The Euthymia Scale: A Clinimetric Analysis

Abstract

Fava and Bech [1] were the first who provided a comprehensive definition of euthymia since they conceptualized it as a condition characterized not only by the absence of affective disorders, as commonly conceived in psychiatry, but, most importantly, as characterized by the presence of psychological flexibility and resistance to stress. In this vein, they [1] developed a self-report rating scale, called Euthymia Scale (ES), which incorporates Jahoda’s conceptualization of euthymia [2], that is the individual’s balance among psychic forces leading to high levels of resilience and frustration tolerance, and the 5-item World Health Organization Well-Being Index (WHO-5) [3], a widely used measure for the assessment of psychological well-being. Indeed, 5 of the 10 items of the ES were derived from the WHO-5.When Fava and Bech [1] developed the ES, they focused on the clinimetric approach [4, 5] which suggests scalability, incremental validity, and sensitivity as main clinimetric parameters for the validation of a scale. Scalability means that each item measures a specific symptom and that the symptoms assessed via the scale belong to an underlying clinical dimension. When the criteria for scalability are fulfilled, the total score of a scale is statistically sufficient for measuring the severity of the clinical condition under assessment [6]. Incremental validity refers to the evaluation of the unique contribution of a rating scale in the assessment and prediction of a specific clinical dimension [4]. Sensitivity is the ability of a scale to discriminate among different groups of subjects who are affected by the same disorder and to detect changes in symptoms following a treatment [4].The aim of the present study was to test two main clinimetric properties of the Euthymia Scale: scalability and incremental validity. They were tested to verify whether ES is a statistically sufficient measure of euthymia and gives a unique contribution in the assessment of euthymia, beyond the WHO-5 score. Concurrent validity between the ES and the Psychological Well-Being Scales (PWB) [7] was also explored.Two samples were enrolled: one consisted of type 2 diabetes outpatients consecutively recruited at the Diabetes Unit of the University Hospital of Chieti (Italy), the other consisted of a convenience sample of healthy subjects recruited from the general population of Florence (Italy).Diabetes outpatients had to meet the following inclusion criteria: (1) a clinical diagnosis of type 2 diabetes according to the diagnostic criteria of the World Health Organization; (2) age 18–64 years. The inclusion criteria applied to the healthy subjects were: (1) being healthy; (2) age 18–64 years. Online supplementary Table S1 (see www.karger.com/doi/10.1159/000496230 for all online suppl. material) reports the exclusion criteria applied.All participants had to provide and sign a written informed consent including a privacy protection disclaimer. The research study was approved by the Institutional Review Board of the University of Florence.Online supplementary Table S2 reports the standard procedure followed to translate and pilot-test the ES.The final version of the ES was administered in a paper-pencil format together with the WHO-5 [3], the PWB [7], and the NEO Personality Inventory-3 (NEO-PI-3) [8]. Online supplementary Table S3 reports a detailed description of the rating scales administered.Scalability was examined via the Mokken analysis; a Loevinger’s coefficient of homogeneity ≥0.30 is acceptable [6].Incremental validity was tested via hierarchical linear regression analyses. The criterion variable was each PWB/NEO-PI-3 subscale. The entry order of predictors was: WHO-5 at step 1, the ES at step 2. The increase in the explained variance from step 1 to step 2 served as a measure of the incremental validity. Hierarchical regressions were adjusted for sex, age, education, and history of psychiatric disorders since they correlated with criterion variables.Pearson correlation coefficients were run to examine concurrent validity.Of the 124 diabetes outpatients who were invited to participate, 100 accepted (response rate: 80.6%). They had a mean age ± SD of 58.59 ± 6.06 years, 42% were females. Of the 110 healthy subjects who were invited to participate, 100 accepted (response rate: 90.1%). They had a mean age ± SD of 32.64 ± 10.48 years, 73% were females.When the Mokken analysis of the ES was run, the coefficient of homogeneity of the first 5 items of the ES, which assess psychological flexibility, was acceptable in diabetes outpatients. The coefficient of homogeneity of the last 5 items of the ES, which overlap with WHO-5 items, was acceptable in both samples (Table 1). However, when all 10 items were included in the analysis, the coefficient of homogeneity was below the level of acceptance (Table 1).In the clinical sample, the ES significantly increased the percentage of the explained variance (from 5 to 15%) in all PWB subscales except