# The Multi-Institutional Research Kickoff Meeting: An Agenda Template By the end of this guide you will be able to draft a complete multi-site research kickoff meeting agenda, budget its time with a simple formula, assign an owner and a concrete output to every item, and close the meeting with dated actions that each site accepts. You will also have a reusable template, a filled 90-minute example, and a checklist of the governance items that cannot be pushed to a later call. Before you start, have the following on hand: - The funded proposal or the current protocol draft - The list of participating sites and each site's principal investigator - A shared document tool the whole consortium can edit, such as Google Docs or Microsoft 365 - A video meeting platform your sites already use, such as Zoom or Microsoft Teams - Your funder's requirements for reporting, ethics review, and data sharing Multi-site work is now a default mode of scientific collaboration. Methodologists describe the multi-site single-study design as one of three main approaches for questions a single institution cannot answer, alongside the single-site single-study design and spatial meta-analysis [1]. A multi-site design only produces pooled knowledge if every site runs the same protocol in the same way, and that alignment has to start somewhere. It starts at the kickoff. ## What the Multi-Institutional Research Kickoff Meeting Must Accomplish The multi-institutional research kickoff meeting is the first structured working meeting of every participating site after the award is made. It is not a launch party and it is not a press event. Its job is to convert a funded proposal into dated, named commitments that survive contact with real calendars. The need for this meeting is well documented. The challenges of collaborative, multi-site research, including coordination burden, uneven capacity across sites, and communication breakdowns, have been recognized in the literature for years [2]. Teams working with academic and community-based partners add a further principle they summarize as "go slow to go fast": time invested in relationships and explicit expectations early in a patient-centered multi-site project prevents far larger delays later [3]. A kickoff that produces nothing written down has failed, so design the meeting around four outputs: 1. **A shared scientific understanding.** Every attendee can state the hypothesis, the primary outcome, and what their site contributes. 2. **Named roles.** A responsibility matrix that says which site enrolls, which site assays, which center houses the data, and who signs which agreement. 3. **An ethics and data governance path.** A decision on the review pathway and named owners for the data use agreement and related documents. 4. **A dated timeline and communication cadence.** Milestones, the recurring meeting schedule, and the platforms the consortium will use. Some programs even publish their kickoff agendas, which shows how central this document is to a consortium's working identity. The combined ENSO workshop and ARiSE kickoff meeting agenda from 2019 is one such example, pairing the formal kickoff with a scientific workshop [4]. You do not need to go that far, but you should treat the agenda as a versioned project document, not a calendar afterthought. ## Before the Meeting: Invitations, Pre-Reads, and Scheduling ### Map every actor first Before you send a single invitation, list every actor whose cooperation the study needs. The STRIPE project, which studied the Global Polio Eradication Initiative, began by theorizing the full universe of polio-related actors in each context, then enumerating the specific individuals within that universe, and only then collecting data from them [5]. Apply the same discipline to your consortium. Your actor universe usually includes site principal investigators, study coordinators, laboratory managers, data managers, a lead biostatistician, regulatory staff at each institution, and, where the design includes them, patient or community partners. Anyone missing from this list becomes a surprise bottleneck at month three. ### Keep the invite list to decision makers Invite the people who can commit their site. The core list is one principal investigator per site, the project manager, the regulatory lead, the data core lead, the lead statistician, and one coordinator or trainee representative per site. Where end users shape the intervention, involve them from the start rather than after decisions harden. In one multi-site formative research project, young people participated directly in developing a self-care intervention across the study sites [6]. A kickoff that includes community partners looks different from one that does not, and the engagement literature suggests that difference deserves explicit planning [3]. ### Send a pre-read packet The highest-yield preparation is a pre-read packet sent at least five business days before the meeting. Include the current protocol draft, the draft site responsibility matrix, the draft data flow diagram, and the agenda itself. If attendees read, the meeting becomes a discussion of decisions. If they do not, the meeting becomes a read-aloud, and read-alouds are why people