N-of-1 studies
Chapter 14: Other Designs and Approaches to Know — Video 1
https://ghrbook.com/videos/n-of-1-studies/

[Slide 1] This chapter covers a grab bag of designs. They share little with each other, except that each one arose because the standard toolkit couldn't answer the question someone needed to ask. The aim is breadth: recognizing what each design is built to do, so you can spot one in a paper or know when to call someone who specializes in it. We'll start with N-of-1 studies.
[Slide 2] Here's a scenario you might face. You're working with a patient, let's call her Maria, who has a chronic condition that responds differently to treatments than the typical patient in clinical trials. The trial evidence says Treatment A should work better than Treatment B, but Maria seems to do worse on A. How do you figure out what's actually best for her? What works for this specific person?
[Slide 3] That's the question an N-of-1 study is built to answer. It's also called single-case design, or single-subject research. Instead of comparing outcomes across many people, we systematically compare conditions within a single individual over time, and the same person experiences both conditions in a structured sequence.
[Slide 4] The classic structure has four phases. We measure the outcome repeatedly with no intervention, then introduce the treatment and keep measuring, then remove it, and then apply it again. This design lets us see if the outcome changes when, and only when, the intervention is present. If Maria's symptoms improve every time Treatment A is introduced and worsen when it's removed, that's pretty good evidence that Treatment A is helping her, regardless of what the population-level trials showed.
[Slide 5] You might be thinking, isn't this just trying different things and seeing what happens? Yes, but the key difference is the systematic structure. A well-designed N-of-1 study has multiple measurement points in each phase, replicates the intervention effect across phases, randomizes when phases occur when that's possible, and blinds the outcome assessors. Each transition between phases becomes a mini-test of the hypothesis.
[Slide 6] In a randomized phase-order design, baseline doesn't always have to come first. We randomize the sequence of phases, which reduces threats to validity from temporal factors and history effects. If Maria's symptoms improve every time she's on Treatment A, whether A comes first or second, that's stronger evidence that the treatment is causing the improvement rather than, say, seasonal changes or the natural course of her disease.
[Slide 7] Another variation is the multiple baseline design, where we stagger when the intervention starts. Imagine three patients with similar chronic conditions. The first starts the intervention after two weeks, the second after four, and the third after six. If all three improve precisely when their intervention begins, and only then, we've replicated the effect three times at different time points. That's pretty compelling evidence, even with just three people.
[Slide 8] N-of-1 designs are most useful when the condition is rare enough that assembling a large trial would take years, when patients show substantial heterogeneity in treatment response and we need to optimize care for a specific individual, or when resources don't permit a full-scale trial but we still need rigorous evidence. The condition needs to be chronic and stable enough that changes in the outcome likely reflect the intervention and not natural disease progression, which rules out acute illnesses and rapidly evolving conditions.
[Slide 9] A study from Eldoret, Kenya used this design to test whether a brief intervention delivered by lay counselors could reduce harmful drinking and improve family functioning among fathers. No proven intervention existed locally, and the population was too dispersed to assemble easily into a conventional trial. Nine men ages 30 to 48 who screened positive for problem drinking were randomized to staggered start dates for a five-session intervention combining behavioral activation, motivational interviewing, and gender role content aimed at fathers. Each man served as his own control, with weekly alcohol consumption measured across baseline, the intervention, and follow-up.
[Slide 10] Eight men completed treatment, and the effects replicated across the staggered starts. The odds of any given day being alcohol-free were 5.1 times higher after treatment than at baseline, with a 95% confidence interval from 3.3 to 7.9. On the days men did drink after treatment, they drank half as much. Partners and children independently reported improvements in family functioning, and the men reported fewer depression symptoms, suggesting effects that extended beyond drinking itself.
[Slide 11] When nine separate baselines all show drops aligned with their individual treatment start dates, history or maturation are weak alternative explanations. The design produced individual-level evidence in a setting where a full randomized trial would have been impractical.
[Slide 12] N-of-1 designs can provide strong causal evidence at the individual level, which makes them ideal when large samples aren't feasible and when results need to apply directly to the person being studied. They can also be more ethical than assigning many patients to potentially inferior treatments when we only need to optimize care for one.
[Slide 13] The trade-offs are real. Results don't generalize to other individuals without replication. The design requires stable chronic conditions, because highly variable conditions make it impossible to attribute changes to the intervention. Carryover effects between phases can confound results if treatments have lingering effects, so we build in washout periods. And establishing a stable baseline before intervening takes time, which means starting the intervention too quickly undermines the entire design.