Elsevier

Journal of Affective Disorders

Volume 207, 1 January 2017, Pages 38-46
Journal of Affective Disorders

Reconsidering the definition of Major Depression based on Collaborative Psychiatric Epidemiology Surveys

https://doi.org/10.1016/j.jad.2016.09.014Get rights and content

Highlights

  • Answer has been called for the question “when does depression become a disorder?”.

  • Often presence of disability is a requirement for a relevant diagnostic definition.

  • A novel definition is derived using nationally representative data of US population.

  • The novel definition captures the cases with disability, excluding those without.

  • A minor adjustment of diagnostic definition may increase clinical relevance a lot.

Abstract

Background

Diagnostic definitions for depressive disorders remain a debated topic, despite their central role in clinical practice and research. We use both recent evidence and nationally representative data to derive an empirically-based modification of DSM-IV/-5 Major Depressive Disorder (MDD).

Method

A modified MDD diagnosis was derived by analyzing data from Collaborative Psychiatric Epidemiology Surveys, a multistage probability sample of adults (n=20 013; age ≥ 18 years) in coterminous USA, Alaska and Hawaii. The old and the newly suggested MDD definitions were compared for their associated disability (WHO Disability Assessment Schedule and number of disability days in past month), suicide attempt, and other covariates.

Results

Our data-driven definition for major depression was “lack of interest to all or most things” plus four other symptoms from the set {weight gain, weight loss, insomnia, psychomotor retardation, fatigue, feelings of worthlessness, diminished ability to think/concentrate, suicidal ideation/attempt}. The new definition captured all the disability implied by MDD and excluded cases that showed no greater disability than the general population nor increased risk of suicide attempts. The lifetime prevalence of the new diagnosis was 14.7% (95% CI=14–15.4%) of the population, slightly less than for the old definition (16.4%; CI=15.4–17.3%).

Limitations

Only conservative modifications of MDD could be studied, because of restrictions in the symptom data.

Conclusions

With only small adjusting, the new definition for major depression may be more clinically relevant than the old one, and could serve as a conservative replacement for the old definition.

Introduction

Clinical definitions for depressive disorders remain a debated topic, despite the high prevalence and burden of disability of these disorders. Recent studies have investigated the depressive disorders at the level of individual symptoms instead of composite clinical definition of Major Depressive Disorder (MDD) (Bringmann et al., 2015, Cramer et al., 2012, Fried and Nesse, 2015, Keller et al., 2007, Keller and Nesse, 2005, Lux and Kendler, 2010, Oquendo et al., 2004). The symptom-level analysis is attractive for basic research because the definition of the MDD “syndrome” is not well established empirically (Haslam et al., 2012, Keller et al., 2007; van Loo et al., 2012; Lux and Kendler, 2010; Solomon et al., 2001). At the same time, clinical practice requires criteria by which to assign individuals to treatment groups, and there is some evidence to support a temporal clustering of symptoms which is consistent with sudden phase transitions characteristic to syndromatic states (Hosenfeld et al., 2015, van de Leemput et al., 2014). Accordingly, there is a challenge to provide empirically based answers to the question “when does depression become a mental disorder” (Maj, 2011a, Maj, 2011b). Treatment assignment based on single symptoms may be infeasible and reliable biomarkers identifying depression are lacking, implying that symptom combinations need to be considered in both research and treatment. This paper draws from the new symptom-level findings and uses United States national-level estimates to derive an empirically-based recommendation for a more clinically salient definition for the Major Depression (MD; we drop the “Disorder” to distinguish between the suggested new ideas and the old definition, and to encompass both episodes and disorder).

The current Major Depressive Disorder (MDD) diagnosis requires the presence of at least one of the two core symptoms: (1) depressed mood and/or (2) markedly diminished interest or pleasure in all, or almost all, activities (American Psychiatric Association, 2013). However, it can be difficult to define depressed mood or sadness without referring to some other symptoms. For example, according to DSM-5 depressed mood is indicated “by either subjective report (e.g., feels sad, empty, hopeless) or observation made by others (e.g., appears tearful)” (American Psychiatric Association, 2013). But, how does one know when one is feeling sad? In April 18th, 2016, Wikipedia defined sadness as “emotional pain associated with, or characterized by feelings of disadvantage, loss, despair, helplessness, disappointment and sorrow”, which seems many things rather than one. In more quantitative terms, we have previously found that, after taking a sadness item into account, other self-report items provided little predictive value for the WHO-CIDI diagnosis of MDD (Rosenström et al., 2015); this is an expected finding when the “sadness” item implicitly implies multiple other symptoms. In a prospective network analysis of depressive symptoms, Bringmann et al. (2015) showed that sadness has a high “indegree” but a low “outdegree” and “betweenness” in relation to other symptoms, suggesting that (statistically) it mostly summarizes other prevailing symptoms instead of predicting them. Those with a lot of problems (symptoms) are likely to become sad (endorse the symptom in future), but the sad who currently lack the other problems are relatively unlikely to get lot of problems in the future.

