The value of one: what can we learn from case studies?
Despite their size, researchers say studies of one person can uncover results that would be lost in a large-scale trial.
SOMETIMES IT TAKES multitudes to reveal scientific truth. Researchers followed more than 7,000 subjects to show that a Mediterranean diet can lower the risk of heart disease. And the Women’s Health Initiative enlisted more than 160,000 women to show, among other findings, that postmenopausal hormone therapy put women at risk of breast cancer and stroke.
But meaningful, scientifically valid insights don’t always have to come from studies of large groups. A growing number of researchers around the world are taking a singular approach to pain, nutrition, psychology and other highly personal health issues. Instead of looking for trends in many people, they’re designing studies for one person at a time.
A study of one person — also called an N of 1 trial — can uncover subtle, important results that would be lost in a large-scale study, says geneticist Nicholas Schork of the Translational Genomics Research Institute in Phoenix. The results, he says, can be combined to provide insights for the population at large. But with N of 1 studies, the individual matters above all. “People differ at fundamental levels,” says Schork, who discussed the potential of N of 1 studies in a 2017 issue of the Annual Review of Nutrition. And the only way to understand individuals is to study them.
Case studies of individuals in odd circumstances have a long history in medical literature. But the concept of a clinical medicine N of 1 study gathering the same level of information as a large study goes back to an article published in the New England Journal of Medicine in 1986. Hundreds of N of 1 studies have been published since then, and the approach is gaining momentum, says Suzanne McDonald, N of 1 research coordinator at the University of Queensland in Brisbane, Australia.
McDonald says that N of 1 trials can seem meager and even trivial, at least at first glance, and especially compared with large, randomized control trials, considered the gold standard of research methods. But when they’re done properly, she says, one-person studies can have all of the statistical power and scientific rigor of studies involving hundreds or thousands of people. She notes that in 2011, the Oxford Center for Evidence-Based Medicine, which promotes reliable study methods, determined that N of 1 trials deserved a “level 1” rating — the highest level of evidence possible — for assessing treatments of individual patients.
“There has been a lot of misunderstanding in the research world,” McDonald says. “People think that N of 1 studies aren’t scientifically rigorous or that they’re just anecdotal. But we can measure someone over time repeatedly and very scientifically.”
When they’re done properly, one-person studies can have all of the statistical power and scientific rigour of studies involving hundreds or thousands of people.
As with most research studies, N of 1 studies gain their power through data points. But instead of taking a few measurements from many people, researchers can conduct many measurements from one person over time. A study might compare the results of two different interventions, perhaps two different pain medications or two different diets. Or it might compare an intervention to a placebo to see if a treatment had any real effect beyond wishful thinking. Throughout the process, the subject often has the power to change the approach and goal of the study to meet their own needs.
Combining those results can provide the same sort of insights as a multiple-person clinical trial. In 2018, a team of researchers in the Netherlands, the U.K., and the U.S. aggregated results from placebo-controlled N of 1 trials of 27 patients to show that the drug mexiletine effectively relieved myotonia (prolonged muscle contractions) in patients with a rare muscle disease. To validate the approach, the researchers compared the results to a more typical randomized control trial. The results were similar, adding a layer of confidence in the N of 1 results. The researchers noted that N of 1 studies could be especially helpful for testing treatments for rare conditions that, by definition, don’t lend themselves to large-scale studies involving many patients.
McDonald is the co-chair of the International Collaborative Network for N-of-1 Clinical Trials and Single-Case Experimental Designs, a community of about 200 researchers, clinicians, and members of the public who share study ideas and highlight recent results from around the world. For example, the network’s website recently posted a 2019 study from China, published in the journal Nutrition and Cancer, that combined results of N of 1 trials to provide preliminary evidence that amino acid supplements might boost the immune systems and reduce inflammation in patients with non-small-cell lung cancer.
Underscoring the diverse possibilities of N of 1 studies, McDonald is currently involved in investigations of insomnia in Parkinson’s disease and symptom fluctuations in chronic fatigue syndrome, among other things. In 2017, she and colleagues published a study that used a series of N of 1 investigations to see if predictors of physical activity changed after retirement. “For some people, it was tied to how much sleep they got the night before, and for others it was their mood,” she says.
Other potential studies are still under discussion. In the July 2019 issue of Frontiers in Nutrition, Spanish researchers made the case that N of 1 studies could be an effective way to investigate nutritional interventions for improving cognitive function and slowing decline in patients with dementia. As the authors note, creating a classic randomized control trial for this issue would be daunting: Even without the complication of dementia, it’s difficult to get people to adhere to diets they were assigned at random. But by using computer-assisted food diaries or perhaps blood tests to confirm dietary patterns, researchers could conduct cognitive tests and compare results to a baseline.
The authors acknowledge that the N of 1 approach would have limitations — limitations that would arise in just about any study of one person. Importantly, there wouldn’t be any sort of control group; the best researchers can do is track changes in cognitive skills over time. And any variations in diet between subjects would complicate attempts to combine results. Still, as more and more patients participate, the aggregated results would provide a more detailed picture of the link between food and cognitive function, which might not be possible with a more traditional study.
