Comorbidity
Comorbidity between Major Depressive Disorder and Dysthymia
- 75-90% of people with Dysthymia have a lifetime history of MDD - 25% of people with MDD report having a history of Dysthymia - Among people with dysthymia, the risk of having a first lifetime MDE is 77%, Klein et al., 2000.
Spurious Association Models
A third variable explains the comorbidity. Diagnosis have the same risk factor or etiology. Can test this model with epidemiological studies.
Is it a subtype or a distinct disorder?
Distinct disorders have have differing developmental courses, etiologies, brain mechanisms, clinical features, and treatments. Example: Schizophrenia and ADHD Subtypes have different clinical features but the same developmental courses, etiologies, brain mechanisms, and treatments. Example: ?
Information Processing Models of Comorbidity
Increased attention to threat material in anxiety. Increased memory for mood-congruent infomation in depression. Studies examining their interpretation of text and expectancies find: Depressed: predict increase in negative events & decrease in positive events Anxious: only predict increase in negative events.
Broadband Psychopathology Models Explaining Comorbidity
1. Same brain areas are implicated in the different disorders - NIDA, 2010 2. Broad genetic, shared liability 3. Attachment theory/interpersonal models - Bowlby; Arieti & Bemporad 4.Phenotypic Groupings - Mineka et al. 1998 >>2 Factor Model >>Clark & Watson Tripartite Model >>Barlow's 3-Factor Model 5. Barlow's Hierarchical Model of Anxiety Disorders - Triple Vulnerability Model 6. Cognitive/Affective models 7. Information Processing Models
Bivariate models of comorbidity
1. Spurious Association Models 2. Associated Liabilities Models 3. Multiformity or Multifinality 4. Causation Models 5. Independence Models
Questions to keep in mind when developing theories about why MDD and Anxiety are so comorbid
1. What is the temporal relationship between anxiety and depression development? > It is much more common for anxiety to precede depression because people with anxiety are more distressed. > Anxiety symptoms and their impacts, e.g. isolation, may cause depression. >Anxiety involves hyper-arousal whereas depression involves hypo-arousal, indicating a cyclical quality to distress. 2. Why is depression more comorbid with some anxiety disorders? PTSD and GAD are most associated with depression, other anxiety disorders are also highly associated; some evidence of social but not simple phobia (maybe less interference?). 3. Why is pure depression so uncommon compared to pure anxiety?
Attachment theory/interpersonal models of comorbidity
Bowlby; Arieti & Bemporad: Anxiety leads to depression: threat of loss leads to anxiety, while actual loss leads to depression with some anxiety features
Definition of Comorbidity
Traditional definition: the correlation OR co-occurrence of two or more disorders. The co-occurrence of more than one disorder within an individual Feinstein, 1970. The co-occurrence of two or more disorders with distinct etiologies or, if etiologies are not known, distinct pathophysiology Vella et al., 2000 The covariation of two disorders Lilienfeld, 2003
Comorbidity between PTSD and Borderline Personality Disorder
33% of people with Borderline Personality Disorder also have PTSD. Potential explanation: most individuals with Borderline personality disorder have a childhood history of abuse, i.e., trauma. It is possible that if trauma occurs early in development, it is integrated into the individual's personality structure through changing coping skills, interpersonal skills, likely decreased social support, and neurobiological/neurochemical changes.
