Demonstration article

This sample was written to show how Northstar Briefing presents journalism. It does not report real events and contains no real quotations, data or named sources.

Many illnesses improve on their own, and people who expect to feel better often do. Doctors, too, can see what they hope to see. These facts make it surprisingly difficult to know whether a treatment works. The modern clinical trial is a set of methods built to remove those confounding influences.

The building blocks

The central tool is randomisation. Participants are assigned by chance to receive either the treatment being tested or a comparison — a placebo or an existing standard treatment. Because assignment is random, differences between the groups at the end are more likely to reflect the treatment than pre-existing differences between the people in them.

Blinding adds another layer of protection. In a single-blind trial, participants do not know which group they are in; in a double-blind trial, neither do the researchers assessing them. This prevents expectations from shaping either how patients report symptoms or how clinicians judge outcomes. Blinding is not always possible — surgery and physical therapies are hard to disguise — and in those cases researchers often ask independent assessors who do not know the group assignments to evaluate results.

A well-designed trial is an argument against our own wishful thinking.

  1. Pre-clinical: laboratory and animal studies assess safety and plausibility.
  2. Early phase: small groups test safety, dosing and how the body handles the drug.
  3. Middle phase: larger groups give early evidence on effectiveness and side effects.
  4. Late phase: large randomised trials compare the treatment with alternatives.
  5. After approval: monitoring continues to detect rare or long-term effects.

Choosing what to measure

Trials must specify in advance what counts as success, known as the primary endpoint. Ideally this is an outcome that matters to patients, such as survival or relief of symptoms. Because those can take a long time to observe, trials sometimes use surrogate endpoints — measurements such as blood pressure or tumour size that are expected to predict real outcomes. Surrogates speed up research, but they do not always translate into genuine benefits.

Pre-registration of trial designs and endpoints in public databases helps prevent researchers from shifting goalposts after seeing the data. Independent monitoring committees review accumulating results and can stop a trial early if a treatment proves clearly harmful or clearly beneficial.

Trials also have limits. Participants may be younger or healthier than the patients who will eventually use a treatment, so results may not generalise. Rare side effects may appear only after many more people have taken a drug. That is why evidence continues to accumulate after approval, and why regulators and doctors weigh trial results alongside other sources of information.