dread consortium calls. ### Schedule around the sites you actually have Work the time zones on paper before you send the poll. A coordinating center in Houston with sites in Boston, Denver, and Seattle gives this map: ``` Coordinating center: Houston (Central Time) Proposed start: 9:00 AM CT Boston site: 10:00 AM ET Denver site: 8:00 AM MT Seattle site: 7:00 AM PT ``` 9:00 AM Central is the latest start that keeps Seattle out of the pre-7 AM zone. If the scheduling poll returns no hour that works for everyone, split the meeting into two sessions by workstream and record both. State the time zone on every invitation, because a bare "9:00" in a consortium email is an invitation to miss the meeting. ## The Multi-Site Research Kickoff Meeting Agenda Template Use the template below as the starting point for your own agenda. The default format is a 90-minute video meeting for up to about six sites. Larger consortia need a half day, a full day, or separate sessions per workstream. ### Budget the time with a formula Reserve 10 minutes for the welcome and the close, then split the remaining working time 60/30/10 across presentations, discussion, and buffer: ``` Meeting length: 90 minutes Reserve for welcome and close: -10 minutes Working time: 80 minutes Working time split 60 / 30 / 10: Presentations: 0.60 x 80 = 48 minutes Discussion: 0.30 x 80 = 24 minutes Buffer: 0.10 x 80 = 8 minutes ``` In practice you fold discussion into each item rather than holding one discussion block, but the ratio keeps the meeting from collapsing into a seminar. If your presentation items total more than 60 percent of working time, cut content or lengthen the meeting. ### The template ```markdown # Kickoff Meeting: [Project short name] Date: [YYYY-MM-DD] Time: [HH:MM to HH:MM] [time zone] Chair: [Coordinating center PI] Note taker: [Project manager] ## Objectives (3 to 5, each testable) 1. Confirm site roles and enrollment or sample targets 2. Agree the IRB pathway and data governance owners 3. Approve the training and certification plan 4. Approve the milestone timeline and communication cadence ## Pre-reads (sent at least 5 business days ahead) 1. Protocol draft, version [number] 2. Site responsibility matrix, version [number] 3. Data flow diagram, version [number] ## Decisions needed today 1. [Decision, with the options written out] 2. [Decision, with the options written out] ## Agenda 1. Welcome and objectives (5 min, Chair) 2. Introductions and site roles (10 min, all) 3. Scientific overview and specific aims (15 min, Lead PI) 4. Site workflows and targets (10 min, Site PIs) 5. Ethics, IRB, and data protection (15 min, Regulatory lead) 6. Standardization, training, and certification (10 min, Training lead) 7. Timeline, milestones, and checkpoints (10 min, Project manager) 8. Communication plan and collaboration tools (5 min, Project manager) 9. Risks and open issues (5 min, all) 10. Actions, owners, and next meeting (5 min, Chair) ## Parking lot - [Item, who raised it, where it goes next] ``` ### The same agenda as a table with outputs Every item needs a lead and an output. An item with no output is a conversation, and conversations drift. | # | Agenda item | Minutes | Led by | Required output | |--|-------|-----|----|---------| | 1 | Welcome and objectives | 5 | Chair | Objectives confirmed | | 2 | Introductions and site roles | 10 | All | Updated responsibility matrix | | 3 | Scientific overview and aims | 15 | Lead PI | Shared statement of hypothesis and primary outcome | | 4 | Site workflows and targets | 10 | Site PIs | Per-site targets and workflow owners | | 5 | Ethics, IRB, and data protection | 15 | Regulatory lead | IRB pathway decision and agreement owners | | 6 | Standardization and training | 10 | Training lead | Training plan with certification dates | | 7 | Timeline and milestones | 10 | Project manager | Approved milestone list | | 8 | Communication and tools | 5 | Project manager | Meeting cadence and platform choices | | 9 | Risks and open issues | 5 | All | Parking lot list | | 10 | Actions and close | 5 | Chair | Action list with owners and due dates | The minutes column sums to 90. Run the same check on your own version before sending it, because an agenda that sums to 110 minutes is a promise you cannot keep. Three checks make an agenda ready to send: the minutes sum to the meeting length, every item names an owner and an output, and every decision lists its options in writing. ## A Worked Multi-Site Research Kickoff Meeting Agenda Example Here is the template filled in for a hypothetical supportive care trial at the scale of published multi-site trials. As a size reference, the PRISM-AC trial is a two-arm, multi-site randomized controlled trial planning to enroll 144 adolescents and young adults with advanced cancer across four academic medical centers [7]. A study of that shape has a four-site kickoff that looks like this: ```markdown # Kickoff Meeting: Multi-Site Supportive Care Trial (worked example) Date: 2026-03-04 Time: 9:00 to 10:30 AM Central Time Chair: Coordinating center PI Note taker: Project manager ## Objectives 1. Confirm per-site enrollment targets summing to the protocol target 2. Agree the single-IRB pathway and the data use agreement owner 