Enter the anchor date, then each visit’s offset in days, weeks or months and its window. Get the target date, the first and last in-window date for every visit, and a check of any actual date you enter.
Free sandbox · No credit card · 21 CFR Part 11 aligned
| Assessment | Wk 2 | Wk 4 | Wk 8 | Wk 12 | Mo 6 |
|---|---|---|---|---|---|
| Target date | 15 Mar | 29 Mar | 26 Apr | 24 May | 1 Sep |
| Window | ±3 d | ±3 d | ±5 d | ±7 d | ±2 wk |
| Actual (demo) | 16 Mar | 31 Mar | 3 May | ||
| Check | in | in | late |
Demo data. Week 8 was 2 days after the window closed on 1 May.
Visit windows in brief
Free tool
The demo is a 12-month study anchored on 1 March 2027. Edit the anchor, add or remove visits, and type an actual date to see it checked. Nothing is saved or sent.
| Visit | Offset | Window − / + | Target | Window | Actual date | Check | |
|---|---|---|---|---|---|---|---|
/ | Mon, 1 Mar 2027 | Mon, 1 Mar 2027 to Mon, 1 Mar 2027 | In window (on target) | ||||
/ | Mon, 15 Mar 2027 | Fri, 12 Mar 2027 to Thu, 18 Mar 2027 | In window (+1 d) | ||||
/ | Mon, 29 Mar 2027 | Fri, 26 Mar 2027 to Thu, 1 Apr 2027 | In window (+2 d) | ||||
/ | Mon, 26 Apr 2027 | Wed, 21 Apr 2027 to Sat, 1 May 2027 | Late by 2 days | ||||
/ | Mon, 24 May 2027 | Mon, 17 May 2027 to Mon, 31 May 2027 | No actual date | ||||
/ | Wed, 1 Sept 2027 | Wed, 18 Aug 2027 to Wed, 15 Sept 2027 | No actual date | ||||
/ | Wed, 1 Mar 2028 | Wed, 16 Feb 2028 to Wed, 15 Mar 2028 | No actual date |
Planning aid, not validated software. Target date = anchor plus the offset: days are calendar days (with Day 1 numbering, Day 29 is the anchor plus 28 days), a week is 7 days, and a month is a calendar month with the day moved back to the last day when the month is shorter (31 Aug plus 6 months is 28 Feb). The window runs from the target minus the first number to the target plus the second, both ends included. Whether the anchor, the window and the day numbering match the protocol is for you to confirm. Dates are computed in your browser; nothing is stored or sent.
The method
Target date. For a day-based offset the target is the anchor plus that many days (minus one if you number the anchor Day 1). For weeks it is the anchor plus 7 times the number of weeks. For months it moves the calendar month and keeps the day of the month, moving back to the last day of the month when the target month is shorter. Week 4 from Monday 1 March 2027 is Monday 29 March. Month 6 is Wednesday 1 September 2027.
Window. The window opens at the target minus the first number and closes at the target plus the second, with both ends included. You can enter the allowance in days or weeks. Week 8 in the demo has a target of Monday 26 April and a window of ±5 days, so it runs from Wednesday 21 April to Saturday 1 May.
Check. An actual date before the opening is early by the number of days before it, and after the closing is late by the number of days after it. A date inside the window shows how far it is from the target. In the demo the week 8 visit on Monday 3 May is late by 2 days, whereas week 4 on 31 March is 2 days after target and in window.
Take an anchor of 1 March 2027, a week 2 visit and a ±3 day window. The target is 15 March, the window runs 12 to 18 March, and a visit on the 19th is late by 1 day. Now enter the offset as 14 days and switch to Day 1 numbering: the offset is read as Day 14, which is 13 days after Day 1, so the target becomes 14 March instead of 15 March. That one-day shift is the most common source of disagreement between a schedule of assessments and a spreadsheet.
