Participant Counts in Separate Two-Arm Trials

Counting Participants Across Separate Trials

When several treatments are assessed in separate two-arm trials, each trial has its own treatment group and its own control group. The overall recruitment target is the sum of the participants recruited to all those groups. The key is to count each group once and keep treatment and control totals distinct.

This is a recruitment-counting exercise, not a method for choosing the sample size needed to detect a specified treatment effect. Take each trial’s planned per-arm sample size as given. The earlier material on shared-control recruitment explains why a common control can reduce recruitment; here, the defining feature is that every separate trial has its own control group.

Definition: Suppose there are \(k\) separate two-arm trials. Trial \(i\) assigns \(n_i\) participants to its treatment arm and \(n_i\) participants to its own control arm. Its total is \(2n_i\). If every trial uses the same per-arm sample size \(n\), the total across all trials is \(2kn\): \(kn\) treatment participants and \(kn\) control participants.

“Equal-sized” here means equal numbers in the treatment and control arms of each trial. The same per-arm number \(n\) may also be used across all trials, but that is an additional condition. If the trials have different targets, use \(n_i\) for trial \(i\) and add their counts. In either case, the treatment and control totals are sums over the corresponding arms.

Why the Total Is a Sum

In a separate-trial setup, a participant in one trial is not also counted as a participant in another trial, and a control group belongs to only its own trial. Therefore, the total number recruited is the sum of all trial-specific arm counts. For trial \(i\), the two arms contain \(n_i+n_i\) participants. Summing those trial totals gives:

N_{\text{total}}=\sum_{i=1}^{k}(n_i+n_i) =2\sum_{i=1}^{k} n_i.

The treatment-arm total is \(\sum_{i=1}^{k}n_i\), and the control-arm total is also \(\sum_{i=1}^{k}n_i\). Thus, the overall total is twice either arm-type total. If all trials have the same per-arm target \(n\), then \(\sum_{i=1}^{k}n_i=kn\), giving \(N_{\text{total}}=2kn\).

This calculation assumes that the stated arm sizes are the planned numbers to recruit, each trial has exactly one treatment arm and one control arm, and the trial populations do not overlap for purposes of counting recruited participants. If the targets are analyzable sample sizes rather than recruitment targets, additional allowance may be needed for participants who do not provide usable outcome data. That allowance is not part of the basic formula.

Worked Example: Three Treatments in Three Separate Trials

Suppose three hypothetical treatments are each being assessed in a separate two-arm trial. Each trial plans to recruit 120 participants to its treatment arm and 120 to its own control arm. Find the treatment total, control total, and overall recruitment target.

Here \(k=3\) and \(n=120\) participants per arm. Each trial therefore recruits \(2(120)=240\) participants. Across all three trials:

N_{\text{total}}=2kn=2(3)(120)=720\text{ participants}.

There are \(3(120)=360\) treatment participants, one treatment group of 120 in each trial. There are also \(3(120)=360\) control participants, one separate control group of 120 in each trial. The check \(360+360=720\) agrees with the overall total.

The fact that there are three treatments does not mean that the total is three groups of 120. Each treatment is paired with its own 120-person control group, so there are six groups of 120 altogether.

When the Trial Targets Differ

Separate trials need not have identical recruitment targets. Equal allocation within each trial still means that its treatment and control arms have the same size; it does not require every trial to use the same size. In this case, calculate each trial’s two-arm total, then sum those totals. Keep a separate running sum for treatment arms and control arms to make the accounting transparent.

Worked Example: Adding Unequal Trial Targets

Four hypothetical independent trials have per-arm recruitment targets of 85, 100, 120, and 95 participants. Each trial has equal-sized treatment and control arms. Calculate the total number of treatment participants, control participants, and participants across the four trials.

The sum of the per-arm targets is:

85+100+120+95=400\text{ participants}.

Each trial’s treatment arm contributes its per-arm target, so the treatment total is 400 participants. The control arms have the same respective sizes, so the control total is also 400. The total recruitment target is:

N_{\text{total}}=2(85+100+120+95) =2(400)=800\text{ participants}.

As a trial-by-trial check, the totals are \(170+200+240+190=800\) participants. The equal treatment and control totals follow from equal allocation within every trial, even though the four trial sizes differ.

Checking a Recruitment Plan

A useful check is to write down the number of trials, the per-arm target for each trial, and the resulting counts before adding them. This makes visible whether each treatment has its own control and whether both arms have been included. The following steps are a practical way to audit the arithmetic.

