Enter the target, the number of sites, the per-site rate, how fast sites open and ramp up, and your screen-failure and dropout rates. Get the number to consent, a projected last-patient-in date in slow, base and fast cases, and the number of sites you would need to hit a date.
Free sandbox · No credit card · 21 CFR Part 11 aligned
Consent / screen
179
Randomise
134
Completers
120
Last participant in
30 Nov 2028
Dec 2029
Nov 2028
Jul 2028
Forecasting enrollment in brief
Free tool
Change any assumption and the funnel, the three scenarios and the sites-needed answer update. Dates are calendar dates computed in your browser. Nothing is saved or sent.
| Scenario | Rate / site / month | Last participant in |
|---|---|---|
| Slow | 0.72 | Thu, 13 Dec 2029 |
| Base | 1.20 | Thu, 30 Nov 2028 |
| Fast | 1.56 | Sat, 22 Jul 2028 |
To reach the target by Wed, 1 Mar 2028 in the base case you need about 16 sites (you entered 8).
Planning aid, not validated software. Assumptions: the first site opens on the date entered and the other sites open evenly spaced until the activation window ends; each site ramps linearly from zero to its full rate over the ramp period and then recruits at a constant rate; the rate is consented participants per open site per month; screen failures and dropouts are fixed percentages. Completers needed ÷ (1 − dropout) gives the number to randomise, and that ÷ (1 − screen failure) gives the number to consent. Real enrolment is lumpy, seasonal and usually slower than sites predict, so use the slow case for commitments. Dates are calendar dates computed in your browser; nothing is stored or sent.
The method
Step 1, the funnel. Suppose the protocol needs 120 participants to complete the study. With 10% dropout after randomisation you must randomise 120 ÷ 0.90 = 133.3, so 134. With 25% screen failure you must consent 134 ÷ 0.75 = 178.7, so 179. If your target is already a randomised number (many sponsors count enrolment at randomisation), switch the tool to "Randomised" and the dropout step is skipped. The dropout and screen-failure rates themselves are inputs, not outputs: take them from comparable trials or from your own pre-screening log, and see the dropout-adjusted sample size calculator if you are still sizing the study.
Step 2, sites open over time. The first site opens on the date you enter and the others are spaced evenly until the end of the activation window. In the demo, eight sites open over 12 weeks, about 1.7 weeks apart. A site does not recruit at full speed on day one: it needs to train staff, finish its first consents and find its rhythm, so the tool ramps each site linearly from zero to its full rate over the ramp-up period (two months by default).
Step 3, the daily simulation. Every day the tool adds up the current rate of all open sites, converted from a monthly figure (rate × 12 ÷ 365.25 per day), and accumulates consented participants. The first day the running total reaches the number to consent is the projected last-patient-in date. First randomised is the first day that total times (1 − screen failure) reaches one participant.
Step 4, scenarios and the reverse question. Slow and fast cases multiply the per-site rate by the percentages you choose (60% and 130% by default) and rerun the simulation. If you also enter a date you need last patient in by, the tool searches for the smallest number of sites that reaches the target by that date in the base case, using the same activation window and ramp-up.
Take one site, no ramp-up, no screen failure and a rate of 1 consented participant per month, and a target of 12. The site produces 12 × 365.25 ÷ 12 = 365.25 days of recruitment at one per 30.44 days, so the twelfth participant arrives on day 366. Add a second identical site and the date halves to day 183. The tool returns exactly those numbers, which is a quick sanity test of any enrolment projection you build in a spreadsheet.
What moves the date
Demo trial: start 1 Mar 2027, 120 completers, 8 sites at 1.2 consented per site per month, 25% screen failure, 10% dropout, 12-week activation window, 2-month ramp-up. Only the listed assumption changes.
| Change | To consent | Last participant in | Months from start |
|---|---|---|---|
| Base case | 179 | 30 Nov 2028 | 21.0 |
| Rate 0.72 per site per month (60%) | 179 | 13 Dec 2029 | 33.4 |
| Rate 1.56 per site per month (130%) | 179 | 22 Jul 2028 | 16.7 |
| Screen failure 40% | 224 | 22 Apr 2029 | 25.7 |
| Dropout 20% | 200 | 5 Feb 2029 | 23.2 |
| 12 sites instead of 8 | 179 | 25 May 2028 | 14.8 |
| 16 sites instead of 8 | 179 | 20 Feb 2028 | 11.7 |
| Activation window 24 weeks | 179 | 11 Jan 2029 | 22.4 |
| Ramp-up 4 months | 179 | 30 Dec 2028 | 22.0 |
| No screen failure, no dropout | 120 | 27 May 2028 | 14.9 |
Calendar dates are model output, so a result can land on a weekend. Sites that open late hurt less than sites that recruit slowly: halving the rate is far more costly than doubling the activation window.
