Randomized trials on populations measure average effects in heterogeneous groups. N-of-1 trials measure what actually happens to one specific person. For individual decision-making, the latter is usually more relevant.
DoOperator Research · May 26, 2026
Decision takeaway
Randomized trials on populations measure average effects in heterogeneous groups. N-of-1 trials measure what actually happens to one specific person. For individual decision-making, the latter is usually more relevant.
Randomized trials on populations measure average effects in heterogeneous groups. N-of-1 trials measure what actually happens to one specific person. For individual decision-making, the latter is usually more relevant — and almost never used.
The gold standard in clinical evidence is the randomized controlled trial: randomly assign participants to treatment or control, measure outcomes, compare. This design, properly executed, eliminates confounding and produces an unbiased estimate of the average treatment effect in the study population. It is a genuine intellectual achievement, and the evidence it produces is more reliable than the alternatives.
But there is something the RCT cannot tell you, by construction: what effect the treatment has on you.
When a drug trial reports that a treatment reduces blood pressure by 8 mmHg on average, that number summarizes what happened across everyone in the trial. It does not mean that every participant experienced exactly an 8-point reduction. Some may have had no response. Some may have had a 20-point reduction. Some may have experienced an increase. The average is a summary statistic over a distribution of individual responses.
For a population-level intervention — a public health campaign, a policy change, a drug approved for mass prescription — the average effect is the right quantity to care about. You cannot tailor the campaign to every individual. You need to know the aggregate.
For an individual making a personal decision, the average is often the wrong quantity. You are not the average person in the trial sample. Your age, genetics, baseline physiology, diet, sleep patterns, stress levels, and the specific way you take or implement the intervention may all shift your individual response away from the population mean. The relevant question is not what happens on average — it is what happens to you.
An N-of-1 trial is a randomized experiment conducted on a single participant. The participant receives the treatment and control conditions in randomized order over multiple periods. Outcomes are measured consistently throughout. The analysis compares treatment periods to control periods for that individual, controlling for time trends and carryover effects.
This design produces evidence about individual response rather than population average. It cannot be generalized to other people, but it does not need to be. Its purpose is to answer the question "does this work for me?" rather than "does this work on average?"
N-of-1 trials have been used in clinical medicine for decades, primarily for conditions where individual response is highly variable: pain management, ADHD, insomnia, dietary interventions. Systematic reviews of N-of-1 trials have found that they frequently reveal that individual patients respond differently from what population averages would predict — in both directions. Some patients who would have been prescribed a treatment based on population evidence turn out not to benefit from it. Others who would have been denied a treatment do benefit.
N-of-1 trials have real limitations. They require an outcome that can be measured repeatedly and consistently. They require an intervention that can be turned on and off — chronic treatments with permanent effects cannot be studied this way. They require enough repeated measurements to distinguish treatment effects from noise. And they require the individual experimenter to have enough methodological care to avoid the systematic biases that undermine self-reported evidence: confirmation bias, regression to the mean, placebo effects.
These constraints are real. They rule out N-of-1 trials for many clinically important questions. But for the behavioral, nutritional, and lifestyle interventions that most people are actually trying to evaluate — sleep schedules, dietary changes, supplementation, exercise protocols, cognitive practices — the constraints are manageable. The interventions can be reversed. The outcomes are measurable at the individual level. The time horizon is short enough for repeated cycles.
Population RCTs and N-of-1 trials answer different questions. Population evidence tells you what is likely to work on average in people like you. N-of-1 evidence tells you what actually worked for you under the conditions of the experiment.
Neither is sufficient alone. Population evidence without individual experimentation leads to applying average-effect recommendations regardless of personal response. Individual experimentation without population evidence leads to unstructured self-help with no prior on what to expect or what has been tried.
The combination — using population evidence to generate prior beliefs about plausible interventions, then using careful individual experimentation to test whether those interventions work in your specific case — is more powerful than either approach alone. It is what evidence-based personal decision-making actually looks like when done rigorously.
The cultural gap is that population evidence is extensively documented, peer-reviewed, and widely reported, while individual experimentation is usually unsystematic, uncontrolled, and invisible. Closing that gap is not a technical problem. The methods for rigorous N-of-1 research are well established. It is a matter of applying them.