Cognitive screening in electroconvulsive therapy: Don’t forget subjective cognition

We read with interest the recent narrative review of brief cognitive screening instruments for electroconvulsive therapy (ECT) (Martin et al., 2020). While ECT is a highly effective treatment, the potential cognitive side effects can be problematic for patients. Assessing for cognitive side effects, and making appropriate changes to treatment (parameters, frequency), reduces the risk that long-term cognitive issues will develop. As discussed in the review by Martin et al., there is no consensus on routine assessment for cognitive side effects in clinical practice. Therefore, we were interested in their recommendations for routine screening for cognitive problems after ECT. The suggested approach by Martin et al. includes (1) measuring recovery of orientation following ECT and (2) use of a brief generalised cognitive screening measure (e.g. Montreal Cognitive Assessment [MoCA], Modified Mini Mental State Examination [3MSE]). While these are reasonable and pragmatic steps, there are some limitations with this strategy that merit further consideration. As Martin et al. note, the suggested cognitive measures were not designed for use in ECT and are screening instruments rather than standardised neuropsychological tests. They do not address autobiographical memory impairment, one of the most common and distressing side effects of ECT. Additionally, measures of subjective cognitive side effects were not included in the review.

Martin et al. excluded subjective cognitive measures from their review due to concerns that these may be conflated by patients’ perceptions and correlate poorly with objective cognitive outcomes. Subjective cognitive impacts are however, by definition, of importance to patients. Furthermore, it is part of the clinical situation that subjective impairment is linked with perception. Patients with more severe depressive symptoms are more likely to undergo ECT and have a greater risk of significant cognitive impairment as a feature of their illness. They may also be likely to report more negative expectations and side effects of treatment, due to the negative cognitions associated with a severe depressive illness. Patients with more severe depressive symptoms do, in fact, report more severe subjective cognitive impairment. Furthermore, subjective cognitive impairment may correlate more strongly with objective measures than previously thought (Brakemeier et al., 2011).

While there is limited data measuring long-term objective cognitive issues post-ECT, there are accounts indicating that subjective side effects continue and can be severe in some people (Vamos, 2008). Patients who have undergone ECT describe a sense that their subjective experience of memory loss is poorly understood and that clinicians’ attempts to minimise this experience can be perceived as invalidating (Vamos, 2008). Significant difficulties in functioning can result from cognitive side effects in the period immediately after ECT, but also longer term (Vamos, 2008). These effects are experienced in a number of different ways in everyday life, including impacts on communication, self-confidence, self-image and ability to function at work (Vamos, 2008). Taken together, this indicates that an increased emphasis on subjective cognitive impairment as part of the routine practice of ECT is important and may have a valuable therapeutic effect for those undergoing treatment.

We suggest that assessment of subjective cognitive side effects, in particular memory impairment, should be part of routine assessment during ECT and present supporting evidence for this.

Autobiographical memory impairment is a particularly important side effect of ECT, but it is difficult to assess. It is most commonly measured using the Colombia University Autobiographical Memory Interview (CUAMI), or its shorter form (CUAMI-SF). There are several difficulties with this, including the time taken to administer (15–20 minutes for the short form) and the issue of a ceiling effect (memory is reported as a percentage recalled from baseline and is therefore unable to improve following the initial assessment). An additional limitation is that the CUAMI-SF only measures more recent retrograde memory (i.e. the previous 2 years). The relationship between subjective and objective assessment of autobiographical memory impairment has been studied. While longer, more detailed scales measuring subjective memory impairment (Cognitive Failures Questionnaire [CFQ] and Squire Memory Complaints Questionnaire [SMCQ]) do not correlate well with the CUAMI, a high level of correlation has been demonstrated between the CUAMI and the Global Self Evaluation of Memory (GSE-My) (Brakemeier et al., 2011). As well as items aimed at remote memory, the SMCQ includes items measuring immediate memory and new learning. However, there is little evidence of impairment beyond the first 2 weeks following ECT in most cognitive domains (Semkovska and McLoughlin, 2010), with the exception of autobiographical memory. The CFQ and SMCQ do not include a measure of memory for recent events prior to ECT. Therefore, the patient’s global subjective assessment of memory (i.e. GSE-My) may better reflect retrograde rather than current memory function. Additionally, the GSE-My appears to be more sensitive to treatment parameters (type and amount of ECT) than the CFQ and SMCQ (Brakemeier et al., 2011).

The GSE-My consists of two simple questions. First, patients are asked to rate their expectations about the likely effect of ECT on their memory prior to treatment using a 7-point Likert-type scale. Second, post-ECT, patients evaluate the overall effect ECT has had on their memory, using the same scale. This is very similar to the routine clinical practice of directly asking patients about their memory and could easily be incorporated into screening for cognitive side effects of ECT. The GSE-My has the benefit of more systematic monitoring of either deterioration or improvement in subjective memory, compared with simply asking about perceived memory impairment. Of note, the high level of correlation with a detailed objective assessment of autobiographical memory suggests that the patient experience of memory impairment is important and closer to objective measures of memory than has been suggested.

A recent national registry study (Sigström et al., 2020) examined long-term subjective memory impairment after ECT using the GSE-My alongside a subjective memory measure (the Comprehensive Psychopathological Rating Scale [CPRS]). Importantly, subjective memory impairment was measured both prior to and following ECT, addressing a common limitation in many ECT studies. Sigström et al. demonstrated that in most (83.7%) participants, subjective memory remained unchanged or improved post-ECT. Of interest, subjective memory worsening was associated with more negative expectations of the likely memory effects of ECT prior to treatment. This finding suggests that expectations of side effects (and therefore outcome) may be modifiable. A pre-ECT assessment of patients’ expectations of the effect of ECT on their memory may be an opportunity for clinicians to address negative expectations and improve subjective memory outcomes as a result.

Inclusion of a subjective measure of cognitive side effects as part of routine assessment during ECT allows for important adjustments to treatment to be made, as with objective screening measures. Obtaining a baseline sense of the degree of current memory impairment is likely to be useful. Additionally, understanding and addressing patients’ expectations of ECT may help to improve subjective memory outcomes. Another potential advantage is that there may be therapeutic value in the clinician taking interest in and validating the individual experience for that patient.

In summary, most clinicians will, as part of assessing progress in a course of ECT, ask about the subjective experience of memory problems. We suggest that this common practice be made systematic by the inclusion of a simple, subjective measure of cognitive function in routine screening for cognitive side effects of ECT. Current evidence suggests that use of a simple, direct question about memory correlates well with detailed objective measures and therefore is a valid, time-efficient approach.