[ ENTERPRISE PILLAR 04 ]
Protect participants while producing reliable clinical evidence through controlled study design, execution, oversight, data, and reporting.
This pillar owns trial strategy and execution; safety case management after and across trials connects to the safety-surveillance pillar.
The capability framing below, its failure modes and the boundary with neighbouring pillars are SPEQ’s practitioner reading — not a regulatory requirement, and not an assessment of any organization.
THE CAPABILITY
This is the only capability in the enterprise whose raw material is other people's willingness to take part, and whose output has to be reconstructable years later by a stranger, from records made largely by people the organization does not employ. It behaves like a distributed manufacturing operation whose product is evidence and whose plants are hospitals. The scientific half and the operational half are one capability rather than two because they fail into each other: a study that cannot be run as written yields a dataset that does not answer the question it was designed to ask, and no amount of operational excellence rescues a design that asked for something impossible in a clinic. Everything else here — sites, providers, the data, the essential records — exists to keep one chain intact, running from the intent behind an endpoint to the number that eventually appears in a report somebody else will read.
Almost none of the work is done by the organization answerable for it. A sponsor is accountable for conduct performed by investigators it does not manage, inside institutions with their own priorities and their own queues, through providers it selected but does not run, involving people who may stop at any moment and owe nobody an explanation. Authority and accountability are separated by contract, so every control has to be exercised as persuasion, documentation or detection rather than instruction, and the strength of any control is bounded by how fast the sponsor can see what has already happened. The evidence is perishable in a way nothing else here is. A measurement not taken at a visit is gone: no retained sample, no repeat run, no second pass at that hour in somebody's life. Error is spent forward, and a quality problem found late can be described accurately but never repaired. Then there is the record. It has to let somebody who was not present reconstruct events several years afterwards, and it is written at the point of maximum operational pressure by clinical staff for whom the study is one demand among many and rarely the most urgent one. Each of those constraints is structural, and none of them yields to running the constituent activities well.
Each of these happens with the individual branches below being run competently. That is what makes them capability failures rather than performance problems.
The schedule is scientifically defensible and unaffordable inside a real clinic day. Sites absorb the difference with sequencing workarounds, and the departures that follow are read as site quality rather than as a design signal being returned by the network. The diagnostic is simple and rarely applied: if the same departure appears at most sites, it is a property of the protocol, not of the sites.
Visits check what is checkable — fields against source, consent dates, product accountability — and often check it well. Meanwhile the thing that will undermine an analysis is a pattern across sites: an entry criterion consistently read one way at the highest-enrolling centres, an assessor whose scoring drifts across a year. No single visit can see a pattern that exists only in aggregate, and the visit report has nowhere to put one.
Documents are collected ahead of an inspection instead of being filed as events happen. What emerges is a plausible history rather than a contemporaneous one, and the gaps land exactly where events were unusual, because unusual events were handled by people who were busy. Timing cannot be added afterwards: a file that became complete last month says something quite different from one that was complete throughout.
Governance meetings run on measures the provider produced about its own performance, so escalation depends on the provider deciding an issue is worth escalating, and the categories depend on how the provider defines its own terms. The sponsor learns about the problems the provider is comfortable reporting, on the provider cadence, and attendance at the meeting gets mistaken for the oversight it was meant to constitute.
WHERE THIS STOPS
This pillar owns the trial. It does not own the safety picture across and beyond it, and that is the seam that causes the most argument at the worst possible moment. An event reported at a site is at once a study event and a safety case, and the two obligations begin on different documents, in different systems, on clocks that start at different instants — so the dispute is rarely about what happened and almost always about when somebody first knew. The second seam is investigational supply: this capability owns whether the right participant received the right kit at the right time, the manufacturing and supply capabilities own whether it was made, labelled and shipped under control, and an excursion in transit is the point at which each assumes the other is judging fitness for use. The third is data, where clinical data management owns the study database and the meaning of what sits in it while the platforms, access model and retention belong to the digital pillar. The fourth is upstream: development characterizes the thing being tested, and this pillar produces the evidence of what it does in people.
Independent sight of the underlying work, at a frequency proportionate to what could go wrong: access to the data rather than a summary of it, sight of the provider quality events rather than a count of them, and at least one route by which a problem can reach the sponsor without passing through the party it concerns. A contract allocates responsibility; it does not create visibility.
It is broader than monitoring, and calling it a monitoring model tends to shrink it to the part that is easiest to buy. It starts at design, by deciding which few factors genuinely determine whether the study can answer its question, and then shapes the protocol, the provider scope, the data flow and the oversight around those factors. Monitoring is one instrument it selects, not the thing itself.
