One recruitment workflow and one enrollment funnel across your sites

Why is enrollment hard to predict across a CRO's site portfolio?

IQVIA estimates that biopharma spent about $159 billion worldwide on clinical development in 2025 and outsourced about $75 billion of it.^⁠[1] Enrollment risk comes with that work, spread across sites that perform very differently.

Tufts CSDD found that 89% of trials eventually meet their enrollment goals, but getting there typically means nearly doubling the original timeline.^⁠[2] The data are global and predate 2013, so treat them as a pattern, not a current benchmark.

Vendor count matters too. In a Tufts CSDD benchmarking study of 13 sponsors and 133 trials, announced by Syneos Health on September 16, 2026, no sourcing model (in-house, full-service CRO or FSP) consistently outperformed the others, and more service providers were associated with longer timelines from protocol approval to database lock.^⁠[4]

What can Bond add to feasibility and site selection?

Site feasibility often means a questionnaire and a qualification visit, and sites are typically not compensated for the work. A 2024 industry task-force paper puts the industry-wide scale of the feasibility problem at $1.6 billion, extrapolated from an ASCO survey of oncology sites. It adds that not all sites have good data and tools for estimating enrollment, leaving sponsors and CROs with inaccurate timelines.^⁠[5]

Where a site already runs Bond, Identify screens structured and unstructured records against the protocol's criteria and shows the chart evidence for each one.^⁠[12] Bond runs feasibility from the same EHR screening, giving eligible-patient counts for a protocol backed by the chart evidence behind each match, which a coordinator can check.^⁠[14] Diagnosis codes alone express few criteria: in Bond's preprint, across 51,055 eligibility criteria from ClinicalTrials.gov, ICD-10-CM resolved at least one exact multi-token concept in 5.7% of criteria, against 82.0% for Bond's terminology graph.^⁠[13]

• Same criteria at every site. When each site screens against the same configured criteria, count differences are less likely to reflect method.
• Know the limits. Counts exist only where Bond is connected to the EHR. Criteria outside the chart, such as willingness to take part, still need a conversation.
• Use them in selection. Pair the counts with the feasibility questionnaire template and the guide to how sponsors choose sites.

What changes day to day at each site?

Bond runs three stages in one workflow. The site team keeps every decision that needs a person.^⁠[12]

1. 1

   Identify

   EHR screening ranks candidates, with the chart evidence for each criterion. Coordinators review a ranked list instead of building one.

2. 2

   Engage

   Engage runs voice and text outreach, pre-screening and scheduling, with scripts set per site and study. Patients are told AI assistance is used and can reach a person at any time: the agent transfers the call live to a coordinator or books a human callback, whichever the site prefers.^⁠[12],[14]

3. 3

   Consent

   Consent support gives plain-language explanations and answers patient questions before the visit, with escalation to staff. The site and PI still obtain consent.

Bond can also set up Meta and Google ad campaigns for a study across the sites it supports.^⁠[14] Leads from those ads flow straight into the same voice and text agents that pre-screen and schedule patients found in each site's EHR.^⁠[14]

After enrollment, the same voice and SMS agents send visit reminders, book transportation, collect symptoms and diaries, run side-effect check-ins and flag participants at risk of dropping out.^⁠[14]

Which metrics will you see, and are they the same at every site?

For every site and study, Bond reports the same funnel: patients matched, contacted, pre-screened, consented and randomized. It adds time to enrollment, matching accuracy, screen-failure signals and coordinator hours saved, and keeps an audit trail. A patient counts as enrolled when randomized.^⁠[12]

Can Bond help rescue a study that is behind plan?

We found no reliable published count of rescue studies, but registry data shows how often enrollment falls short. Of 2,579 Phase 2 and Phase 3 trials on ClinicalTrials.gov that closed in 2011, 19% were terminated for failed accrual or finished below 85% of target enrollment.^⁠[6]

A rescue has three levers: add sites, amend the protocol, or get more from active sites. New sites restart selection and start-up, which industry still describes as more than two months to select sites and another eight months to activate them.^⁠[7] New sites do not guarantee patients either. In the STICH heart-failure trial, low enrollment forced an expansion from 32 planned US and Canadian sites to 171 sites; 44 approved sites never enrolled a patient, at an estimated $10,000 per deactivated site.^⁠[8]

Bond works on the third lever. With the EHR connected, it screens charts at active sites and contacts the candidates it finds. Full EHR integration takes 48 hours per site. A pilot without EHR integration does not read charts: it contacts and pre-screens patients from a list the site already holds.^⁠[12]

What does implementation ask of the CRO and its sites?

Start-up is already the slow part. In WCG's 2025 survey of 611 sites, 54% of independent sites and physician practices said they can start a study in under 60 days, against 9% of academic medical centers, community hospitals, health systems and site networks.^⁠[9] In WCG's 2024 survey, budgets and contracts were the most common cause of start-up delay, affecting 69% of sites.^⁠[10]

Each site signs a BAA, completes a security review, connects its EHR, validates the criteria with its coordinators and, where required, sends outreach scripts to the IRB. The 48 hours per site depend on the EHR, the IT review and the interface method.^⁠[12] The implementation plan lists each step.

• Choose the first studies and sites where enrollment risk is highest and IT can respond quickly.
• Start Bond setup during site start-up, next to contracts and budgets, so a validated match list is ready when the site activates.
• Agree funnel definitions and reporting cadence with the sponsor, including enrolled as randomized.
• Name one owner for the dashboard and for site escalations.

How does pricing line up with how a CRO is measured?

Bond uses two-part pricing. A volume-based platform fee covers EHR integration, implementation and ongoing operation: the EHR connection, security review, audit logging setup, workflow configuration and ongoing support. There is no separate integration fee. A success fee is paid per enrolled patient, where enrolled means randomized, and per-visit milestones can be added.^⁠[12]

The success fee follows randomized patients, the number you report to the sponsor, rather than outreach volume. Pricing is custom. See pricing for the structure and what every engagement includes.