- A visit window is the range of days, such as Day 29 ± 3, in which a visit still counts.
- The ICH M11 template, adopted in November 2025, asks for allowable windows on all visits and procedures (ICH, 2025).
- Phase III protocols averaged 118.5 protocol deviations, involving 32.8% of enrolled patients (Tufts CSDD, 2022).
- Each window needs six inputs in the protocol: reference date, numbering, target day, tolerance, type of days, anchor.
- Counting each visit from the previous one lets delays add up; counting from Day 1 keeps the planned weeks.
A protocol line that reads "Week 4: Day 29 ± 3" can give two dates for a participant randomized on Monday, March 1, 2027: March 29 if the site counts randomization as Day 1, March 30 if it counts it as Day 0. In the second case, the site's window ends on April 2, one day after the protocol window closes. In a study published in 2022 covering 187 protocols approved between 2013 and 2018, Tufts CSDD counted a mean of 118.5 protocol deviations per phase III protocol, involving 32.8% of enrolled patients. Its authors name participants' difficulty in keeping to the schedule of visits among the usual causes. The date of a visit depends on six inputs the protocol should state: the reference date, the Day 0 or Day 1 convention, the target day, the tolerance on each side, the type of days counted and the anchor for later visits.
What is a visit window in a clinical trial?
A visit window is the range of days around a visit's target day in which the visit still counts as performed on schedule. The protocol expresses it as a target day and a tolerance, such as Day 29 ± 3: the visit is on time anywhere from Day 26 to Day 32. A visit outside that range is a protocol deviation.
The ICH M11 template, the structured protocol format that ICH adopted on November 19, 2025, places the windows in the Schedule of Activities and states that "allowable windows should be stated for all visits and procedures". M11 is a recommended format. Our article on what the M11 structured protocol changes for EDC teams explains what its Schedule of Activities, in Section 1.3, means for the eCRF build. ICH E6(R3), adopted on January 6, 2025, gives the reason for this margin in its Appendix B: including "acceptable ranges for specific protocol provisions" can reduce the number of deviations.
The FDA makes the same point in its draft guidance on protocol deviations of December 2024. Intentional departures for a single participant should be rare, it says, because the protocol should include "appropriate flexibility regarding trial conduct (e.g., eligibility criteria, reasonable visit windows)". The guidance asks sponsors to run a risk assessment before or during protocol development to find the elements, visit windows included, that can be made more flexible.
What should the protocol say about each visit window?
For each window, the protocol should state the reference date, the numbering convention, the target day, the tolerance before and after, the type of days counted and the anchor for later visits. With these six elements written down, a coordinator who reads the protocol reaches the same date as the sponsor, the CRA and the scheduling tool.
The table gives, for each element, a wording that leaves no room for interpretation.
| Element | What the protocol states | Example |
|---|---|---|
| Reference date | The event that starts the count | Randomization, first dose, or the procedure date in a device study |
| Numbering | Whether the reference date is Day 0 or Day 1 | “Randomization is Day 1. There is no Day 0.” |
| Target day | The day each visit aims for | Week 4 visit on Day 29 |
| Tolerance | Days allowed before and after the target, which can differ | ± 3 days, or −0 / +7 days for a visit that cannot take place early |
| Type of days | Calendar days or business days | “All windows are expressed in calendar days.” |
| Anchor | Whether each visit counts from the reference date or from the previous visit | “All visit days count from Day 1.” |
The width of a window follows the purpose of the visit. At the Week 12 visit that measures the primary endpoint, a wide window spreads the measurements around the timepoint the analysis expects. A safety follow-up visit can accept a wider range, such as ± 7 days at Day 113 in the schedule below. Flexibility still has its place: among the ways to reduce deviations, the FDA draft guidance lists "using flexible time frames for collection of essential data where feasible".
How do you calculate a clinical trial visit window?
With randomization as Day 1, the target date of a visit is the Day 1 date plus the target day minus one. The earliest date is the target date minus the tolerance before, and the latest date is the target date plus the tolerance after. If the protocol calls randomization Day 0, add the target day to the Day 0 date.