Multiformity
AKA Multifinality Heterogeneity in the expression of uncorrelated liabilities. A single liability can drive development of multiple different disorders and multiple liabilities can cause a single disorder. Comorbid cases are therefore those individuals high on liability one, high on liability 2, or having a mixture of both liabilities. A number of conditions have the same genetic single-nucleotide polymorphisms (SNPs), child temperament, autonomic functioning, parental psychological difficulties, parent-child conflict in common. Widiger and Clark, 2000 explore this model and important aspects the person, genetic makeup or developmental stage, and environment contribute - child stress, i.e. liability A + B, predict depression and anxiety but adult stress, liability B, only predicted depression, not anxiety. Can test this model with different phases: X, Y, XY predict X, Y, XY and studies on liabilities
Barlow's Hierarchical Model of Anxiety Disorders
AKA Triple Vulnerability Model Negative Affect is a higher order factor seen across anxiety disorders and accounts for overlap with depression Lower order factors that distinguish each anxiety disorder from others: hyper-arousal specific to panic disorder
Barlow's 3-Factor Model
Anxiety: distress/negative affect Fear/panic: autonomic arousal Depression: anhedonia, low positive affect, hopelessness
Comorbidity between generalized social phobia & Avoidant Personality Disorder
Comorbidity between General Social Phobia and Avoidant PD is 59% across studies. People with Generalized Social Phobia are more anxious and depressed than people with circumscribed Social Phobia. People with Generalized Social Phobia + Avoidant PD are more depressed and anxious than people with circumscribed SP. People with Generalized Social Phobia + Avoidant PD are more depressed and anxious than individuals with Generalized SP only, but this finding is less consistent and sometimes not significant. Chambless, Fidrych & Rodebaugh, 2008: There are few social differences between individuals with Generalized SP vs. Avoidant PD, but authors indicate that this may be an artifact of poor social skills assessment across studies. Conducted a study to test whether Avoidant PD was a severe form of Generalized SP or a distinct disorder. Participants were primarily Generalized Social Phobics, 65% of whom had comorbid Avoidant PD. GSP + APD individuals had more social phobic symptoms, more negative self-evaluations, lower self-esteem, and fewer social skills than individuals with GSP alone, but these differences were accounted for by social phobia symptoms. APD + GSP individuals had the same level of anxiety as individuals with GSP alone when performing task. Cluster analysis does not result in a separate group of Avoidant Personality Disorder individuals. Conclusion: Avoidant personality disorder, as it is currently defined, is just a more severe form of generalized social phobia, as there is no qualitative difference, only quantitative differences.
What is the prevalance of comorbidity?
Comorbidity is the rule rather than the exception in clinical and epidemiological populations Krueger and Markon, 2006; Kessler et al., 2011; Widiger & Clark, 2000 50% of individuals diagnosed with mental illnesses have a lifetime history of 3+ disorders aka multi-morbidity. Kendler et al., 1994. Furthermore, in clinical populations, the majority of people have more than one disorder e.g. Brown et al, 2001 found that 95% of clinical sample meeting criteria for lifetime history of MDD or dysthymia also met criteria for current or past anxiety disorder. Brown et al, 2001 found that 95% of clinical sample meeting criteria for lifetime history of MDD or dysthymia also met criteria for current or past anxiety disorder. There are several pairs of disorders for which the comorbidity rate is significantly higher than chance, so chance does not explain the overlap, Zimmerman & Chelminski, 2003. The number of disorders coded in the DSM has greatly increased over the past years; with this proliferation of disorders comes an increase in the overlap and chances for comorbidity Klein & Riso, 1993.
Clark & Watson's Tripartite Model
Depression specific: low positive affect Anxiety specific: physiological hyperarousal/somatic symptoms Shared symptoms: general distress & negative affect
Examples of the same brain areas being implicated in different disorders
Dopamine: the circuits in the brain that use the neurotransmitter dopamine are typically affected by addictive substances and may also be involved in depression, schizophrenia, and other psychiatric disorders. Dopamine pathways have also been implicated in the way in which stress can increase vulnerability to drug addiction. Stress is also a known risk factor for a range of mental disorders and therefore provides one likely common neurobiological link between the disease processes of addiction and those of other mental disorders. Use of substances → release of dopamine → positive mood states → change in threshold for pleasure/incentive sensitization → increasing amounts are used to maintain positive mood state Depressed people will find this extra reinforcing!
Independence Model
Each disorder is influenced by a latent factor, i.e. liability, that is unrelated to the latent factors, or liabilities, of other disorders. In this definition, co-morbid disorders are not simply the co-occurrence of two disorders at the same time, but rather a separate disorder that has not yet been defined. Can test this model with a longitudinal study showing stability of pure disorders X and Y and comorbid XY.
Associated Liabilities Model
Each disorder is influenced by a latent liability factor and these latent factors are correlated. This would explain why individuals with more risk factors often have greater comorbidity rates. A and B are separate liabilities; if these risk factors are correlated 0 <r <1; the cases would reflect correlation between risk factors. If r = 0, comorbidity is by chance alone. If 0 < r < 1, 'correlated liabilities model' - comorbid cases reflect correlation between liabilities. If r = 1, 'Alternate Forms Model ' A and B are interchangeable and can lead to disorder 1 or 2. E.g. MDD and GAD; genetic predisposition and family discord are liabilities for both
How comorbid are depression and anxiety?