3. Approve the intervention training and certification plan 4. Approve the 18-month milestone timeline and monthly call cadence ## Decisions needed today 1. IRB pathway: single IRB of record vs local review at each site 2. Data use agreement owner: data core vs each site's counsel ## Agenda 1. Welcome and objectives (5 min, Chair) 2. Introductions, 2 min per site (10 min, all) 3. Scientific overview: design, arms, primary outcome (15 min, Lead PI) 4. Site workflows: recruitment, delivery, app support (10 min, Site PIs) 5. Ethics, IRB, and data protection (15 min, Regulatory lead) 6. Training: session manuals, certification, calibration calls (10 min, Training lead) 7. Timeline: startup, enrollment, follow-up, closeout (10 min, Project manager) 8. Communication: monthly all-sites call, shared folder, tracker (5 min, Project manager) 9. Risks and open issues (5 min, all) 10. Actions and next meeting (5 min, Chair) ``` Two details in this example deserve comment. First, the introductions math: 10 minutes across 4 sites is 2.5 minutes per site, so ask each site for a 2-minute summary and keep the remainder as slack. Second, item 6 exists because standardization across sites never happens by accident. In diagnostic assay work, one published three-site study shipped five replicate sets of specimens to every site and had them scored blinded to site, specifically to measure inter-site and intra-site reproducibility [8]. Plan the equivalent exercise for whatever your study runs at every site, whether that is an assay, a structured interview, or a manualized intervention. ## Step by Step: Building and Running the Meeting ### Step 1. Write three to five testable objectives Each objective must be a sentence you could verify at the end of the meeting. "Discuss data sharing" is not testable. "Agree the data use agreement owner and the storage location" is. If you cannot write testable objectives, the consortium is not ready for a kickoff, and you should hold a smaller planning call first. ### Step 2. Budget the time Apply the 60/30/10 formula from the template section. Then sanity check the result: the item minutes must sum to the meeting length, and no single item should exceed 20 minutes without a break or a written pre-read behind it. ### Step 3. Assign an owner and an output to every item An owner is a named person, not an institution. "The data core will present" fails when the data core has four people. Write the required output next to each item, as the table above shows. ### Step 4. Send the pre-read and the agenda five business days out Use a message like this: ``` Subject: Pre-read for the [Project] kickoff, [Date] Attached: 1. Protocol draft v0.9 (please read Sections 3 and 5) 2. Site responsibility matrix v0.5 (send corrections by [Date]) 3. Data flow diagram v0.5 4. Agenda v1.0 Two decisions need votes on the call: the IRB pathway and the data use agreement owner. Please come ready to decide. ``` ### Step 5. Facilitate the decisions, not the discussion For each decision item, state the options, take one round of comments, then call the question. For items where many voices must be heard, borrow structure from consensus methods. The co-production of a medicines optimisation research agenda used a modified Nominal Group Technique to generate preliminary questions, electronic consultation and ranking with a wider multi-stakeholder group, and a final one-day consensus meeting that ranked 92 questions into four theme areas [9]. A kickoff rarely needs the full machinery, but silent written generation followed by a round-robin read-out gets quieter sites into the conversation within minutes. Keep a parking lot for anything that is real but not decidable today, and say aloud where each parked item goes next. ### Step 6. Close with actions and send minutes within 48 hours End the meeting by reading the action list back. Then send minutes within two business days, while memories are fresh: ```markdown ## Actions from kickoff, 2026-03-04 1. Submit single-IRB reliance packet. Owner: Regulatory lead. Due: 2026-03-18 2. Draft data use agreement. Owner: Data core lead. Due: 2026-03-25 3. Circulate training and certification schedule. Owner: Training lead. Due: 2026-03-11 4. Open shared folder and project tracker. Owner: Project manager. Done in meeting 5. Schedule first monthly all-sites call. Owner: Project manager. Due: 2026-03-08 ``` The recurring meetings you schedule here are not bureaucracy. In a cancer genetics improvement program disseminated from a lead team to five participating health systems, the facilitation-associated meetings and meeting resources were the most frequently used supports, and the majority of the 21 surveyed site team members rated them useful or helpful [10]. The kickoff sets that rhythm. ## Ethics, Data Protection, and Governance Items You Cannot Skip Three agenda items have long lead times and belong at the kickoff, not after it. ### The IRB pathway Decide, in principle, whether the study will use a single institutional review board of record or local review at each site. The National Cancer Institute operates a Central Institutional Review Board for multi-site research precisely to reduce duplicate review across sites [11]. Networks have