M. de Groot, M. Drangsholt, F. Martín-Sánchez
2017
Citations
0
Influential Citations
18
Citations
Quality indicators
Journal
Methods of Information in Medicine
Abstract
Single subject research design, also known as N-of-1 research, is a scientific method in which an individual person serves as the research subject. We treat “N-of-1” and “single subject” as synonyms encompassing all scientific practice which focuses on observations made about a single person. Other names for similar and overlapping approaches include: single case experiments [1–3] single case research [4, 5], single case designs [6], and single patient trials [7]. Some authors distinguish between single subject research in general, which may be descriptive and exploratory in character, and single subject experiments that are prospectively planned and use formal methods such as randomization, blinding, or crossover comparisons. Here, we use N-of-1 and single subject research as synonymous, high level general terms for research focused on an individual rather than a group. N-of-1 research is common in applied fields of psychology, education, and human behavior where it has benefited from extensive methodical research and practical guidance for practitioners [8, 9]. However, over a half-century of study and advocacy, including pioneering publications by Guyatt et al., Larson et al., Mahon et al., and others, have failed to establish single subject science as central to research and practice in medicine [10–13]. A systematic review of 122 eligible N-of-1 studies published between 1985 and 2013 showed wide variation in methodology and reporting, reducing the power of these studies to influence practice [14]. Researchers advocating N-of-1 techniques have noted that the practical obstacles to design, conduct, analyze and apply the results for single subjects have simply been too high [15, 16]. Nevertheless the rise of personalized medicine and patient-centered research create new opportunities for using N-of-1 methods [17, 18]. Recent key publications include an extensive and comprehensive user guide for the design and implementation of N-of-1 trials [19], an update of the standard (CONSORT) for reporting N-of-1 trials [20, 21], and a special issue of the Journal of Clinical Epidemiology devoted to individual patients as the primary source and target of clinical research [22]. General public interest in gathering data about health is also growing. A Pew Internet study conducted in 2013 found that 1 in 5 Americans use some form of technology to track their health [23]. In 2016, the number of consumers in the United States who use mobile health apps increased from 16 percent in 2014 to 33 percent and the number of consumers who use health wearables increased from 9 percent to 21 percent [24]. According to data from the International Data Corporation (IDC), 104.3 million wearable devices were shipped in 2016, a number that is likely to be almost doubled by 2021 [25]. The increasing availability of home blood testing kits, wearable glucose monitors, and heart rate monitors, among other consumer health tools and services, suggest a large scale transformation of the measurement context for N-of-1 research. The combination of increased public interest and reliable measurement technologies broadly available may reduce the barriers to application of N-of-1 methodology [16, 26]. These consumer technologies have already attracted research attention. For instance, activity trackers made by Fitbit, Inc, have been deployed as instrumentation in over 450 public scientific studies [27]. Of course, application of wearables for clinical or research practice requires the technology to be valid and reliable. Research has found considerable variation of accuracy in different consumer wearables, including activity trackers [28–30], sleep trackers [31, 32], and wrist worn heart rate monitors [33, 34]. Despite this variation, there have been some notable successes. For instance, in an innovative two year study published in 2017, Li et al. demonstrated that measurement of heart rate and skin temperature using consumer wearables could predict inflammatory response as revealed by laboratory blood work showing elevated hs-CRP and onset of symptoms [35]. In presenting the articles in this focus theme, we aim to encourage attention to single subject research from from both scholars and researchers in health and biomedical informatics who may play a key role in advancing its practical methods and resolving doubts about its power and validity.