Why Admission Needs More Than a Scientific “Yes”
A candidate treatment can be scientifically compelling and still be unsuitable to enter a platform now. Sites may lack trained staff, investigational product may not be available, or the platform may have no capacity to support the arm’s data and safety requirements. Conversely, operational convenience cannot justify adding an arm whose scientific question is weak or whose rationale is not adequately supported.
Admission criteria make these judgments explicit before a candidate is considered. They also help distinguish two decisions that are easily conflated: whether an arm is approved to join the platform, and whether it is ready to begin operating. Approval can be conditional on completing specified prerequisites; activation should occur only when the required prerequisites are verified.
The earlier material on governance layers, decision rights, scientific committees, and governance charters provides the context for assigning responsibilities. Here the focus is the decision standard: what evidence should be reviewed, which conditions are mandatory, and how an approval becomes an operational activation.
Three Domains of Admission Criteria
A practical admission framework examines scientific rationale, operational readiness, and required resources. The criteria should be written in terms of evidence that can be reviewed, not broad aspirations such as “high potential” or “adequate support.” For each criterion, the governance documents should identify what counts as evidence, who verifies it, and how a shortfall is handled.
- Scientific rationale. The candidate should address a clinically relevant question, have a defensible evidentiary basis, and fit the platform’s scope. Review should establish why the question merits study, whether the proposed comparison is interpretable, and whether the available evidence supports conducting the study in the proposed setting. The committee should record important uncertainty rather than treating absence of evidence as evidence of suitability.
- Operational readiness. The arm must be implementable under the platform’s procedures. This includes a workable study plan, defined responsibilities, feasible site procedures, appropriate data collection, and an executable safety-monitoring approach. Readiness is supported by concrete checks—for example, confirmation that sites can perform required assessments—not by a general expectation that problems can be solved after recruitment starts.
- Resources. The platform and responsible parties must identify and commit the personnel, funds, product supply, systems, and oversight capacity that the arm requires. Resource review should include both arm-specific needs and the effect of those needs on shared platform capacity. A nominal budget line is not sufficient if the funding is unavailable, restricted, or does not cover the planned activities.
The criteria should distinguish mandatory conditions from considerations that inform judgment. For example, a documented rationale may be mandatory, while the relative priority of two scientifically credible candidates may be a comparative consideration. A platform may define minimum evidence thresholds, but it should not disguise a discretionary judgment as a numerical score unless the scoring method is justified, prespecified, and suitable for the decision.
From Admission Review to Activation
A useful decision structure separates assessment, approval, and activation. At the admission review, the authorized body evaluates whether the candidate meets the scientific standard and whether the proposed arm is sufficiently specified to enter the platform. If some operational details remain incomplete but are resolvable, the decision may be approval in principle with explicit conditions and owners. It is not permission to recruit.
Before activation, the responsible operational authority checks that each condition designated as a prerequisite has been completed and documented. Examples include confirmation of staff training, verified product availability, functioning data workflows, and completed required local or institutional authorizations. The decision record should state who verified each item, the evidence reviewed, the date, and the effective activation date.
This logic can be expressed simply. Let \(S\), \(O\), and \(R\) indicate whether the scientific, operational, and resource admission criteria, respectively, are satisfied. Let \(V\) indicate that all designated activation prerequisites have been verified. In a framework where all three domains are mandatory, admission approval requires all three, while activation additionally requires verification of prerequisites.
The conjunction matters: if even one mandatory domain is false, the overall condition is false. A conditional approval can be recorded separately as a governance outcome, but it must not be mistaken for satisfaction of the activation rule.
Worked Example: A Strong Rationale With an Unresolved Supply Plan
All figures and circumstances in this example are hypothetical. A platform is considering an arm to evaluate Treatment A for a serious condition. The scientific committee finds a clear clinical question, a plausible supporting evidence base, and a proposed comparison that fits the platform’s scope. It records the scientific criterion as met. The operational team confirms that the proposed assessments and site procedures are feasible. However, the sponsor has not confirmed that sufficient product will be available for the planned initial enrollment.
