10 key take-home messages from an expert workshop on single-case methodology
Author: Dr Suzanne McDonald
Single-case (or ‘N-of-1’) methodology provides researchers and clinicians with a flexible and rigorous alternative to using group-based research designs. There is a growing recognition of the benefits single-case methods offer in health-related research. In November 2018, the Recover Injury Research Centre and the Centre of Research Excellence in Recovery Following Road Traffic Injuries hosted a two-day, ‘hands on’ workshop on single-case methods at The University of Queensland, Brisbane, Australia. Over the course of the two days, participants were shown how to design, conduct and analyse scientifically rigorous single-case studies by single-case experts. 52 people attended the workshop, coming from a wide range of disciplines including medicine, psychology, physiotherapy, neurorehabilitation, statistics, nutrition, complementary and alternative medicine, sport science/exercise physiology, digital health, occupational therapy, nursing, speech pathology and dentistry.
Summary of workshop presentations
Professor Robyn Tate opened the workshop with an overview of the different variants of single-case experimental designs (SCEDs) and the current methodological standards for design and evaluation. Next, Dr Suzanne McDonald introduced single-case observational designs (SCODs) – a novel type of single-case design that does not involve the implementation of an intervention – and the key methodological issues that should be considered when designing a SCOD study. Dr Michael Perdices then described traditional visual analysis and statistical analysis and their advantages and pitfalls in evaluating SCED data. Day one of the workshop ended with a presentation from Associate Professor James McGree, who discussed concepts in Bayesian statistics and provided examples of using Bayesian statistics for analysing individual and aggregated single-case data.
On the second day, Professor Tate and Dr Perdices focused on how to get the most out of single-case experiments and to maximise their scientific rigour. This included discussion about how to select designs that are suited to answering the research question and testing the intervention of interest, as well as key issues related to increasing the internal and external validity of SCEDs. Dr Percdices provided worked examples to enable workshop participants to try out a range of visual and statistical techniques themselves. Professor Tate and Dr Perdices then described a 10-step procedure devised to facilitate the design and conduct of high-quality single-case studies. The workshop closed with a group session, where workshop participants discussed and developed their own proposals for single-case research under the expert guidance of the workshop presenters.
10 take-home messages about single-case designs
There are two main categories of single-case designs; experimental designs (SCEDs) and observational designs (SCODs).
There are many design sub-types within the SCED category. Some of these include intervention withdrawal (ABA and further permutations), multiple baseline designs (across participants, settings or outcomes), and changing-criterion designs
The quality of SCEDs is improved if investigators follow published methodological standards for the design, conduct and reporting of SCEDs such as RoBiNT and SCRIBE
SCODs are a novel type of single-case design that can be used to describe patterns and predictors of naturally-occurring phenomena over time, such as symptoms and behaviours.
The results from a study using a SCOD can inform the design of highly personalised health interventions, which, in turn, can be evaluated using a SCED.
In studies using SCEDs and SCODs, it is essential for investigators to fully define and operationalise the target behaviour/outcome a priori.
SCODs and SCEDs are flexible research tools that can be tailored to the needs and preferences of participants.
The best approach to evaluating single-case data is to use both visual and statistical analysis and there are a number of online resources that researchers can use to analyse single-case data themselves.
A series of single-case studies can be statistically aggregated to estimate population effects making it possible to make inferences that are equivalent to randomised controlled trials.
There is support available for researchers or clinicians who would like to conduct single-case studies from the International Collaborative Network (ICN) for N-of-1 Trials and Single-Case Experimental Designs. The ICN also provides a platform to facilitate discussion and collaborative activities between members, with the goal of furthering the science of individualised research.
About the author
Suzanne McDonald is a clinical researcher and chartered psychologist interested in the development and application of N-of-1 trials, SCEDs and SCODs in medicine, psychology and digital health. Suzanne is co-chair of the ICN.