Vitalis Therapeutics Inc. — How sites are selected, activated, monitored and remediated across the VitaFlow (VTX-401) pivotal program: 260 planned sites in eight countries, a 18-week activation sequence running at 22, and a risk-based monitoring model under ICH E6(R2).
260
Sites planned
22 wk
Actual activation
34
Held in reserve
12 wk
Routine visit cadence
Contents
1. Why Sites Are the Program
A pivotal trial is not run by the sponsor. It is run by 260 independent medical practices, hospitals and research units, in eight countries, each with its own staff, patient population, competing commitments and institutional review board.
The sponsor writes the protocol. The sites decide whether it happens.
The consequence for program management. Every schedule commitment in this program
ultimately resolves to a question about sites: whether enough of them open, whether they find
eligible participants, and whether the data they produce will withstand inspection. The
Enrolment Dashboard traced the current
76-participant shortfall to site activation rather than to
screening. This document governs the thing that diagnosis points at.
2. Selection
Sites are scored against six weighted criteria before selection. The weighting is the interesting part — it encodes what the program believes actually predicts performance.
| Criterion | Weight | Why it is weighted that way |
|---|---|---|
| Access to eligible population | 30% | Documented patient volume meeting BMI and comorbidity criteria. The single strongest predictor of enrolment performance, and the one most often asserted rather than evidenced. |
| Prior performance in comparable trials | 25% | Randomizations per month achieved on previous metabolic or endocrine studies. Past performance predicts future performance better than any stated intention. |
| Investigator engagement | 15% | Whether the principal investigator is personally involved or has delegated the study to a coordinator without oversight. |
| Staffing and coordinator capacity | 15% | Dedicated study coordinator time. A site running eight concurrent trials on one coordinator will deprioritize whichever sponsor complains least. |
| Regulatory and quality history | 10% | Prior inspection findings, protocol deviation rates, data query rates. |
| Start-up speed | 5% | Historical time from contract to first patient in. Weighted low deliberately: a fast site that cannot enroll is worse than a slow site that can. |
Note what is weighted highest and what is weighted lowest. Access to an eligible
population carries 30% and start-up speed carries
5%. That ordering is deliberate and slightly counter-intuitive: a program
under schedule pressure is tempted to select for sites that open quickly.
A fast site that cannot enroll is worse than a slow site that can. It consumes start-up cost, occupies a slot in the network, generates monitoring overhead, and returns nothing. The 17 activated-but-zero sites in this program are that failure mode, and they were selected on criteria that under-weighted population access.
A fast site that cannot enroll is worse than a slow site that can. It consumes start-up cost, occupies a slot in the network, generates monitoring overhead, and returns nothing. The 17 activated-but-zero sites in this program are that failure mode, and they were selected on criteria that under-weighted population access.
Prior performance carries 25% because it is measurable and predictive. Feasibility questionnaires ask sites to estimate how many participants they can enroll; those estimates are consistently optimistic, across the industry, in a way that is well understood and still routinely relied upon. Historical randomization rates on comparable studies are worth more than any stated intention.
3. Activation
From selection to first screening visit is a sequence of dependent steps, most of which are outside the sponsor's direct control.
| When | Step | Owner |
|---|---|---|
| Week 0 | Feasibility questionnaire returned and assessed | Sponsor / CRO |
| Week 2 | Site selection visit — population, staffing and facilities verified in person | CRO |
| Week 4 | Contract and budget negotiation opens | CRO contracts |
| Week 8 | Regulatory document package collected (Form FDA 1572, CVs, licenses, financial disclosure) | Site / CRO |
| Week 9 | IRB submission | Site or central IRB |
| Week 14 | IRB approval received | IRB |
| Week 15 | Contract executed | CRO contracts |
| Week 16 | Site initiation visit — protocol training, systems access, delegation log | CRO |
| Week 17 | Investigational product shipped and received | Depot / site pharmacy |
| Week 18 | Green light issued — site may screen | Sponsor |
Planned 18 weeks; running at 22. That
4-week average slip, multiplied across the network, is the origin of the enrolment shortfall
reported at the status date. It did not come from screening, recruitment or protocol design. It
came from contracts and ethics submissions taking longer than the curve assumed.
