Best practice in clinical trial site selection is to choose each site against evidence it has already produced, rather than against the estimate it supplies on a feasibility form. A clinical trial site is the organisation and location where participants are screened, consented, treated and followed up under one protocol. The site’s screening logs, staffing records and set-up history predict delivery more reliably than any patient number written on a questionnaire.
The stakes changed on 28 April 2026. The amended UK clinical trials regulations require every trial of an investigational medicinal product to recruit its first UK participant within two years of approval, and the approval lapses where recruitment does not start and no extension is granted. Site selection now carries a regulatory consequence alongside the commercial one.
This guide covers who confirms a site can run a study, what the 2026 UK rules put on the selection decision, where selection fails in practice, the nine criteria that predict delivery, how to test a site’s real access to its population, and the measures that show afterwards whether the choice was right.
What Is a Clinical Trial Site, and Who Confirms One Can Run a Study?
A clinical trial site is the place where the protocol meets a participant. The category covers NHS trusts and health boards, primary care practices, university teaching hospitals, clinical research facilities and dedicated commercial research units. Each type brings a different mix of population access, staffing depth and set-up speed.
- NHS trusts and health boards hold the largest secondary care populations, and their set-up runs through a research and development office with its own queue.
- Primary care practices reach large unselected populations for prevention and public health studies, and clinical session time limits their capacity.
- University teaching hospitals combine specialist services with experienced research teams, and they carry the heaviest load of competing studies.
- Clinical research facilities hold purpose-built space, trained delivery staff and established procedures, and they depend on referral routes for participants.
- Commercial research units hold volunteer panels and early phase capability, and their fit narrows to the protocols their panel matches.
Two separate decisions sit behind every active site. The sponsor selects the site and appoints the investigator. MHRA guidance states that the appointment of an investigator is the responsibility of a sponsor, and that the sponsor must ensure each investigator is appropriately trained under regulation 3B, as well as qualified by education and experience. The participating organisation then makes the second decision and confirms it holds the capacity and capability to deliver the study.
The two decisions run in sequence, and the first one is made early. The Health Research Authority expects a sponsor to have identified potential participating sites and discussed the project with local research teams before the IRAS application is submitted. Site selection is therefore the earliest committed decision in a UK study, and it is made on the least evidence.
What Did the April 2026 UK Regulations Change for Site Selection?
The Medicines for Human Use (Clinical Trials) (Amendment) Regulations 2024 came into force on 28 April 2026 and apply across all four UK nations. Five of the changes land directly on the sites a sponsor picks, because each one attaches a dated obligation to a study record that individual sites populate.
| What changed on 28 April 2026 | What it means for the site you choose | The record that proves it |
| Every trial of an investigational medicinal product recruits its first UK participant within two years of approval, and the approval lapses where no extension is granted | Early recruitment capability outranks total capacity, because a slow site puts the authorisation at risk | Approval date, site confirmation date and first participant date on one study record |
| Registration in a public registry and publication of a summary of results became a legal requirement, with results due within 12 months of the end of the trial | Sites that stall recruitment extend the whole reporting chain behind them | A defensible trial end date and a milestone history |
| Amendments became modifications, classified as substantial, a modification of an important detail, or minor | A late site swap is a governance event, and it costs approval time as well as set-up time | A per-site version history of which documents were live on which date |
| A notification scheme and a faster route for eligible modifications took effect, with those modifications approved automatically unless concerns are raised within 14 calendar days | Regulatory time shrank, and local site set-up became the binding constraint | Set-up milestones tracked per site rather than per study |
| Every trial follows the principles of Good Clinical Practice, and the sponsor must ensure each investigator is appropriately trained under regulation 3B | Investigator training and delegation move from a green light formality to a selection criterion | Training records and a delegation log tied to the protocol version in force |
The two-year condition is the one that reframes the whole exercise. Government figures published in April 2026 record that the combined MHRA and HRA review takes an average of 41 days, and that set-up time for studies going through combined review fell from 169 days to 122 days. Approval arrives quickly, local set-up holds the delay, and the two-year clock runs through both. The UK clinical research delivery KPIs track that split across the national portfolio.
Selection used to risk a slow study. It now risks the authorisation.