Among the depressive symptoms, anhedonia (low positive affect) has been found to be relatively specific to depression e.g. when comparing depression, anxiety, and schizophrenia, whereas “depressed mood” represents “a mixture of relatively high NA [negative affect] and moderately low PA [positive affect]”, therefore being less specific to depression and more related to general distress (Clark and Watson, 1991, Joiner et al., 2003). Ideally, constructing a data-driven definition should start from the most elementary (specific) components available rather than use variables that already are a priori given compositions of many elements. Furthermore, sadness is frequently seen as a normal, adaptive response to loss (Kleinman, 2012, Wakefield and First, 2012, Wakefield and Schmitz, 2013). In contrast, lack of interest in “all, or almost all, activities” should reliably intervene with goal-directed behaviors. Therefore, anhedonia may also be intrinsically more disabling than sadness.

Based on the findings cited above, we take only “diminished interest in all, or almost all, activities” as the core feature of depression in our analysis, as learning models of depression have done (Griffiths et al., 2014, Trimmer et al., 2015). For ease of reference, we call this “lack of interest” also as “anhedonia”, although strictly speaking, it refers to lack of pleasure. We then study the following empirical questions: How do the other depressive symptoms distribute in the anhedonic population and what would be a sensible definition of depressive disorder based on that distribution? How much this new definition of depression overlaps with the old one? Which one of the definitions, the old or the new, is more clinically relevant? A diagnostic definition that implies higher level of disability, longer episodes, and greater probability of suicide attempts than an alternative definition is an example of comparatively “clinically relevant” diagnosis. These questions are analyzed here using a representative population sample of psychiatric symptoms in the United States, the World Health Organization’s (WHO’s) Collaborative Psychiatric Epidemiology Surveys (CPES) (Alegria et al., 2015).

Section snippets

Sample and procedures

CPES data joins together three multi-stage area probability samples, the National Comorbidity Survey Replication (NCS-R), the National Study of American Life (NSAL), and the National Latino and Asian American Study of Mental Health (NLAAS). Collection of the samples were funded by the National Institute of Mental Health and they were selected using the sampling frames and sample selection procedures that are common to the University of Michigan Survey Research Center's National Sample design,

A data-driven definition for MD

Altogether 82% of those who suffer from sadness or anhedonia, suffer from both. Overall, the prevalence of the other symptoms is at least as high in the anhedonic as in the sad people (Supplementary Table S1). Given these data and the arguments in the Introduction, the decision to limit the core symptoms to anhedonia seems reasonable and will be in effect from this point onwards.

While all the symptoms are correlated in the general population, many were virtually uncorrelated in the anhedonic

Discussion

Here we propose an alternative, data-driven definition for MD that is similar in spirit to the well-known DSM-IV definition of MDD, but based on less conceptually overlapping symptom descriptions. The new MD is slightly less prevalent in the population than the familiar MDD, but significantly more associated with disability and suicide attempts than the MDD. In fact, those who satisfied MDD criteria without satisfying the MD criteria did not significantly differ from the rest of the US

Limitations

A major strength of the empirical estimates provided in this study is that they are representative of the entire US population. In contrast, generalizability to the other populations is a limitation and a topic for further study. However, there is evidence for the same underlying structure of major depression despite significant cultural differences (Kendler et al., 2015), and depressive disorders have been described throughout the written history (Beck, 1967). Likewise, we mainly restricted

References (50)

  • L.A. Clark et al.

    Tripartite model of anxiety and depression: psychometric evidence and taxonomic implications

    J. Abnorm. Psychol.

    (1991)
  • A.O.J. Cramer et al.

    The pathoplasticity of dysphoric episodes: differential impact of stressful life events on the pattern of depressive symptom inter-correlations

    Psychol. Med.