The two steps that dominate are contract negotiation (weeks 4–15) and IRB approval (weeks 9–14). Neither is clinical work. Both are administrative processes running at other organizations' pace, and both are systematically under-estimated because they feel like they should be quick.
The two steps that dominate are contract negotiation (weeks 4–15) and IRB approval (weeks 9–14). Neither is clinical work. Both are administrative processes running at other organizations' pace, and both are systematically under-estimated because they feel like they should be quick.
Two consequences the program acted on:
- Contract and IRB run in parallel, not in series. Regulatory document collection and IRB submission begin while the budget is still under negotiation. Sequencing them would add six weeks per site.
- A site is not counted as activated until green light. Counting at contract execution or at site initiation visit produces a flattering number and an inaccurate forecast — a site that cannot screen has not activated in any sense the enrolment curve cares about.
4. Monitoring Model
The program uses risk-based monitoring under ICH E6(R2), rather than the historical model of frequent on-site visits with 100% source data verification.
| Activity | Frequency | Purpose |
|---|---|---|
| Centralized monitoring | Continuous | Statistical surveillance of all sites: enrolment rate, screen-failure rate, query rate, protocol deviation frequency, unusual data patterns. Cheap, and catches what on-site visits miss because it sees every site at once. |
| Remote source data review | Monthly | Targeted review of critical data points against source, conducted remotely where site systems permit. |
| Routine on-site visit | Every 12 weeks | Standard cadence for a site performing normally. Reduced from the historical 4–6 week interval, which consumed monitor capacity without proportionate quality benefit. |
| Triggered on-site visit | On signal | Escalated frequency where centralized monitoring flags a site: deviation clustering, query ageing, enrolment anomaly, or a consent-process concern. |
| For-cause audit | As required | Quality Assurance, independent of clinical operations. Triggered by suspected serious non-compliance or data integrity concern. |
Why the industry moved away from 100% source verification. The traditional model sent
monitors to every site every four to six weeks to check every data point against source records.
It was enormously expensive, and the evidence showed it caught relatively few errors that
mattered — because errors that matter are not evenly distributed either. They cluster at
particular sites, in particular data domains, at particular times.
Centralized statistical monitoring sees every site simultaneously and flags the anomalies. On-site effort then goes where the signal is. This is the same logic as the site-tier analysis in the Enrolment Dashboard: uniform effort against non-uniform risk wastes most of the effort.
Centralized statistical monitoring sees every site simultaneously and flags the anomalies. On-site effort then goes where the signal is. This is the same logic as the site-tier analysis in the Enrolment Dashboard: uniform effort against non-uniform risk wastes most of the effort.
5. What Gets Verified
Risk-based does not mean less rigorous. It means rigor concentrated where an error would change something.
| Data | Verification | Rationale |
|---|---|---|
| Informed consent — presence, version, date, signature | 100% | Non-negotiable. A participant dosed without valid consent is a reportable violation. |
| Eligibility criteria at randomization | 100% | An ineligible randomization contaminates the analysis population. |
| Primary endpoint data | 100% | The measurement the entire program rests on. |
| Serious adverse events | 100% | Regulatory reporting obligation attaches. |
| Investigational product accountability | 100% | Chain of custody from depot to participant. |
| Secondary endpoints | Risk-based sample | Sampled, with sample size increasing where centralized monitoring flags the site. |
| Non-critical data | Centralized only | Reviewed statistically rather than visit by visit. 100% source verification of everything is expensive and does not improve data quality proportionately. |
The five 100% categories are not negotiable and are worth understanding individually.
Consent, because a participant dosed without valid consent is a reportable violation regardless
of outcome. Eligibility, because an ineligible randomization contaminates the analysis
population. Primary endpoint, because it is the measurement the program exists to produce.
Serious adverse events, because a regulatory reporting clock attaches. Product accountability,
because chain of custody must be demonstrable.
Everything else is sampled. A monitoring plan that treats all data as equally critical has not made a risk judgment — it has avoided making one.
Everything else is sampled. A monitoring plan that treats all data as equally critical has not made a risk judgment — it has avoided making one.
