What clinicians and patients want to know is not always what the existing evidence-base is equipped to say. The randomised controlled trial (RCT) is commonly regarded as the ‘gold standard’ for evaluating health interventions and findings from an RCT are used to select which interventions and treatments to implement in clinical practice. However, this is not always the case. For example, narrow inclusion criteria can result in the recruitment of patients that are not representative of those seen in routine clinical practice (e.g. older adults and/or patients with multi-morbidity are sometimes excluded) which limits the applicability of RCT findings.
There are other factors which limit the applicability of RCT findings. RCTs involve randomly allocating individuals to intervention(s) or control conditions and comparing outcomes between the two groups. Therefore, they identify whether an intervention works on average, with less regard for whether the intervention works for a given individual in the trial (although, some RCTs may estimate and report heterogeneity of treatment effects to address this). The focus on aggregated results for each group assumes that the average treatment effects for the group reflect the characteristics of the individual (also known as ecological fallacy). For example, an intervention which tests the efficacy of a new behavioural treatment for weight loss may be effective at the group level but the average weight lost may not reflect the weight loss of all (or any) individuals in the trial. In some cases the ‘effective’ intervention may be ineffective or harmful to some individuals (e.g., some people may gain weight while participating in a weight loss study). Furthermore, findings from RCTs reveal little about the timing of response to interventions. The timing at which an individual ‘responds’ to an intervention may also vary substantially between individuals. This information is often not well-represented or well-reported in the findings from RCTs, reporting only aggregated results.
To understand individual response to interventions it is necessary to adopt alternative methods. N-of-1 methods play a key role in evaluating interventions and theory, and have been recommended by the Medical Research Council for this purpose. They involve the repeated measurement of an individual over time to draw conclusions about that specific individual and have a long history of use in various disciplines such as medicine, neuropsychological rehabilitation, education and psychology. In n-of-1 RCT designs individuals are randomised to intervention and control conditions across different periods of time and outcomes are compared across the different periods. This design can be used to evaluate how individuals respond to health interventions such as different health behaviour change interventions (e.g. self-monitoring vs. goal-setting for increasing physical activity) or pharmacological treatments (e.g. celecoxib vs. paracetamol for reducing pain in osteoarthritis). As a result, individuals act as their own control.
In recent years we have seen a movement towards personalised medicine, shared-decision making and patient-centred care. Therefore, it is critical and timely to develop and extend research methods which can advance our understanding about individuals. N-of-1 methods offer a range of opportunities. For example, we can use N-of-1 methods as a tool to personalise health interventions to individuals by first identifying the unique predictors of a given outcome in one individual and tailoring the intervention to target these predictors. Most health interventions are complex, involving multiple components. N-of-1 methods provide the opportunity to examine the effectiveness of individual components, as well as different sequences, combinations and doses of components and this information can be used to optimise interventions. Although n-of-1 methods have been generally under-recognised and under-used in health research, we know that they these methods are perceived as acceptable and empowering to patients suffering from a range of chronic diseases, who recognise the value in participating in n-of-1 research to learn about the determinants of their conditions and symptoms. Therefore, n-of-1 methods represent an important component of the research toolbox and may facilitate a move away from the ‘one-size-fits-all’ assumption that has dominated health and social research.
Picture credit: Sharon Cronin. Original here: https://www.flickr.com/photos/66127779@N04/24935522849/