Desire for greater clarity when defining ‘cognitive remediation’ in reviews of treatment efficacy for schizophrenia

To the Editor

In the last 15 years, cognitive remediation (CR) has emerged as a promising therapeutic intervention for improving cognitive and psychosocial outcomes in people with schizophrenia. As defined by the Cognitive Remediation Experts Working Group (2012), CR represents a training-based therapy ‘targeting cognitive deficits (attention, memory, executive function, social cognition, or meta-cognition) … with the ultimate goal of improving functional outcomes’ that are durable over time (McGurk et al., 2013: 134). Similar iterations have been cited previously. Nevertheless, we raise the possibility of inconsistencies in the application of this definition.

Cognitive rehabilitation, at least in the context of traumatic brain injury, can be grouped according to three general approaches: restorative, compensatory (internal compensation) and environmental (external compensation; Velikonja et al., 2014). In the psychiatric literature, while the greatest outcomes may involve a combination of these approaches, it is generally accepted that CR involves directly targeting an area, or areas, of cognitive impairment and/or teaching strategies for the self-management of these skills. Therefore, by definition, a CR intervention in schizophrenia could involve a restorative ‘drill-and-practice’ programme that theoretically targets distributed neural circuitry underpinning specific cognitive functions and/or a ‘strategy-based’ intervention providing internal strategies to support cognitive difficulties (i.e. contextualisation and semantic clustering for verbal learning and memory).

Importantly, interventions that use ‘alternate strategies for teaching psychosocial skills (e.g. Work skills), such as errorless learning’, or those involving environmental modification, are considered to be distinct from CR (McGurk et al., 2013: 134). This becomes problematic, therefore, when reviews of treatment efficacy, such as one of the most cited by Wykes et al. (2011), include an errorless learning paradigm as well as a combined internal and external compensatory intervention. This is further problematic given that the weighted effect sizes of these types of studies are among those providing evidence in support of functional generalisation of CR (i.e. 0.32–0.99). It is noted that CR is a broad concept and that multiple interventions fall under this umbrella. Nevertheless, consistency between commonly accepted definitions of CR and the therapeutic elements of studies considered for systematic review or meta-analyses will be essential during ongoing discussions of treatment effectiveness and suitability for clinical translation.

In the interest of transparency going forward, we encourage the authors of new reviews to explicitly detail the type(s) of interventions they are evaluating (as well as the types excluded) to promote greater clarity and specificity of reviewed research findings.