Suppose the prespecified resource criterion requires a confirmed supply of at least 120 treatment courses before activation. The sponsor documents 90 courses currently available and 40 more expected, but the additional 40 are not contractually confirmed. The documented, confirmed supply is therefore \(90\), not \(90+40=130\). Since \(90<120\), the resource criterion is not met on the available evidence.
The gate results are \(S=1\), \(O=1\), and \(R=0\). Therefore, admission approval under the conjunctive rule is \(1\land1\land0=0\). The governance body may defer the decision or record a conditional outcome if its charter allows one, with a named owner and deadline for confirming supply. It should not authorize activation on the basis of an expected but unconfirmed delivery.
If a binding confirmation later establishes the full 130 courses, the resource threshold is met because \(130\ge120\). That resolves this particular resource condition, but activation still depends on verifying every other designated prerequisite. The calculation does not establish that the arm is scientifically appropriate; that is a separate gate.
Justifying the Gate Structure
A weighted total can be attractive because it appears to summarize several kinds of evidence in one number. But it can permit an unacceptable trade-off: a very high score in one domain might offset failure in another domain that is essential. When the decision requires every mandatory condition, a conjunction is the direct representation of the rule.
The justification follows from the meaning of “required.” If the arm is suitable only when all three requirements hold, then the set of acceptable cases consists exactly of the cases in which \(S=1\), \(O=1\), and \(R=1\). There are \(2^3=8\) possible combinations of binary gate results. Only one combination satisfies all three:
Thus the rule \(S\land O\land R\) approves precisely the cases meeting every mandatory condition. A sum such as \(S+O+R\) would not preserve that meaning: a total of 2 could represent either a missing scientific criterion, a missing operational criterion, or a missing resource criterion. Those failures are not necessarily interchangeable, so the total alone is insufficient for a sound decision.
This argument does not mean that every relevant consideration must be binary. Committees may need structured judgment for evidence quality, clinical importance, or comparative priority. The point is narrower: where a condition is genuinely mandatory, it should remain visible as a gate rather than being diluted into an aggregate score.
Worked Example: Approval in Principle Is Not Activation
All details are hypothetical. A platform’s admission criteria require a defensible scientific rationale, a feasible operational plan, and identified resources. Its activation checklist separately requires trained site staff, a tested data-entry workflow, and verified availability of study product.
The candidate arm meets all three admission criteria, so \(S=1\), \(O=1\), and \(R=1\), giving \(S\land O\land R=1\). The committee approves the arm to join the platform. At the activation review, the data workflow has passed testing and product availability has been verified, but only 7 of the 10 designated site coordinators have completed training.
Let the three activation checks be \(V_T\) for training, \(V_D\) for data workflow, and \(V_P\) for product availability. The checklist requires \(V=V_T\land V_D\land V_P\). Here, \(V_T=0\), \(V_D=1\), and \(V_P=1\), so \(V=0\land1\land1=0\). The arm is approved for admission but is not activated.
The appropriate record identifies the three untrained coordinators, the person responsible for completing their training, and the evidence required to verify completion. Once all 10 have completed the required training and the other checks remain valid, the operational authority can recheck the full activation checklist and record the effective activation date. Approval in principle does not itself start recruitment.
Applying Criteria Consistently
Criteria are useful only when they are applied consistently and supported by an auditable record. A review should connect each decision to the evidence actually available at that time. A short record can use the following sequence:
Identify whether it is mandatory for admission, a consideration for prioritization, or a prerequisite for activation.
Record the document, confirmation, or assessment used, including its date and any material uncertainty.
Mark the criterion met, not met, or unresolved, and explain the decision in context rather than relying on a score alone.
For an unresolved condition, name the responsible party, required evidence, and decision-maker for the next review.
Distinguish deferral, rejection, approval subject to conditions, and activation; record the date and any restrictions.
Worked Example: Testing Capacity Rather Than Assuming It
All numbers are hypothetical. A platform estimates that a candidate arm needs 1.5 full-time-equivalent (FTE) data-management capacity for its first 6 months. The platform has 2.2 FTE available for additional work during that period. A separate, already approved activity is forecast to require 0.9 FTE. If those estimates use the same time period and measure, the remaining capacity is \(2.2-0.9=1.3\) FTE.