The table applies this calculation to a fictitious participant randomized on Monday, March 1, 2027, with randomization as Day 1 and all windows in calendar days. In this convention there is no Day 0: the day before randomization is Day −1.
| Visit | Target day | Window | Target date | Earliest date | Latest date |
|---|---|---|---|---|---|
| Screening | Day −28 to Day −1 | None | None | Mon Feb 1, 2027 | Sun Feb 28, 2027 |
| Randomization | Day 1 | None | Mon Mar 1, 2027 | None | None |
| Week 2 | Day 15 | ± 2 days | Mon Mar 15, 2027 | Sat Mar 13, 2027 | Wed Mar 17, 2027 |
| Week 4 | Day 29 | ± 3 days | Mon Mar 29, 2027 | Fri Mar 26, 2027 | Thu Apr 1, 2027 |
| Week 8 | Day 57 | ± 3 days | Mon Apr 26, 2027 | Fri Apr 23, 2027 | Thu Apr 29, 2027 |
| Week 12 | Day 85 | ± 5 days | Mon May 24, 2027 | Wed May 19, 2027 | Sat May 29, 2027 |
| Follow-up | Day 113 | ± 7 days | Mon Jun 21, 2027 | Mon Jun 14, 2027 | Mon Jun 28, 2027 |
When you lay out the dates, three constraints appear. Every weekly target falls on a Monday, because Days 15, 29, 57, 85 and 113 are whole weeks after Day 1: a participant randomized on a Friday gets every target on a Friday. The Week 2 window opens on a Saturday, so a site closed on weekends has three usable days out of five. The Week 4 target, March 29, 2027, is Easter Monday, a public holiday in France and in England and Wales. A site in most of France books that visit on Friday, March 26, or between Tuesday, March 30, and Thursday, April 1. In England and Wales, and in the Alsace-Moselle towns that observe Good Friday, March 26 is a holiday too, which leaves three days out of seven.
You can build the same visit window calculator in a spreadsheet. With the Day 1 date in cell B1, the Week 4 target is =B1+29-1, the earliest date =B1+29-1-3 and the latest date =B1+29-1+3.
The Day 0 convention renames the days and keeps the dates, as long as everyone applies it. In a Day 0 protocol, Week 4 is Day 28 and still falls on March 29, 2027. Mixing the two conventions shifts the dates: a site that reads a Day 1 protocol as Day 0 books every visit one day late, and the last day of its window falls one day after the protocol window closes.
Business days widen the window and tie it to a holiday calendar. Counted as ± 3 business days, the Week 4 window above runs from Wednesday, March 24, to Thursday, April 1, 2027: nine calendar days instead of seven. In a spreadsheet, =WORKDAY(target, -3, holidays) returns the first day and =WORKDAY(target, 3, holidays) the last, and the result changes with the holidays each site enters. Calendar days give every site in every country the same dates. Write the choice into the protocol.
Should a visit window count from Day 1 or from the previous visit?
Count from Day 1 unless the interval between two visits matters more than the timepoint in the trial. Anchoring every visit on the reference date keeps it at its planned week, whatever happened at earlier visits. Anchoring each visit on the previous one keeps the gap between visits constant, and each delay then moves every visit that follows.
The table follows a participant who comes two days after each target, inside every window, under the two anchors.
| Visit | Window | Anchored on Day 1 | Anchored on the previous visit |
|---|---|---|---|
| Week 2 | ± 2 days | Target Day 15, done Day 17 | Target Day 15 (Day 1 + 14), done Day 17 |
| Week 4 | ± 3 days | Target Day 29, done Day 31 | Target Day 31 (Day 17 + 14), done Day 33 |
| Week 8 | ± 3 days | Target Day 57, done Day 59 | Target Day 61 (Day 33 + 28), done Day 63 |
| Week 12 | ± 5 days | Target Day 85, done Day 87 | Target Day 91 (Day 63 + 28), done Day 93 |
Under both rules, each visit stays inside its own window. With the previous-visit anchor, the Week 12 assessment takes place on Day 93, eight days after the Day 85 timepoint the analysis expects, and each new delay adds to the gap. The previous-visit anchor suits visits tied to dosing, such as a dose given every 21 days, where the interval itself protects the participant. When a protocol uses both, it should name the anchor visit by visit.
What happens when a visit falls outside its window?
A visit performed outside its window is a protocol deviation. The investigator documents it, the sponsor's criteria decide whether it is important, and with a Day 1 anchor the next visits keep their planned dates. ICH E6(R3) asks the investigator to document all protocol deviations and to explain those deemed important (section 2.5.3).
The ICH E3 questions and answers of July 2012 define a protocol deviation as "any change, divergence, or departure from the study design or procedures defined in the protocol".
The table sets out what the site does in each case.
| Situation | What the site does | What the site records |
|---|---|---|
| The participant cancels and the window is still open | Rebooks the visit inside the window | The new date and the reason for the change |
| No date is possible inside the window | Applies the protocol rule for that visit: late visit or missed visit | A protocol deviation: planned window, actual date, days outside, reason |
| The visit takes place after the window | Performs the assessments the protocol requires for a late visit | The deviation, reported to the sponsor within the agreed time frame |
| The visit does not take place | Marks it as missed and keeps the next visits on their planned dates | The missed visit, the reason and the assessments not collected |
| The late or missed visit carries a primary endpoint or a safety test | Informs the sponsor without waiting for the next monitoring visit | A deviation the sponsor may classify as important |
The FDA draft guidance of December 2024 gives "small deviations from protocol-specified visit windows" as an example of deviations that are not important, and a "missed scheduled visit" as a deviation at participant level. Importance depends on what the visit measures: the same guidance lists "failure to collect data to evaluate important study endpoints" among important deviations. Under ICH E6(R3), the sponsor sets the trial-specific criteria for classifying a deviation as important (section 3.9.3), so a visit two days outside its window may be minor at Week 2 and important at the primary endpoint visit.