Conventions to settle
| Choice | Options | Effect | What to do |
|---|---|---|---|
| Anchor | Baseline, randomisation, first dose, previous visit | Moves every date, and decides whether delays accumulate | State it in the schedule of assessments |
| Day numbering | Anchor is Day 0 or Day 1 | One day, on every day-based offset | Use the protocol’s definition and show an example |
| Week and month length | 7 days, calendar month, or 28 or 30 days | Up to 3 days at 12 months | Define it; avoid “monthly” without a rule |
| Window shape | Symmetric or asymmetric (for example −2 / +7) | Which side deviations fall on | Match the biology: dosing visits often allow late, not early |
| Window anchor drift | Relative to baseline or to the previous visit | Late visits can shift later ones | Relative to baseline keeps the schedule fixed |
The tool anchors every visit to one date. If your protocol resets the window from the previous visit, work out each target from the actual date of the previous visit by running the tool again with that date as the anchor.
Build the visit schedule in the free sandbox, set windows around each visit and see late visits flagged as they happen.
Choosing a window
Narrow windows protect interpretability and wide windows protect feasibility. Use the endpoint to decide. The primary endpoint visit needs the tightest window, because a week 12 measurement taken at week 15 is a different measurement. Safety visits that follow a dose need windows tied to the drug’s pharmacokinetics. Routine follow-up visits can be wider, particularly for participants who travel. A ±3 day window around a 2-week visit is common, ±7 days at week 12 is also common, and ±2 weeks at six months or later is typical, but these are conventions, not rules. Take the numbers from your protocol and the biology.
Windows interact with the analysis. If a visit outside the window is excluded from the per-protocol set, the window defines who is analysed, and the sample size needs to allow for it. The schedule of assessments builder is a good companion for laying out which procedures happen at which visit, and the clinical trial timeline calculator turns the longest follow-up into a study end date. For patient-completed diaries, window rules apply per task: see reminders and completion windows.
In Capture
In Capture you define visit windows in the schedule, so out-of-window visits are visible when they happen rather than at database lock. The same builder holds the forms for each visit, so the schedule and the data collection stay in one place.
Window -3 / +3 days around the anchored target date
Soft window
Stays available with a protocol deviation flag
Hard window
Locks outside the window
Warning only
Stays available with a warning
Auto-close
Marked missed when the window closes
Before you lock the schedule
One named anchor event for each window, written in the protocol.
Day 0 or Day 1 stated, with a worked example in the schedule of assessments.
A week is 7 days. A month is either a calendar month or a fixed number of days, stated.
The visit that gives the primary measurement has the narrowest window.
Check how many targets fall on a weekend and whether the window lets the site book a weekday.
What counts as a deviation, which class, and whether data from the visit are used.
It is the range of dates around the target date of a scheduled visit in which the visit still counts as on schedule, usually written as plus or minus a number of days from the target, for example week 4 ± 3 days.
Add the visit’s offset to the anchor date to get the target date, then subtract the allowed days before and add the allowed days after. Both ends are normally included. The anchor and the day numbering must match the protocol.
If the anchor is Day 0, Day 28 is 28 days after it. If the anchor is Day 1, Day 29 is 28 days after it. The tool has a switch because mixing the two shifts every day-based visit by one day.
It depends on the protocol. A calendar month keeps the day of the month, and moves to the last day when the month is shorter. Some protocols use 28 or 30 days instead. Define it, because the difference reaches several days at 12 months.
It is a protocol deviation. Whether it is an important deviation depends on the visit and the data it provides, as set out in the protocol deviation plan. The data are not necessarily discarded, but the analysis plan should say how they are handled.
Yes. Enter different numbers for before and after, for example 2 days before and 7 days after, which suits visits where late is more acceptable than early.
No. All dates are calculated in your browser and nothing you enter is saved or transmitted.
Keep exploring
Schedule of assessments builder
Lay out procedures by visit.
Protocol deviation severity classifier
Classify an out-of-window visit.
Clinical trial timeline calculator
From follow-up length to study end.
Reminders and completion windows
Window modes for participant tasks.
Protocol to study setup
Draft the schedule from the protocol.
Set up visits, windows and forms in a free sandbox with every feature. No credit card, pay only when you go live.