As Schork explains, the N of 1 approach appreciates and embraces the fact that what works for one person may not work for another. Differences in our genes, our habits, and even our microbial communities can influence the impact of all sorts of interventions and treatments. By studying individuals instead of groups, researchers can learn more about which quirks really matter for the big picture. If a large number of N of 1 studies showed, for instance, that people with certain genetic markers responded well to a particular diet, that diet could be recommended to others with that marker. “Finding the differences that matter versus the ones that we can ignore — that’s where all of the action is,” Schork says.
Some patients are using N of 1 studies to get a better understanding of their own conditions. Sara Riggare, a Ph.D. student in health informatics at the Karolinska Institute in Stockholm, designed a study in which she was the one and only patient. Riggare, who has Parkinson’s disease, tested the effects of nicotine on her levodopa-induced dyskinesia, a common and troubling side effect of Parkinson’s medication. Using an electronic cigarette, she received either a dose of nicotine or a pure water vapor and noted the immediate effects.
“Finding the differences that matter versus the ones that we can ignore — that’s where all of the action is.”
Riggare found that the treatment worked, at least in her case. “As soon as that nicotine reached my brain, I felt a sense of calm,” she says. Still, she doesn’t think every Parkinson’s patient should start using e-cigarettes, especially given the current health concerns about vaping. For her, the real take-home message was the approach, not the specific result. “I wanted to create a model for other patients to find out what works for them,” she says. “You can’t know how something works on an individual until you try it on an individual.”
In a larger example of self-experimentation, nearly 1,600 patients with type 1 diabetes are sharing results of their do-it-yourself “closed-loop” insulin pumps that are modified to allow the pumps to receive commands from their glucose monitors. The so-called OpenAPS systems haven’t been FDA approved, so they can’t be sold commercially. But — ideally with guidance and permission from their doctors — patients can go online for hardware and software checklists and step-by-step instructions. The patients then have the option to collect daily data and post the results for other users to see. A compilation of real-world results from these devices, presented on a poster at the annual meeting of the American Diabetes Association in 2016, showed that 18 patients who used the device over months were able to keep their blood glucose under control with no incidences of severe hypo- or hyperglycemia.
Not every attempt at N of 1 research has been a complete success. A 2018 study published in JAMA Internal Medicine compared N of 1 trials to standard care for a group of 215 people with chronic musculoskeletal pain. The 108 patients in the N of 1 group essentially ran their own studies. They chose two treatments for comparison — such as acetaminophen and acupuncture — and they also chose the study length. Subjects used smartphones or tablets to track their results. The other 107 subjects received usual care. Both groups reported better control of pain one year later, and the N of 1 group consistently fared better on a wide array of outcomes at different times (for example, pain intensity at six and 12 months). But by the study’s end, there was no significant difference in pain-related interference with daily activities.
Still, the study authors, led by Richard Kravitz, a physician at the University of California, Davis, saw the results as a partial vindication for the potential of N of 1 studies, not a setback. In a letter published in JAMA Internal Medicine in 2019, they wrote that widespread use of N of 1 trials could help people learn to improve their health through a better understanding of their diet, exercise, and medications. As a bonus, they would get a firsthand experience with the rigors of scientific research, which would improve their scientific literacy as well as their self-care.
Smartphone apps like those used in the JAMA Internal Medicine study should help open up a new world of N of 1 studies, says Christopher Schmid, a biostatistician at Brown University in Providence, Rhode Island, allowing individuals to do their own studies in collaboration with their physicians. “Advances are being made in the way that people collect their own data,” he says. “It brings research to people who don’t normally think about doing research.” Schmid is currently working with the Patient-Centered Outcomes Research Institute, a nonprofit that investigates medical treatments, on a project that uses smartphone apps to test different diets on children with inflammatory bowel diseases. “We’re looking to see which diet works best for each kid,” he says.
Individual differences can be especially important when it comes to nutrition. An Israeli study published in the journal Cell in 2015, for example, found “high interpersonal variability” in blood sugar responses to identical foods in 800 nondiabetic people consuming a total of more than 46,000 meals. The average blood sugar reading after eating a piece of bread, for example, was 44 milligrams of sugar per deciliter, but some people had less than 15 milligrams while others scored around 80.
Such wide variation underscores the need to study nutrition on an individual basis, Schork says. In the near term, he hopes to conduct nutritional studies on people with rare, poorly understood conditions that aren’t easily treated through medications — such as syndromes that predispose people to cancer. Roughly speaking, the studies would collect thousands of measurements while the subjects try different diets. All of those data points should add up to a clear picture of which foods would work best for their particular needs. Schork also has plans to investigate people with metabolic diseases that change the way they process nutrients.
Riggare has published one more study on herself — an exploration of self-tracking published in 2018 in the Journal of Parkinson’s Disease — since that nicotine experiment, and she continues to keep a close watch on her Parkinson’s and overall health. She uses phone apps and wearable sensors to help track her mood, symptoms, and physical activity. She encourages other patients to learn about themselves too, whether it’s through a formal N of 1 trial or informal experimentation with advice from a doctor. Large randomized studies may be the gold standard of science, but self-knowledge can be the most valuable insight of all.
by Chris Woolston
This article originally appeared in Knowable Magazine, an independent journalistic endeavor from Annual Reviews.