    (2012)
  • E.I. Fried et al.

    The volumes of subcortical regions in depressed and healthy individuals are strikingly similar: a reinterpretation of the results by Schmaal et al

    Mol. Psychiatry

    (2015)
  • E.I. Fried et al.

    Depression sum-scores don’t add up: why analyzing specific depression symptoms is essential

    BMC Med.

    (2015)
  • E.I. Fried et al.

    Measuring depression over time … or not? Lack of unidimensionality and longitudinal measurement invariance in four common rating scales of depression

    Psychol. Assess. Epub Ahead of Print.

    (2016)
  • A. Gelman et al.

    Data Analysis Using Regression and Multilevel/Hierarchical Models

    (2007)
  • K.R. Griffiths et al.

    Translational studies of goal-directed action as a framework for classifying deficits across psychiatric disorders

    Front. Syst. Neurosci.

    (2014)
  • N. Haslam et al.

    Categories versus dimensions in personality and psychopathology: a quantitative review of taxometric research

    Psychol. Med.

    (2012)
  • K. Hayashi et al.

    On the likelihood ratio test for the number of factors in exploratory factor analysis

    Struct. Equ. Model.

    (2007)
  • S.G. Heeringa et al.

    Sample designs and sampling methods for the Collaborative Psychiatric Epidemiology Studies (CPES)

    Int. J. Methods Psychiatr. Res.

    (2004)
  • J.L. Horn

    A rationale and test for the number of factors in factor analysis

    Psychometrika

    (1965)
  • B. Hosenfeld et al.

    Major depressive disorder as a nonlinear dynamic system: bimodality in the frequency distribution of depressive symptoms over time

    BMC Psychiatry

    (2015)
  • L.G. Humphreys et al.

    An investigation of the parallel analysis criterion for determining the number of common factors

    Multivar. Behav. Res.

    (1975)
  • R.I. Jennrich et al.

    Exploratory bi-factor analysis: the oblique case

    Psychometrika

    (2012)
  • M. Jokela et al.

    Inflammation and specific symptoms of depression

    JAMA Psychiatry

    (2015)
  • Cited by (20)

    • Up-regulation of SETD3 may contribute to post-stroke depression in rat through negatively regulating VEGF expression

      2022, Behavioural Brain Research
      Citation Excerpt :

      Additionally, SETD3 knockdown relieved the depressive symptom of PSD through enhancing the sucrose preference, higher locomotor and rearing activities as well as reducing the immobility time of PSD rats. Depression is a common disorder worldwide, and has been linked to the increased risk of suicide, social withdrawal, and the impaired social skills [26]. More and more studies have supported that there is a strong relationship between the neurogenesis of adult hippocampal and the formation of new neurons in the dentate gyrus and depression [27,28].

    • The impact of social distancing during COVID-19: A conditional process model of negative emotions, alienation, affective disorders, and post-traumatic stress disorder

      2021, Journal of Affective Disorders
      Citation Excerpt :

      For example, feeling uncertain about loss tends to be associated with anxiety, while being sure of loss is linked to depression (Ehlers and Clark, 2000), which potentially influence the effect of negative emotions and alienation. Specifically, people with a high level of depression have been quite likely to suffer from negative emotions, feelings of loneliness, worthlessness, and hopelessness (American Psychiatric Association, 2013; Deprince et al., 2010; Rosenström and Jokela, 2017), which may aggravate the feeling of alienation. Depression is also characterized by social avoidance, rumination, and these outcomes will further make one indulge in negative thinking and undermine social support, enhancing the maintenance of PTSD (Ebert and Dyck, 2004; Fernández-Theoduloz et al., 2019; Berman et al., 2011; Spinhoven et al., 2015; Pugach et al., 2020).

    • Instrumental assessment of balance and gait in depression: A systematic review

      2020, Psychiatry Research
      Citation Excerpt :

      Moreover, existing ratings of motor symptoms are typically based on the subjective judgment formulated by clinicians on the basis of observed or recalled behavior (Fried, 2017; Kendler, 2017). Despite being under-represented in the assessment of depression, motor symptoms often comprise an independent factor from the affective, cognitive and neurovegetative domains (Drysdale et al., 2017; Rosenström and Jokela, 2017; Vrieze et al., 2014). In recent years, novel technologies have become available to reliably assess motor features, both in quantitative and qualitative terms.

    View all citing articles on Scopus
    View full text