The candidate arm’s estimated need exceeds the remaining capacity: \(1.5>1.3\) FTE, a shortfall of \(1.5-1.3=0.2\) FTE. Under a criterion requiring identified capacity for the full planned workload, the resource gate is not met unless the platform documents an additional 0.2 FTE or revises the supported plan through the authorized process. Simply saying that staff will “absorb the work” does not resolve the quantified shortfall.
The arithmetic is meaningful only if the inputs are comparable and credible. The review should check that both estimates cover the same 6-month interval, include the relevant data tasks, and reflect staff who are actually available. If the 2.2 FTE figure is an annual average or includes work already committed elsewhere, subtracting 0.9 would produce a misleading capacity estimate. The correct governance response is to clarify the evidence before marking the gate met.
Alternatives, Diagnostics, and Limitations
A purely discretionary case-by-case review can accommodate unusual circumstances, but it risks inconsistent decisions and undocumented shifts in standards. Fixed criteria improve predictability and auditability, though criteria that are too rigid may fail to recognize a genuinely important exception. If exceptions are allowed, define who can authorize them, what rationale must be recorded, and whether an exception can waive a mandatory safety or legal requirement. Do not imply that an exception exists if the governance documents do not permit one.
A numerical readiness score can help organize discussion, especially for nonmandatory comparative considerations. It is not a substitute for checking each essential condition. If scoring is used, weights and thresholds should be specified in advance, the underlying evidence should remain visible, and a low score in a mandatory domain should not be hidden by a high score elsewhere.
Warning signs that the gate process is failing include criteria being rewritten after a candidate is reviewed, “met” findings without supporting evidence, different standards for similar candidates without explanation, resource assumptions that count unconfirmed commitments, and approval language that sites interpret as permission to recruit. A useful diagnostic is to ask whether an independent reviewer could reconstruct the decision from the criteria, evidence, and record.
Admission gates also have limits. They structure governance judgment; they do not guarantee that a scientifically sound arm will produce useful results or that operational risks will not arise. Forecasts of capacity and supply can be wrong, evidence can be incomplete, and a checklist cannot replace expert review. The framework is strongest when it clarifies responsibility and uncertainty rather than presenting a complex decision as mechanically certain.
Common Mistake
Wrong reasoning: “The candidate has excellent scientific support and strong sponsor interest, so we can approve it now and resolve staffing and product supply after sites begin enrolling.” This treats essential conditions as compensable and collapses admission into activation.
Correction: Evaluate the prespecified mandatory domains separately. Record the scientific finding, determine whether operational and resource criteria are met, and distinguish any conditional admission outcome from activation. Recruitment should not begin until the required activation evidence has been verified and the authorized decision-maker has recorded the activation.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions.
- A candidate meets the scientific criterion and has identified funding, but its required site procedure has not been tested. If operational readiness is mandatory, can the arm be approved under a three-gate conjunctive rule?
- A resource criterion requires 80 confirmed treatment courses. A candidate has 55 in stock and 30 expected without a binding confirmation. How many courses count as confirmed, and is the criterion met?
- All admission criteria are met, but the activation checklist requires completed staff training and only 18 of 20 designated staff are trained. Is the arm admitted, activated, both, or neither?
- Why might a weighted score of 9 out of 10 be inadequate evidence for admission if one mandatory domain is not met?
Solutions. (1) No. With the scientific and resource gates met but the operational gate unresolved, the conjunction is \(1\land0\land1=0\); the procedure must be tested or the governance outcome otherwise resolved under the prespecified rules. (2) Only 55 courses are confirmed. Since \(55<80\), the criterion is not met; the expected 30 do not count as confirmed supply. (3) The arm is admitted but not activated. Admission criteria are met, but the training prerequisite is not, so the activation checklist is incomplete. (4) A weighted score can conceal which domain failed and may let strength in other domains offset a mandatory failure. The reviewer must inspect each gate and apply the stated mandatory conditions.