Also Read: Capacity and Capability Confirmation: Timelines and Delays
Where Does Clinical Trial Site Selection Go Wrong?
Site selection fails at the question, not at the answer. A feasibility questionnaire asks how many eligible patients a site sees. The site answers honestly from the data it can reach quickly, which is usually a diagnosis code count from an electronic patient record. The number is accurate and the question was the wrong one.
A worked example shows the gap. Northgate General NHS Foundation Trust is a fictional site selected for an interventional study on the strength of a feasibility return stating approximately 240 eligible patients per year. Twelve months after green light the site has screened 11 and randomised 3.
- The 240 figure came from a prevalence count of everyone in the trust carrying the diagnosis code, including patients last seen four years ago.
- The protocol excludes two common comorbidities, which removes roughly half of that group before any clinic contact takes place.
- The delegation log records that two of the four staff named on the feasibility return moved posts before the site opened.
- The only research pharmacist covers three other studies, and the investigational medicinal product needs preparation inside a two-hour window.
- A competing study opened in the same clinic six weeks earlier and screens from the same list.
Each of those five facts existed on the day the site was selected. None appeared on the form, because the form asked for an estimate and the site supplied one. A screening log, a current delegation log and a pharmacy workload list would have exposed all five before contract signature.
A feasibility estimate is a forecast. A screening log is a record.
Which Criteria Predict Whether a Site Delivers?
Nine criteria carry most of the predictive weight in UK studies. Each one has a version the questionnaire asks and a version the record answers, and the difference between the two is where selection quality sits. Work the third column and the site’s estimate becomes checkable.
| Criterion | What the questionnaire asks | What to verify instead |
| Population access | Number of eligible patients seen per year | A screening log from a comparable protocol, with screened, consented and randomised counts |
| Investigator time | Whether the principal investigator has capacity | Open studies the investigator currently holds, and the clinical sessions committed to each |
| Coordinator capacity | How many research nurses and coordinators the site employs | Funded whole time equivalent per open study, and which posts are vacant or fixed term |
| Pharmacy and IMP handling | Whether a research pharmacy exists | Preparation windows, temperature-controlled storage, out-of-hours cover and current study load |
| Specialist services | Whether imaging, laboratory and other services are available | Written agreement from each supporting department, plus the HRA technical assurance position |
| Competing studies | Whether competing studies are running | Every open and planned study drawing from the same clinic list, with target and end date |
| Set-up performance | How quickly the site sets up studies | Elapsed days from approval to first participant on the site’s last five studies |
| Data and documentation quality | Whether the site follows good clinical practice | Query counts, deviation rates per participant, and any audit findings with their CAPA status |
| Retention | Whether participants complete the study | Withdrawal and lost-to-follow-up rates on comparable protocols, with the visit burden of each |
Three of these deserve a note. Investigator time causes more slow starts than any other factor, and it is the criterion most often assessed by asking the investigator. Competition operates at the level of the clinic list rather than the therapy area, because two protocols with different indications still screen the same Tuesday morning cohort. Set-up performance transfers most reliably of the nine, since a site that took 240 days on its last three studies has a structural constraint a fourth study inherits.
Also Read: 5 Proven Patient Recruitment Strategies in Clinical Trials
How Do You Test a Site’s Real Access to the Population?
Eligible on a database and able to attend every visit are two different populations, and the protocol decides how far apart they sit. A single screening visit keeps almost everyone in the catchment. A protocol with fortnightly attendance and a four-hour pharmacokinetic day shrinks it to the people who live close enough and have the working pattern to absorb it.
Five checks convert a headline number into a recruitable one.
- Ask which query produced the figure and over what period, which separates an annual incidence count from a standing prevalence count presented as an annual one.
- Apply the protocol exclusions to the site’s own denominator before the call ends, which usually removes between a third and two thirds of the headline group and does so on the site’s data rather than an assumption.
- Map the visit schedule against travel time from the site’s actual referral area, which shows how much of the catchment the visit burden removes.
- Request a screening log from one comparable completed study, which gives a real conversion rate from approached to randomised at that site with that team.
- List every study screening from the same clinic list, which exposes the competition for the same patients on the same day.
