A review of ssri-induced indifference

Discussion

The aim of the review was to study the prevalence of the adverse effect variously termed emotional blunting or apathy syndrome in the earlier literature. As the authors Sansone and Sansone⁵ do, it may be more accurate to regard the loss of emotional reactivity, diminished motivation, indifference and their associated behaviors as comprising a cluster of symptoms.¹ 

All the biases inherent in self-reporting instruments³⁹ such as the participant’s introspective ability, their suggestibility and understanding, their interpretation of terms and the control of the sample (selection bias and with anonymous internet samples the dangers of spurious or multiple responses) exist in these measurements, but perhaps made greater by the complexity and subjectivity of the task asked of them, that of judging their emotional responses to stimuli. Nonetheless, they are a necessary tool for researching mental states and behavior, and there are ways to identify and minimize these biases.⁴⁰ Few of the studies discuss their attempts to either measure or minimize these biases. The study by Mccabe et al.¹¹ presents a novel method to allow for an objective and independent measure of emotional reactivity by measuring imaging findings after delivering stimuli, alongside a self-report measure. While there are problems in the short time span of the study and the analysis of the results, the study presents interest for new alternative study designs to the reliance on self-report instrument.

The symptoms of apathy and emotional blunting are complex mental states that have many possible causes/explanations. A patient on SSRI treatment may have risk factors such as residual symptoms of depression and personality vulnerabilities. Patients taking medications may experience adverse side effects that are not a direct result of the medication but due to expectations of adverse effects⁴¹ or the psychological meaning they ascribe to being on medication.⁴² The various studies generally have not adequately addressed these confounding factors through good study design. The cross-sectional studies cannot tell us if the symptoms are related in time to the SSRI treatment, or rely too much on the recall of the patients, with the attendant self-reporting bias. Many of the studies attempt to adjust for depressive symptoms by applying a depression assessment scale, commonly the Hamilton Depression Scale. As many studies note, this analysis cannot discriminate between whether the outcome reflects partially treated depression or a new effect of the treatment.
As discussed above, a key problem was in the selection bias inherent in the study populations of many of the studies, with large potential pool of subjects in open invitations to recruitment with proportionally smaller population of participants and a failure to account for the bias. It would be expected that the study population would self-select for those who are sensitive to over-reporting, for example. This may account for the unequivocally high rates of positive findings of an association between emotional blunting/apathy and SSRI treatment in the studies reviewed. Finally, another factor accounting for these mostly positive results is the multiple comparisons problem in many of the studies, described in the results section. The studies have rarely established explicitly the outcome measures prior with reasons prior to the study design with consequent cherry picking of the positive findings in several of the papers within the within the multi-item questionnaires (examples include Marazitti et al.¹⁹ and Aydemir²²), and the inadequate discussion of the multiple comparisons problem in the fMRI study (McCabe et al.¹¹ as described above.
This review demonstrates the various levels at which the studies can be methodologically compromised in a complex subject such as this. From the selection bias at the recruitment stage, the validity of the instruments and the manner of their application at the measurement stage, the poor adjustment for confounding factors and the problem of multiple comparisons at the analysis stage, the researcher has many pitfalls to avoid in order to reach a conclusion one way or another.