also expanded the functional scope of their review over time, moving from data-only studies to non-clinical interventions and then to clinical trials, which means the review burden a consortium inherits can grow with the study's ambition [12]. Your kickoff output here is a decision in principle, a named owner for the reliance paperwork, and a date by which each site's local requirements are confirmed. ### Data protection as its own workstream Protecting human subjects and protecting their data are related but distinct obligations in multi-site research [13]. The kickoff should name who owns the data use agreement, where data will physically reside, which sites can access which variables, and how de-identification happens before transfer. Put these as line items under the ethics block of the agenda rather than hoping they emerge from open discussion. ### Decide now versus decide later | Decide at kickoff | Defer to a named workstream | |--|--| | IRB pathway in principle | Protocol amendments | | Owners for each agreement | Full data analysis plan | | Data storage location and access tiers | Variable-level data dictionaries | | Training and certification plan | Publication and authorship policy details | | Milestone dates and go/no-go checkpoints | Site-specific recruitment tactics | Deferring is not avoiding. Each deferred item needs a named workstream, a named lead, and a date by which it returns to the full consortium. ## Research Collaboration Tools and the Cadence That Follows The kickoff is one meeting, but scientific collaboration runs on a rhythm. Set that rhythm before the meeting ends. **Training and calibration.** If sites administer interviews, score behaviors, or deliver a manualized intervention, build the training plan into the agenda. One multi-site childhood trauma research network used synchronous and asynchronous didactic modules, certified raters through critique of videotaped mock scale administrations, held monthly review meetings, and issued ongoing scoring exercises, then estimated inter-rater reliability with kappa, weighted kappa, and intraclass correlations across its key measures [14]. That is the level of detail your training agenda item should produce. **A common phase structure.** Define the study's phases identically at every site, with dates. In one multi-site study spanning seven Veterans Health Administration psychiatric units, every site followed the same 21-week structure, with a three-week pre-treatment period, a fifteen-week treatment period, and a three-week post-treatment period [15]. A shared phase calendar prevents the slow drift in which one site starts enrollment three weeks after the others and nobody notices for a quarter. **The tool stack.** Keep it boring and universal: a shared drive with version control for documents, the video platform your sites already have, a project tracker with owners and due dates, and your electronic data capture system. Add asynchronous tools deliberately. The medicines optimisation agenda project ran an electronic consultation and ranking stage between its in-person events [9], and multi-actor agenda-setting projects commonly use Delphi-style survey rounds to converge. In the UK agroforestry agenda, 156 stakeholders submitted 238 candidate priorities through a survey and workshop, and a second workshop with 48 participants distilled them to 40 [16]. If your consortium will set its own priorities over time, plan those survey tools at kickoff. **Evaluate your own process.** The cancer genetics program evaluated its multi-site rollout with a logic model covering inputs, activities, outputs, fidelity, and reach [10]. A one-page version for your kickoff, tracking whether pre-reads went out on time and whether actions closed on their due dates, tells you within a quarter whether your cadence is real. ## Common Mistakes and Limitations Mistakes that recur in multi-site kickoffs, with fixes: 1. **Inviting entire site teams.** Forty attendees cannot make decisions. Invite decision makers and ask each site to cascade details afterward. 2. **No pre-read.** The meeting becomes a read-aloud. Send the packet five business days out and say in the invitation which sections need reading. 3. **Items without owners or outputs.** Fix this in the agenda itself, in writing, before anyone joins the call. 4. **Deferring ethics and data governance.** IRB reliance and data use agreements have lead times measured in months. Deciding them "later" delays study startup. 5. **Ignoring time zones.** A start time reasonable at the coordinating center can be a 6 AM start for a West Coast site. Compute local times before sending the poll and rotate inconvenient slots for recurring calls. 6. **Treating the kickoff as a one-off.** The kickoff's most important output is the cadence that follows, not the meeting itself. 7. **Assuming standardization is automatic.** Identical protocols do not guarantee identical performance, so schedule calibration and reproducibility checks now rather than after the first data mismatch. 8. **Not recording decisions.