The site initiation visit is the most ceremonial event in clinical trial operations. A study lead from the CRO travels to the site, sits in a conference room with the PI and study coordinators, walks them through a slide deck of the protocol, the schedule of events, the case report form, the regulatory binder, the IP handling SOP, and the safety reporting workflow. It runs two hours, sometimes longer. Then everyone signs the SIV log and the site is, in a procedural sense, ready to enroll.
The convention treats SIV duration as a measure of thoroughness. A long SIV is a careful SIV. A short SIV is a rushed one. We have come to believe the opposite. Across more than 300 IVD studies since 2011, the sites whose SIVs ran the longest were rarely the ones that enrolled fastest. The sites whose SIVs ran 20 to 30 minutes — because by the time the visit happened, the site already understood the protocol — were the ones that produced subjects in the first weeks of activation.
This is not an argument for cutting corners on the SIV. It is an argument for moving the work the SIV pretends to do into the weeks before the SIV, where the site has time to absorb the material and the CRO has time to answer questions before they become problems on day one of enrollment. The SIV is the visible tip of a support system that has either done its job already or is about to make up for lost time in front of the PI. The framework below describes the system that makes 20-minute SIVs possible.
Move the work upstream of the SIV.
By the day of the visit, the site should already understand the protocol. The SIV confirms understanding; it does not introduce it.
The first principle is the simplest and the most consequential. A two-hour SIV happens because the site is being introduced to material it has not previously seen. The CRO presents the protocol, the schedule of events, the consent script, the IP handling, the AE reporting flow — and the site absorbs it in the room. Some of it sticks. Most of it does not. The site coordinator who sat through the SIV on Monday will, on Tuesday's first consent appointment, have to look up the inclusion criteria.
The fix is to send the material in advance and create the conversational space for it to land before the SIV happens. A protocol overview that arrives two weeks before the SIV with a structured set of follow-up touchpoints — a 30-minute call with the coordinator, a separate 20-minute call with the PI, a written Q&A — does the work that the SIV otherwise tries to do in two hours under fluorescent lighting. The PI walks into the SIV with questions already answered. The coordinator walks in confident about the consent flow. The visit becomes a confirmation, not a training session.
On a recent multi-cohort rescue study, nine site initiation visits cleared in January 2025 with a median turnaround of four days from IRB approval to SIV completion. The SIVs themselves were short — most under 30 minutes — because the protocol had been pre-shared in early January and the substantive PI conversations had happened by phone in the days before the visit. The fastest SIV cleared in a single day. The conventional benchmarks for SIV completion timelines assume the SIV is the first substantive site engagement. It does not have to be.
Schedule a 30-minute pre-SIV call with the coordinator and a separate 20-minute call with the PI no later than ten days before the SIV. Send the protocol synopsis, the consent script, and the schedule of events in advance. The SIV agenda should be: confirm IP receipt, confirm freezer logistics, sign the SIV log. If anything else takes longer than five minutes, the upstream system failed.
Site support is on-call, not on-calendar.
The questions that matter on a study come up at consent, at the draw, at shipment — not at the SIV. The CRO has to be available when those moments happen.
The second principle generalizes the first. If the SIV is not where the substantive site work happens, where does it happen? The honest answer is: at every consent, every draw, every shipment, every monitoring visit, every protocol amendment. The site has questions when the questions arise — not on the calendar the CRO has scheduled.
Most CROs structure site support as a series of milestone events: SIV, first monitoring visit, interim monitoring, close-out. The model assumes that questions can wait until the next scheduled touchpoint. They cannot. The coordinator who is sitting in front of a patient who does not fit the inclusion criteria as written needs an answer in the next five minutes, not at the next interim visit. The PI who is on hold with the central lab over a sample receipt discrepancy needs the CRO to call the lab, not to schedule a follow-up.
This requires staffing the support system to be reachable rather than scheduled. A named project manager whose phone the site can call. A study lead who answers email within hours rather than business days. A clinical operations team that escalates issues without waiting for the next monitoring visit. None of this is glamorous. All of it is what site support actually means.
The pattern is observable across our rescue engagements. Sites we inherit on stalled studies often report that the prior CRO was reachable for monitoring visits but unreachable between them. The questions that go unanswered between visits become protocol deviations at the next one. The same sites, once moved onto an on-call support model, frequently produce more subjects in the first month of new support than they had in the prior six.
Each active site should have a named first-call contact whose phone number the coordinator has saved. Response time targets should be measured in hours, not business days. The SIV is the moment to confirm those expectations explicitly with the PI, not to set them implicitly through scheduling cadence.
SIV duration is a quality signal. Track it.
Sites whose SIVs run long enroll slowly. The signal shows up on day one. Use it.
The third principle is operational. The duration of the SIV — once principles I and II are in place — becomes a leading indicator of how well the site is going to perform. SIVs that run long are usually long for a reason: the PI did not engage in advance, the coordinator did not absorb the materials, the upstream support system did not reach the site. All three are early signals of a site that will struggle to enroll under contract.
This is not deterministic. There are good sites whose SIVs run long because the PI has substantive scientific questions about the protocol, and those are exactly the SIVs you want to take time over. The signal worth tracking is not duration in absolute terms but duration relative to expectation: a site that was expected to be ready and whose SIV nonetheless required two hours to complete the standard agenda is signaling something. Treat the signal as data.
The corollary is the operational discipline this requires from the CRO. Most monitoring systems track the SIV as a binary event — completed or not. SIV duration, captured systematically across a study's full site list, becomes a leading indicator of which sites will need additional support in the first month of enrollment, well before the enrollment data itself confirms the diagnosis.
Capture SIV duration as a tracked metric for every site, alongside enrollment performance. Sites whose SIVs ran materially longer than expected should trigger a 14-day check-in independent of the standard monitoring schedule. The duration signal predicts the recovery work; act on it before the enrollment data confirms it.
What this framework rules out.
The three principles describe what good site support actually looks like. They also rule out a few conventions worth naming.
They rule out the long SIV as a sign of thoroughness. A two-hour SIV is most often a sign that the upstream support system has compressed two weeks of pre-SIV work into a single conference room. Thorough support shows up before the visit, not during it.
They rule out milestone-based site support cadences. Sites do not have questions on a schedule. Support has to be available when the questions arise, which is between the milestones, not at them.
They rule out treating the SIV as the moment a site becomes ready. Readiness is a state the site enters in the weeks before the SIV, through pre-shared materials, structured conversations, and confirmed logistics. The SIV ratifies readiness; it does not produce it.
The framework is not closed. When the study outcome matters, you call RDI. Good site support is a system that runs continuously, before activation and after database lock. The 20-minute SIV is the visible signal that the system is working.