Record the deviation where the CRA will look for it: the visit date in the eCRF, the planned window, the number of days outside, the reason given by the participant or the site, and the assessments done or not done. For drug trials, the FDA draft guidance recommends that investigators report all protocol deviations to the sponsor, with a procedure that highlights the important ones.
[[cta]]
How Datacapt keeps visits inside their windows
In Datacapt Calendar & Scheduling, your team sets the window each visit allows and generates the visit schedule from the time windows in the protocol. Datacapt dates every visit from the participant's reference date, blocks a booking that falls outside the window, and flags a visit that drifts before it turns into a protocol deviation. Participants book and reschedule their own visits online, inside the window the protocol allows, and receive email and SMS reminders in their language.
A missed visit shows on the site calendar and on the coordinator's dashboard the same day, so the coordinator can rebook it while the window is still open. The visit report, one of the EDC features, sorts each participant's visits into done, planned, out of window and missed. A booked visit opens the right eCRF forms, and marking it complete releases the participant's payment, so the visit date, the data and the payment come from one record. See how Datacapt handles scheduling visits within protocol windows.
Test the windows before the first participant
Calculating a clinical trial visit window takes one line of arithmetic. The risk sits in the inputs: a protocol that leaves the reference date, the numbering or the type of days open lets each site fill the gap its own way. Before the site initiation visit, take the Schedule of Activities and compute the dates for one fictitious participant randomized on a Friday before a public holiday. Have the site coordinator and the CRA calculate the same dates, then compare them with the dates the scheduling tool generates. If one set differs, correct the protocol wording or the tool configuration before the first participant is randomized.
[[faq]]
Sources
- International Council for Harmonisation, Clinical Electronic Structured Harmonised Protocol (CeSHarP): M11 Template, 2025
- International Council for Harmonisation, Guideline for Good Clinical Practice E6(R3), 2025
- International Council for Harmonisation, ICH E3 Guideline: Structure and Content of Clinical Study Reports, Questions & Answers (R1), 2012
- U.S. Food and Drug Administration, Protocol Deviations for Clinical Investigations of Drugs, Biological Products, and Devices, draft guidance, 2024
- Getz K, Smith Z, Jain A, Krauss R (Tufts Center for the Study of Drug Development), Benchmarking Protocol Deviations and Their Variation by Major Disease Categories, Therapeutic Innovation & Regulatory Science, 2022
What is a visit window in a clinical trial?
A visit window in a clinical trial is the range of days around a visit's target day in which the visit still counts as on schedule, for example Day 29 ± 3, from Day 26 to Day 32. The protocol sets the reference date, the target day and the tolerance. A visit outside the range is a protocol deviation.
How do you calculate a clinical trial visit window?
Start from the reference date, such as randomization. If it is Day 1, the target date is that date plus the target day minus one; if it is Day 0, add the target day. Subtract the tolerance before and add the tolerance after to get the earliest and latest dates. Count calendar days unless the protocol says otherwise.
Is an out-of-window visit a protocol deviation?
Yes. The ICH E3 questions and answers define a protocol deviation as any departure from the study design or procedures defined in the protocol, and the protocol defines each visit window. The FDA's December 2024 draft guidance cites small deviations from visit windows as deviations that are not important. The sponsor's criteria decide when a late visit becomes important.
Should visit windows use calendar days or business days?
The protocol decides, and it should state the choice in writing. Calendar days give every site the same dates. Business days widen the window and depend on each site's holidays: ± 3 business days around Monday, March 29, 2027, run from March 24 to April 1, nine calendar days instead of seven.
Does a late visit change the dates of the following visits?
It depends on the anchor the protocol sets. When every visit counts from Day 1, a late Week 4 visit leaves Week 8 on its planned date. When each visit counts from the previous one, every delay moves the visits that follow: a participant two days late at each visit reaches Week 12 eight days after Day 85.


With over 10 years of experience working in CROs, Khalil brings deep expertise in clinical trials and a clear understanding of the daily challenges faced by research professionals. His insights are grounded in real-world operations, making his perspective both practical and strategic.
Blog & News Datacapt
News, Articles, Resources et Tutorials.
