Comments on “Sensitivity and Specificity Analysis: Use of Emoticon for Screening of Depression in Elderly in Singapore”

Abstract

To the Editor: This letter is written in reference to an article titled “Sensitivity and Specificity Analysis: Use of Emoticon for Screening of Depression in Elderly in Singapore” published in the Journal of the American Psychiatric Nurses Association (Tan, Toh, Sim, & Low, 2018). Depression has become a worldwide mental health problem in the past two decades (Lim et al., 2018). Depression in older adults is associated with increased risk of morbidity, mortality, and suicides (Rodda & Carter, 2011). Early screening of depression is therefore paramount to ensuring early detection and treatment. Existing screening tools for depression can be long and burdensome for the elderly (Dowell & Biran, 1990) and their accuracy is limited by the quality of the assessor. A simple, user-friendly depression screening tool for the elderly, especially those who have hearing impairments and/or speak only dialects, can help reduce the burden and improve diagnostic accuracy. In our recently published, cross-sectional study on the use of emotion scale as a simple visual assessment tool to screen for depression in the elderly in Singapore, we did not find the tool to be useful. In the study, a total of 77 participants aged 65 and older rated their mood using seven simple emoticons with smiling faces (score 1-3), a neutral face (score 4), and frowning faces (score 5-7). The emoticon scale showed low sensitivity when compared with the gold standard for diagnosing depression, which is the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; American Psychiatric Association, 1994) criteria. We concluded that the emoticon scale is easy to use in the elderly as it took on average 3 minutes to administer, and none of the elderly complained about it being burdensome. However, its ineffectiveness as a depression screening tool may be a result of lack of measurements in the other domains of the DSM-IV criteria such as sleep, energy, and appetite. Similar to the one-item emoticon scale used in our study, the One-Question Screen only measures the mood domain in DSM-IV. However, unlike the emoticon scale that uses pictures, the One-Question Screen uses a simple verbal question “Are you depressed?” to detect depression. The One-Question Screen was first developed in 1999 as an easy to use, practical tool to screen for depression in the primary care setting (Williams et al., 1999). A later study that used this tool to screen for depression in people with multiple sclerosis found it to have a sensitivity of 91%. However, in the same study, if the person does not recognize or denies mood problems, the tool’s sensitivity drops to only 70% (Vahter, Kreegipuu, Talvik, & Gross-Paju, 2007). The Two-Question Screen that adds a question of anhedonia (i.e., having little or no pleasure in activities) has shown to have better diagnostic performance than the One-Question Screen (Tsoi, Chan, Hirai, & Wong, 2017). In a recent, extensive meta-analysis (Tsoi et al., 2017) that examined the diagnostic accuracy of the Two-Question Screen in the elderly and compared it to the other screening tools for depression, the authors recommended the Two-Question Screen (with the cutoff point of one) to screen for depression in the elderly. As the Two-Question Screen is comparable to other screening tools in its psychometric properties and is easy to use, the Two-Question Screen is preferred over other instruments in screening for depression in older people across both clinical and community settings. The study also found that the OneQuestion Screen had the lowest sensitivity among the screening instruments. A Bayesian meta-analysis that aimed to evaluate effectiveness of one or two simple questions in detecting depression in cancer patients also found that the Two-Question Screen is significantly more accurate than either single question (Mitchell, 2008). It therefore seems that the emoticon scale was merely lacking in a second item for increased sensitivity and specificity. The challenge would be to add a pictorial version of anhedonia to the scale. Using faces to represent mood seems like a common concept, but the concept of anhedonia first involves having an expressed interest in something. Despite that, we believe that it is possible to pictorially represent anhedonia. A study to understand pictorial representation of anhedonia would be a prudent step before we combine both scales to form a two-item pictorial scale to screen for depression.