** If it is not in the minutes with an owner and a date, it did not happen. Limitations. This template is a starting point, not a regulation. Consortium size, funder requirements, and whether the study is a clinical trial, an observational cohort, or a multi-lab bench project all change the agenda. Regulatory decisions described here need confirmation from your own IRB and institutional counsel. And no agenda can rescue a study whose scientific design is still moving, so if the aims are unsettled, hold a smaller design meeting first and schedule the kickoff after the protocol stabilizes. ## Frequently Asked Questions ### What is a multi-site research kickoff meeting agenda? It is the timed, ordered list of items a multi-site consortium covers in its first all-sites working meeting, with a named owner and a required output for every item. The agenda turns a funded proposal into dated commitments. ### How long should a multi-site research kickoff meeting last? Plan 90 minutes for up to about six sites. Larger consortia need a half day, a full day, or separate sessions per workstream, and any session longer than two hours needs a break. ### Who should be invited to the kickoff? Invite one principal investigator per site, the project manager, the regulatory lead, the data core lead, the lead statistician, and any patient or community partners the design includes. Everyone else can be reached through a site-level cascade after the meeting. ### What goes in the pre-read packet? The current protocol draft, the site responsibility matrix, the data flow diagram, and the agenda itself, sent at least five business days before the meeting with a note on which sections need reading. ### How do we handle time zones across sites? Compute each site's local start time before you send the poll, avoid starts before 7 AM or after 8 PM local, state the time zone on every invitation, and rotate inconvenient slots for recurring meetings. ### Is a kickoff meeting the same as an investigator meeting? No. The kickoff is the first all-sites working meeting that converts the proposal into commitments, while investigator meetings in clinical trials recur throughout the study to train and update site staff. ### When should the kickoff happen? Within the first month after the award starts, once the site principal investigators are confirmed, and before IRB submissions and data use agreements are finalized. ### What should be decided at kickoff versus later? Decide roles, the IRB pathway in principle, data governance owners, the training plan, and the milestone timeline at kickoff. Defer protocol amendments, the full analysis plan, and authorship policy to named workstreams with return dates. ## Sources 1. [A Research Agenda for Understanding How Social Inequality is Linked to Brain Structure and Function](https://pubmed.ncbi.nlm.nih.gov/38172629/) 2. [Recognising the challenges of collaborative, multi-site research](https://doi.org/10.1177/136140960400900606) 3. [Go Slow to Go Fast: Successful Engagement Strategies for Patient-Centered, Multi-Site Research, Involving Academic and Community-Based Organizations](https://pubmed.ncbi.nlm.nih.gov/30353249/) 4. [ENSO workshop and ARiSE kickoff meeting agenda](https://www.semanticscholar.org/paper/e7ff10c60dfaebc740d9819701c32a0611ab22ee) 5. [Sampling method for surveying complex and multi-institutional partnerships: lessons from the Global Polio Eradication Initiative.](https://pubmed.ncbi.nlm.nih.gov/32787958/) 6. [Young people's participation in the development of a self-care intervention--a multi-site formative research study.](https://pubmed.ncbi.nlm.nih.gov/23221590/) 7. [Protocol for The Promoting Resilience in Stress Management (PRISM) Intervention: a multi-site randomized controlled trial for adolescents and young adults with advanced cancer](https://pubmed.ncbi.nlm.nih.gov/37066150/) 8. [Abstract 684: Multi-site external reproducibility of PD-L1 IHC 22C3 pharmDx performance on Dako Omnis and comparison to Autostainer Link 48 for head and neck squamous cell carcinoma using the Combined Positive Score](https://doi.org/10.1158/1538-7445.am2025-684) 9. [A multi-stakeholder approach to the co-production of the research agenda for medicines optimisation](https://pubmed.ncbi.nlm.nih.gov/33441135/) 10. [Implementation and outcome evaluations of a multi-site improvement program in cancer genetics](https://pubmed.ncbi.nlm.nih.gov/36117454/) 11. [Meeting the Challenge: The National Cancer Institute's Central Institutional Review Board for Multi-Site Research.](https://pubmed.ncbi.nlm.nih.gov/29384720/) 12. [PS1-03: Expanding the Functional Scope of Institutional Review Board Review of Multi-Site Research in the HMO Research Network: from Data-Only Studies, to Non-Clinical Interventions, to Clinical Trials](https://doi.org/10.3121/cmr.2011.1020.ps1-03) 13. [Protecting Human Subjects and Their Data in Multi-Site Research](https://www.semanticscholar.org/paper/d3ebece71c3a2f192b05aeb021976dc6258e48b3) 14. [Establishing a training plan and estimating inter-rater reliability across the multi-site Texas childhood trauma research network.](https://pubmed.ncbi.nlm.nih.gov/36931015/) 15. [The Violence Prevention Community Meeting: A Multi-Site Study](https://pubmed.ncbi.nlm.nih.gov/27256945/) 16. [Co-designing a research agenda for UK agroforestry using a multi-actor approach](https://doi.org/10.1007/s13593-026-01089-8)