6. Performance Management
Underperformance is handled in escalating tiers rather than by immediate replacement.
| Tier | Trigger | Action |
|---|---|---|
| Tier 1 — coaching | Below plan, cause understood | Monitor works with coordinator on referral workflow and pre-screening. No escalation. |
| Tier 2 — action plan | Below plan two consecutive periods | Written plan with dated commitments, signed by the principal investigator. Reviewed at the next monitoring visit. |
| Tier 3 — sponsor intervention | No improvement against action plan | Senior clinical operations and medical monitor engage the investigator directly. |
| Tier 4 — closure | Not remediable, or quality concern | Site closed to further enrolment. Enrolled participants continue to protocol; they are never transferred for convenience. |
Tier 4 carries a provision that is easy to miss and matters enormously. When a site is
closed to further enrolment, participants already enrolled continue to protocol at that
site. They are not transferred for the sponsor's convenience.
Transferring a participant means a new investigator, a new site, disrupted continuity of care, and a fresh consent process — imposed on someone who agreed to take an experimental drug partly on the basis of their relationship with a specific clinician. A sponsor may end its commercial relationship with a site. It does not thereby end its obligations to the people that site enrolled.
Transferring a participant means a new investigator, a new site, disrupted continuity of care, and a fresh consent process — imposed on someone who agreed to take an experimental drug partly on the basis of their relationship with a specific clinician. A sponsor may end its commercial relationship with a site. It does not thereby end its obligations to the people that site enrolled.
Applied to the current position: 94 sites are below plan and 17 have randomized nobody. Those are Tier 1 and Tier 2 populations, not Tier 4 — underperformance is not itself a quality concern, and a site that has not yet enrolled may simply have activated late.
7. The Reserve List
34 sites were qualified at protocol finalization and deliberately held in reserve — assessed, scored and contract-ready, but not activated.
Why this is a planning practice worth copying. Qualifying a site takes
22 weeks. A program that begins qualification when it discovers it is
behind has already lost more time than the shortfall it is trying to recover.
The reserve list converts a five-month problem into a six-week one. It costs the assessment effort for sites that may never be used, which is a real and modest cost, paid against a risk that materialized here exactly as anticipated (R-05).
The reserve list converts a five-month problem into a six-week one. It costs the assessment effort for sites that may never be used, which is a real and modest cost, paid against a risk that materialized here exactly as anticipated (R-05).
Reserve sites are being drawn under the recovery plan in the Enrolment Dashboard §6, sequenced by score with population access weighted first.
8. Sponsor and CRO Responsibilities
Trial execution is contracted to Meridian Clinical Research (Decision D-04). The division of labor is explicit because accountability for trial conduct cannot be contracted away.
| Activity | Meridian (CRO) | Vitalis (sponsor) |
|---|---|---|
| Site identification and feasibility | Executes | Approves criteria and final list |
| Contract and budget negotiation | Executes | Sets budget parameters and approves outliers |
| Site initiation and training | Executes | Approves training content |
| Routine monitoring | Executes | Reviews reports, sets risk thresholds |
| Centralized monitoring analytics | Executes | Defines signals and escalation triggers |
| Medical monitoring | — | Retained by sponsor |
| Protocol interpretation and deviations | Escalates | Decides |
| Site closure decision | Recommends | Decides |
| Regulatory reporting | Supports | Retained by sponsor |
| Inspection response | Supports | Retained by sponsor |
The pattern in the right-hand column. Everything the sponsor retains is a matter of
medical judgment or regulatory liability. Everything delegated is execution capacity.
That line is not a commercial preference — it is the regulatory position. A sponsor remains accountable for trial conduct regardless of what it has outsourced, and a sponsor that has delegated medical monitoring or protocol interpretation has delegated something it will nonetheless be held responsible for at inspection. Oversight mechanics are set out in the CRO Oversight Plan.
That line is not a commercial preference — it is the regulatory position. A sponsor remains accountable for trial conduct regardless of what it has outsourced, and a sponsor that has delegated medical monitoring or protocol interpretation has delegated something it will nonetheless be held responsible for at inspection. Oversight mechanics are set out in the CRO Oversight Plan.