Because its evidential value comes from when documents arrived, not only from whether they are present. A file that was complete and contemporaneous shows an operation that was under control while it ran; a file completed retrospectively shows an operation that could be described afterwards. Only the first supports the reconstruction a later reader needs, and the difference between them is not recoverable.
As soon as it touches something the analysis depends on: eligibility, the timing of a primary assessment, the integrity of blinding, or the conditions under which a measurement was made. The routine that works is to ask, at the point of recording, which analysis population or endpoint the departure could move, and to route it on that basis rather than filing every departure into one undifferentiated queue.
CAPABILITY BRANCH MAP
The plan for what evidence the programme must generate: endpoints, populations, phases, regions, comparators and the decision criteria that determine whether to continue after each stage.
A trial that runs perfectly against the wrong question wastes years and exposes participants for nothing. Strategy is where the evidence a regulator and a payer will each require is reconciled, before either can be asked.
Turning the strategy into an executable study: objectives and estimands, endpoints, sample size, eligibility, assessment schedule, and the identification of what is genuinely critical to quality.
The protocol is the single document that governs everything sites do. Complexity added here multiplies across every participant at every visit, and the burden falls on people who had no part in writing it.
Selecting countries and sites, activating them, agreeing contracts and budgets, confirming they are genuinely ready, and supporting them for the duration of the study.
Sites are where the protocol meets reality, and they are independent organisations with their own pressures. A site activated before it is ready produces the deviations that consume the rest of the study.
The protections owed to the people in the study: ethics review, informed consent as an ongoing process, recruitment and retention practices, privacy, and communication with participants including about results.
This is the ethical foundation the entire discipline rests on, and the one failure that cannot be remediated after the fact. Consent obtained without genuine understanding is not consent, however complete the signature page.
Running the study: conduct against the protocol, monitoring, deviation handling, investigational product accountability, vendor coordination and escalation when something goes wrong.
Execution is where a good protocol either produces reliable evidence or does not. Deviations are the visible signal, and the way they are handled determines whether an issue is contained at one site or discovered across all of them at the end.
Directing oversight at what matters: identifying critical-to-quality factors, setting indicators and tolerance limits, monitoring centrally, and managing the issues that surface — rather than checking everything everywhere equally.
Effort spent verifying non-critical data is effort not spent on the factors that determine whether the trial can be relied on. Risk-based quality management is now the expectation rather than an alternative to full source verification.
Everything that happens to trial data between collection and analysis: capture, coding, query management, reconciliation across sources, database lock, statistical analysis and the traceability that connects result to source.
The analysis is only as trustworthy as the path back to the source record. Where that path is broken, the result cannot be defended — and an inspector who cannot reconstruct a data point will not accept the conclusion built on it.
The essential records that allow the trial to be reconstructed: what exists, whether it is complete and contemporaneous, who can access it, and how it is archived for the retention period.
The trial master file is the trial as far as an inspector is concerned. A study conducted correctly but documented late is indistinguishable, from the outside, from one conducted badly.
The external organisations and technologies the trial depends on — contract research organisations, central laboratories, imaging providers, eClinical platforms — and the sponsor oversight that remains regardless of delegation.
Activities can be delegated; accountability cannot. Sponsors are held to what their vendors did, and the oversight failure is usually not that a vendor performed badly but that nobody was checking in a way that would have noticed.
Ending the study properly: site closeout and reconciliation, the clinical study report, registry disclosure and results posting, retention of records, and communication back to participants.
Disclosure obligations are legal duties with deadlines, and they are enforced independently of how the trial went. Closeout is also the last point at which a missing record can still be recovered from a site that still exists.
CONNECTED OPERATING MODEL
MATURITY ORIENTATION · SPEQ SYNTHESIS
HIGH-VALUE INTERSECTIONS
SOURCE BASIS
REGULATORY BASIS
The 12 standards SPEQ maps to this pillar, and the 6 regulatory bodies behind them. Which standards belong to a pillar is a SPEQ judgement; the bodies, disciplines and industries below are read from the standards themselves.
EC · EMA · FDA · ICH · ISO · VICH
Ranked by how many of this pillar’s standards each one cites — showing 10 of 18.
These 8 standards govern the system classes this pillar depends on rather than the pillar itself. The distinction matters: a standard that governs a system is not thereby a standard of every capability that uses it.
21 CFR Part 11ISPE GAMP 5 (2022)MHRA GxP DI (2018)EU GVP ModulesICH E2AICH E2B(R3)ICH E2C(R2)21 CFR 314.80
PROFESSIONAL · READINESS ORIENTATION
Rate observable operation from 0 (not established) to 4 (adaptive). The protected output prioritizes operating dimensions and evidence—not a compliance score.