Those five checks produce a recruitment forecast built from the site’s own history rather than its expectations. A site that converted 1 in 8 approached participants on a comparable protocol is likely to convert at a similar rate on this one.
What Happens Between Selection and the First Participant?
Selection commits a site to a set-up path with several owners, and the site controls only part of it. The sequence below is the England route for an NHS organisation. Each step produces a named artefact, and a missing artefact stops the step behind it.
- Preliminary discussion and Part C of the IRAS form record the site as a participating organisation before the application is submitted.
- The local information pack reaches the site after HRA approval and carries the documents the organisation needs to assess the study.
- The costing and contracting route follows the study type. Commercial contract studies are priced through the National Contract Value Review, and non-commercial studies attribute activity through a Schedule of Events Cost Attribution Tool.
- Specialist review settles pharmacy and radiation questions, and HRA technical assurance answers them once for every site rather than site by site.
- Capacity and capability confirmation is the organisation’s statement that the arrangements are in place, and the Date Site Confirmed records the last contract signature across all organisations at that site.
- Green light and first participant follow, and the site’s milestone dates flow to the national record through CPMS and the local portfolio management system.
Two properties of this sequence bear on selection. Steps three, four and five sit outside the research team that answered the feasibility questionnaire, so a site’s clinical enthusiasm predicts very little about its set-up speed. The steps also run partly in parallel, and a site that starts contracting during specialist review recovers weeks a sequential site loses.
Also Read: Why NHS Sites Miss the 90-Day Set-Up Target and How to Close It
How Do You Know the Selection Was Right?
Selection quality shows in six measures, all available within the first quarter a site is open. A site heading for a missed target usually declares itself by week twelve, which leaves enough of the two-year window to add a site or revise the target.
- Days from capacity and capability confirmation to first participant measure set-up execution separately from approval time, which isolates the part of the delay a site owns.
- Screened to randomised conversion tests the feasibility estimate against reality, which shows whether the population was misjudged or the outreach was.
- Screen failure reasons grouped by criterion point at the specific exclusion doing the damage, which allows a protocol modification to be argued with evidence.
- Protocol deviations per participant indicate whether the site’s capacity matches its commitments, and capacity planning on spreadsheets is where that mismatch usually hides.
- Visit window compliance shows whether appointments are booked inside the protocol schedule, which prevents deviations that would otherwise be documented after the fact.
- Open query age and volume reveal documentation capacity, which predicts the monitoring load the site will generate for the rest of the study.
Those six measures depend on one condition. The dates, the screening counts and the deviations must sit in the same system, against the same site record, or the comparison across a portfolio of sites turns into a reconciliation exercise. Real-time recruitment visibility makes the week twelve conversation possible at all.
How Does AQ Support Site Selection and Set-Up?
AQ holds site selection, set-up and delivery in one study record, so the evidence a sponsor needs at selection is the evidence the site generates during conduct. The AQ CTMS carries study set-up, site records, milestone dates and recruitment, and the other modules of the AQ Platform hold documentation and quality records against the same sites.
- AQ CTMS records approval, confirmation and first participant dates per site, which turns the two-year recruitment condition into a milestone a research office reviews weekly.
- The visit diary checks every booking against the protocol window at the point of booking, which prevents the deviation rather than recording it.
- AQ Digital DoA holds delegated roles, qualifications and training dates, which shows whether the staff named at feasibility are the staff actually delegated at green light.
- AQ eTMF and AQ eISF hold sponsor and site documentation against one index, which lets a monitor check a site’s file without a separate request to the site.
- AQ ePSF tracks the IMP chain from receipt to destruction, which surfaces the pharmacy constraints that feasibility forms rarely capture.
- AQ QMS and AQ CAPA link deviations to root cause and corrective action, which builds the site performance history the next selection round should read.
AQ is aligned with Good Clinical Practice, UK GDPR and 21 CFR Part 11, and supplies validation, data security and governance evidence for NHS and sponsor procurement. Assurance covers G-Cloud, DSPT and Cyber Essentials. The NHS trust buyer’s guide sets out what a research office checks before go-live.
Book a live demo and the walkthrough runs on your own site list, your set-up milestones and the criteria your feasibility process uses today.