Assumptions you can defend
Per-site rate. Site questionnaires and feasibility surveys are optimistic. A better anchor is the registry record of a comparable completed trial: divide its enrolment by its number of sites and by the months between first and last participant in, and expect your own sites to land near or below that figure. Remember that the tool counts consented participants per open site per month, so a site that sees 4 eligible patients a month and consents a quarter of them is a rate of 1.0, not 4.
Screen failure. Count failures at the same point your enrolment target is counted. A trial with a long run-in or a central eligibility review fails more people after consent than one with a simple checklist. Past screening logs, even from a pilot, beat any rule of thumb. The screening and enrollment log template shows what to record so the rate is measurable by site and by reason, and pre-screening tools can reduce failures after consent by catching ineligible people earlier.
Dropout. Use the rate that matches your analysis population. Dropout after randomisation inflates the number to randomise only when the target is completers or evaluable participants. If you use a conservative estimate here, use the same figure in the statistical section of the protocol so the two documents agree.
Activation and ramp-up. Activation dates come from contracts, ethics or IRB approvals and drug supply, none of which a calculator can know. Use the date each site is realistically able to consent its first participant. If the schedule is a plan rather than a commitment, run the 24-week window as a second case.
| Input | Best source | Weak source | |
|---|---|---|---|
| Rate per site | Comparable completed trial on a registry | Feasibility questionnaire | |
| Screen failure | Own screening log | Rule of thumb | |
| Dropout | Same assumption as the protocol | Different number in each document | |
| Activation dates | Approvals and contracts in hand | Target dates from the plan |
Set up the study, add sites and record screening and randomisation in the free sandbox, then rerun this forecast with your actual monthly rates.
Which tool when
The clinical trial timeline calculator answers a different question: given start-up time, sites and a follow-up period, when do first participant in, database lock and topline results fall? It works on participants enrolled. This forecast stays inside the enrolment phase and goes deeper on it: it starts from completers, adds the screen-failure and dropout funnel, compares slow, base and fast cases, and solves backwards for the number of sites needed to meet a date. Use this page to defend the enrolment assumption, then carry the resulting last-patient-in date into the full timeline.
Two other pages help with the neighbouring steps. The clinical trial cost per patient calculator turns the number of participants you must consent into a budget line, because screen failures cost money even though they never reach the analysis. And the patient recruitment playbook for small trials covers what to do when the slow case is the realistic one.
When the forecast is bad
A high screen-failure rate is often a protocol or pre-screening problem. Tightening referral criteria costs less than opening four more sites.
Shortening the activation window helps modestly (see the table). Fixing the slowest approvals is cheaper than recruiting extra sites.
The tool assumes new sites perform like existing ones. Late-added sites usually do not, so treat the result as a minimum.
After three months, replace the planned rate with the observed one per site. A forecast that is never updated is just a plan.
A date range with stated assumptions holds up better in a progress report than a single date that slips.
Before you rely on it
Completers, evaluable or randomised, matching the protocol statistics section.
Screen failure and dropout from a log or comparable trial, source written down.
Consented participants per open site per month, not eligible patients seen.
Each site's expected first-consent date and the evidence for it.
At least 60% of the planned rate, reported next to the base case.
Re-forecast monthly using observed rates by site.
Convert the target into the number to consent using dropout and screen-failure rates, then project recruitment forward as the sum of all open sites' rates, with each site ramping up after it opens. The date the running total reaches the number to consent is the projected last-patient-in date.
Completers needed divided by (1 minus the dropout rate) gives the number to randomise. That divided by (1 minus the screen-failure rate) gives the number to consent. Both are rounded up. For 120 completers, 10% dropout and 25% screen failure that is 134 and then 179.
It depends on the condition and the design, so use a comparable completed trial on a registry, divided by sites and months, rather than a general figure. The tool counts consented participants per open site per month.
It reruns the simulation with the per-site rate multiplied by the percentage you set (60% by default). Showing it next to the base case makes the uncertainty visible to funders and sponsors.
No. It assumes a constant rate once a site has ramped up. If recruitment is seasonal, use a lower average rate, or run separate forecasts for the busy and quiet periods.
If you enter a deadline, the tool finds the smallest number of sites that reaches the target by that date in the base case, using the same activation window, ramp-up and rate. It assumes added sites perform like the planned ones.
No. The calculation runs in your browser and nothing you type is stored or transmitted.
The timeline calculator maps the whole study from start-up to topline. This page models the enrolment phase in more detail, with the screen-failure and dropout funnel, three scenarios and the sites-needed answer.
Keep exploring
Clinical trial timeline calculator
Start-up to topline.
Dropout-adjusted sample size
Inflate the sample for attrition.
Cost per patient calculator
Price the participants you must consent.
Pre-screening tools
Fewer failures after consent.
Screening and enrollment log
Measure screen failure by site.
Recruitment playbook for small trials
When the slow case is real.
Run the study in a free sandbox with every feature. No credit card, pay only when you go live.