building needs) without relying on central tendencies. Finally, the need for consensus in a Delphi study may force respondents to modify their answers, whereas consensus is not needed in CM.

Limitations

Some limitations need to be acknowledged. First, the use of a convenience sample may limit generalizability, although diverse stakeholder participation mitigated this to some extent [53]. Second, stakeholder perspectives were analyzed collectively. Subgroup analyses may have yielded different priorities, but sample sizes within stakeholder categories were small. Third, implementation science remains relatively nascent in suicide prevention, and participants may have drawn on their own interpretations of capacity building needs despite standardized definitions. Fourthly, the self-report design may have limited insights into how participants interpreted and organized statements across clusters. Finally, our findings should be interpreted with caution, as cluster ratings were grouped closely, suggesting limited differentiation between clusters. Rather than indicating strong distinctions, they suggest that each strategy has a role, while the importance and feasibility ratings help guide more strategic sector initiatives.

Conclusion

This CM study offers an important snapshot of the sector’s current needs around effectively utilizing implementation science within suicide prevention. The findings suggest that capacity building must extend beyond traditional training and individual skill development to encompass organizational and system-level change. Strengthening implementation science in the sector will require coordinated efforts that address structural conditions, stakeholder engagement, and the integration of lived experience alongside research expertise. Further exploration of how different stakeholder groups conceptualize and prioritize capacity building will be critical to designing initiatives that are responsive, inclusive, and capable of driving meaningful and sustained change.

Funding sources

This study was conducted with support from the LIFEWAYS project, which is funded by the Australian Government Department of Health, Disability and Ageing under the National Suicide Prevention Leadership and Support Program (grant/award number: not applicable).

Role of the funder

The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.

Conflicts of interest

Document attached.

Supplementary material

Supplementary material is available at Translational Behavioral Medicine online.

Human rights

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The study involved human participants and was approved by Griffith University Human Research Ethics Committee—GU ref no: 2025/764.

Informed consent

Informed consent was obtained from all individual participants included in the study.

Welfare of animals

This article does not contain any studies with animals performed by any of the authors.

Transparency statements

Study registration

This study was not formally registered.

Analytic plan pre-registration

The analysis plan was not formally pre-registered. De-identified data from this study are not available in a public archive. De-identified data from this study will be made available (as allowable according to institutional IRB standards) by emailing the corresponding author.

Analytic code availability

There is no analytic code associated with this study.

Materials availability

Materials used to conduct the study are not publicly available.

Use of AI

Artificial intelligence was not used in the development of this manuscript.