List the trials and targets.
For each trial \(i\), record \(n_i\), the planned number of participants in each arm.

Count both arms.
Calculate \(2n_i\) for each trial, or record \(n_i\) once in the treatment column and once in the control column.

Add by arm type and overall.
Sum the treatment targets, sum the control targets, and add those two totals to obtain \(N_{\text{total}}\).

Check the structure.
Confirm that every separate trial has exactly one treatment and one control group, and that no participant or group has been counted twice or omitted.

Worked Example: Auditing a Stated Total

A planning note for two hypothetical separate trials says that each trial will recruit 75 participants per arm and that the combined target is 225 participants. Check the stated total and identify the correct treatment and control counts.

Each trial has two arms, so its total is \(2(75)=150\) participants. With two separate trials, the overall total is:

N_{\text{total}}=2(2)(75)=300\text{ participants}.

The treatment total is \(2(75)=150\) participants, and the control total is also 150 participants. Their sum is \(150+150=300\). The stated 225 is too small because it counts only three groups of 75, whereas the two separate trials require four groups: two treatment groups and two control groups.

What This Count Does—and Does Not—Compare

The formula \(2\sum_i n_i\) applies when each trial recruits its own control group. It is not a formula for a design in which several treatment arms use a single common control group: in that arrangement, the control participants belong to one group rather than being recruited afresh for every treatment comparison. The distinction is structural. Before calculating, determine whether the design has separate trials with separate controls or a common control group.

For this participant-counting task, the separate-trial calculation is exact once the arm targets and trial structure are specified. It does not by itself establish that the trials are feasible, that their sample sizes are adequate, or that all participants will provide usable data. Those questions require information beyond the group counts.

Common Mistakes and Limitations

Counting only the treatment groups. Three treatment arms of 120 participants are not 360 total if each is in a separate two-arm trial. The three own-control groups add another 360, making 720 overall.
Counting one control group for several separate trials. That changes the design being counted. In separate trials, each trial contributes its own control-arm target.
Assuming “equal-sized” means every trial has the same size. Equal allocation requires equal arm sizes within a trial. If trial sizes differ, use each \(n_i\) in the sum rather than forcing one common value.
Confusing participants per arm with participants per trial. If a target is 120 per arm, a two-arm trial recruits 240, not 120.
Mixing recruitment targets with analyzable sample sizes. A target for participants with usable outcomes may be smaller than the number that must be recruited if some outcomes are expected to be missing. The basic count does not apply an attrition adjustment automatically.

A final limitation is that the total count says nothing about how participants are distributed across sites, time periods, or other strata. It also does not answer whether the arms are randomized or whether the trials share other operational resources. Those details do not change the arithmetic above unless they change which people belong to which trial-specific groups.

Key takeaway: In \(k\) separate equal-allocation two-arm trials with per-arm targets \(n_1,\ldots,n_k\), the treatment total is \(\sum_i n_i\), the control total is \(\sum_i n_i\), and the combined target is \(2\sum_i n_i\). If every per-arm target is \(n\), this becomes \(2kn\).

Check Your Understanding

Answer each question, then compare your reasoning with the solutions.

Three separate trials each plan for 64 participants per arm. What are the treatment total, control total, and combined total?
Two separate trials have per-arm targets of 110 and 140 participants. What is the overall target, and how many participants are in each arm type combined?
A planner says that four separate two-arm trials with 50 participants per arm need 400 participants total. Is the count correct? Explain.
In a set of three separate trials, the control-arm targets are 70, 90, and 100. Each trial has equal-sized arms. Find the treatment total and the overall total.
Why should the count for separate trials not use one control group total for all treatments?

Solutions. (1) Treatment total \(=3(64)=192\), control total \(=192\), and combined total \(=192+192=384\) participants. (2) The per-arm targets sum to \(110+140=250\), so there are 250 treatment participants and 250 control participants, for \(2(250)=500\) participants overall. (3) The count is correct: there are \(4\) trials, each with \(2(50)=100\) participants, giving \(4(100)=400\). The treatment and control totals are each \(4(50)=200\). (4) Equal-sized arms mean the treatment targets are also 70, 90, and 100; their sum is 260. The overall total is \(2(260)=520\) participants. (5) Each separate trial has its own control arm. Using a single control total would count a different design, not the specified collection of separate two-arm trials.