Trial Adherence Is Low: How to Read the Treatment Actually Received

Short answer: check who was analyzed, and in which group

When trial adherence is low, check who was analyzed and which group each person was counted in. An analysis by assigned group answers what happened to people offered a treatment. It does not answer what happens when treatment is taken exactly as directed. Those are different questions.

By ClinicalStudyConnect.com Research Desk. Last updated: September 26, 2026.

This guide is general education about reading study reports. It is not medical advice, and it cannot tell you whether any treatment is right for you.

Process map: from assignment to analysis

1. Participants are randomly assigned (randomized) to a trial group.
2. Some participants receive the assigned treatment as planned. Others take less of it, stop early, switch, or never start.
3. Some participants drop out, meaning they leave the trial or stop being followed.
4. Researchers choose an analysis population, the set of participants whose data go into a given result.

Steps 2 to 4 are where a low-adherence trial can be read well or badly. (The plain-language descriptions of assignment and dropout here are this desk’s own explanations, not quotations.)

What does “assignment” mean, and why does it matter?

Assignment is the group a participant was randomly placed in, such as treatment or comparison. It is not a record of the treatment actually received.

The combined SPIRIT-CONSORT guidance on the CONSORT-SPIRIT site explains why this matters: “To preserve the benefits of randomisation, all randomised participants need to be included in the analysis and retained in the group to which they were allocated.” An analysis built this way is called an intention-to-treat analysis. It counts participants in the group they were originally assigned to, whether or not they followed the plan.

CONSORT is the reporting guideline for trial results. SPIRIT is the matching guideline for trial protocols.

What is the difference between intention-to-treat and per-protocol?

• Intention-to-treat (ITT): participants are analyzed in the group they were originally assigned to.
• Per protocol (PP): in the guidance’s words, “all randomised participants who received their allocated trial treatment without major protocol deviations.”

The guidance also warns about leaving non-adherers out: “Exclusion of data from protocol non-adherers can compromise the randomisation and introduce bias, particularly if the frequency of and the reasons for non-adherence vary between the trial groups.”

In plain terms, people who follow a treatment plan may differ from people who do not, so a per-protocol result can differ for reasons unrelated to the treatment. Neither analysis is automatically “the truth.” Each answers a different question, and a report should say which one it shows.

How do dropouts fit in?

Dropout and low adherence overlap but are not the same thing. A person can stay in the trial and take little of the treatment. A person can also leave the trial after taking every dose. The CONSORT site says the flow diagram “displays the progress of all participants through the trial.” Reading it shows how many people were assigned, followed, and analyzed in each group.

Adherence context sheet

Use this checklist while reading a trial paper or results record. Print it and keep it beside the paper if that helps.

• Groups: How many people were assigned to each group?
• Flow: How many in each group were followed to the end? Are reasons for leaving given?
• Balance: Did dropout or non-adherence look similar across groups, or very different?
• Analysis population: Is the main result labeled intention-to-treat, per protocol, or something else?
• Group kept: Were participants analyzed in the group they were assigned to?
• Treatment actually received: Does the report say how much treatment people took, and how that was measured?
• Missing data: Does it say how missing results were handled?
• Stated limits: Do the authors say how low adherence may affect the result?

Conditional next step

1. If the report shows only one analysis population, look for the other in the tables or supplement. If it is not there, treat any claim about “the treatment as taken” as unsupported.
2. If adherence or dropout differs sharply between groups, read comparisons cautiously. The guidance names this as a bias risk.
3. If the report gives no adherence information, treat the result as incomplete. Do not fill the gap with assumptions.

What this guide cannot tell you

Low adherence does not by itself prove a treatment works or fails. It also does not tell you how large any effect would be if people took the treatment as directed. Those answers depend on the specific trial, its design, and its reporting. This guide does not rely on any single trial’s data.

If you are thinking about a treatment decision, talk with your own clinician or pharmacist. Do not start, stop, or change any medicine based on a summary like this. For an emergency, contact local emergency services.

Sources and editorial note

• CONSORT-SPIRIT: reporting guidelines for randomized trials and protocols, including the item on analysis populations (item 27b).
• ClinicalTrials.gov, a public registry of clinical studies where results records may be posted.

Clinical Study Connect is an independent educational publication. It is not a clinic, agency, or medical provider, and no credentialed reviewer is implied. Nothing here promotes a product, service, or treatment.