First off, MDD and GAD share symptoms: sleep problems, trouble concentrating, indecisiveness. Mineka, Watson & Clark, 1998: anxiety disorders are as or more comorbid with depression as they are with other anxiety disorders. > In DSM-III, 57-58% of depressed patients also have a lifetime history of anxiety disorder, and about the same percentage of anxious patients have a lifetime history of depressive disorder. > In DSM-IV, 62% of GAD patients also have a lifetime history of MDD, which increased because hierarchical criteria. > 67% of panic disorder with agoraphobia patients also have MDD. > 20% of phobia patients also have MDD. Brown, Campbell et al., 2001: 2/3 of patients with principle MDD (MDD causing the most distress) and almost all patients with dysthymia could be diagnosed with GAD if not for hierarchy rule saying that symptoms are not accounted for by another mood disorder. Only 5% of sample had pure GAD. >55% of patients with an anxiety or mood disorder had at least one other mood or anxiety disorder > Only 5% of those with DYS or MDD had no current or lifetime history of anxiety disorders. More common for someone with depression to have anxiety than from someone with anxiety to have depression > Perhaps anxiety + impairment → depression > Research support for cognitive affective, genetic, symptom based models! Zimmerman & Chelminski, 2003: there is higher than chance comorbidity for GAD and MDD, and there is a worse prognosis for patients with comorbid GAD and MDD than with the individual disorders alone. Also, comorbid MDD & GAD patients differ in terms of symptoms and at the psychosocial level.
Causation Model
Having one disorder confers risk for another; i.e., the disorder develops not as a result of a specific liability for another disorder, but because of the actual other disorder. Directional Causation Model: one disorder is a risk for the other, e.g. ADHD → interpersonal/academic problems → depression or ODD. Reciprocal Model: disorders 1 and 2 cause each other Can test this model with longitudinal studies.
2-Factor Model explaining comorbidity
High negative affect in both depression and anxiety, but high positive affect only in anxiety
Cognitive/Affective models of comorbidity
Hopelessness/helplessness theory: helplessness predicts hopelessness, similar to Brown et al., 2001: showing the temporal relationship between anxiety and depression, Anxiety with lots of helplessness would lead to hopelessness and depression. Beck: Depression involves automatic negative thoughts about self. Anxiety involves automatic negative thoughts about harm or potential danger. Dozios & Dobson (2001): examined 4 groups: pure depressed, pure anxious, comorbid, and control, by having them rate themselves on interpersonal attributes and take tests. All disorder groups had more negative self-descriptions than controls. Pure anxious & control groups had more positive self-descriptions than pure depressed and comorbid. Overall, within subject analyses showed that depressed had equal negative and positive descriptions, while anxiety had greater positive descriptions than negative. Negative self-schemas more ingrained and organized with their roles than positive self-schemas for depressed folks.
How do you test different models of comorbidity?
Klein and Riso, 1993: 1. Epidemiological studies: Look at prevalence of pure and comorbid cases >Identify risk factors that may be shared - Support for independent liabilities if there is a double dissociation in those with pure cases >Look at population stratification 2. Longitudinal studies: Examine course and stability of pure and comorbid conditions >Heterogeneity is supported if both pure groups are stable over time and if comorbid group is unstable 3. Family Studies: Examine family members of probands w/X & Y vs. X or Y >If there is substantial similarity, than it supports the same disorder category >E.g., Kendall et al. twin studies showed phenotypic similarity between GAD & MDD supporting recategorization of GAD into mood disorder 4. Treatment studies: If we treat one disorder specifically and comorbid disorder is down at post-treatment, suggests that two disorders are not completely independent, although it doesn't show us how they are related. >Brown et al. 1995 treated 126 patients with CBT for panic disorder and found that comorbidity of other internalizing disorders went down significantly at post-treatment and noted that this questions the validity of the distinction. > Hein et al. 2010: found that when women with PTSD and comorbid SUD were treated for just the PTSD their substance use went down.
Integrated Liability-Spectrum Model
Krueger and Markon 2006 High rates of comorbidity of a large number of disorders can be explained by alternate expressions of a smaller number of underlying liabilities. There is a broad internalizing liability and an associated broad externalizing liability. The internalizing liability can be further broken down into distress and fear, which separates depressive and anxiety disorders, with GAD being more related to distress than to fear in comparison to other anxiety disorders. The Externalizing liability gives rise to the substance abusing, conduct disordered, and antisocial disorders.
What causes comorbidity?
Krueger and Markon, 2006 separate models into Bivariate Models and Multivariate Models using a structural equation modeling approach to assessing comorbidity.
Phenotypic Groupings explaining comorbidity
Mineka et al., 1998: 1. 2-factor model 2. Clark & Watson's Tripartite Model 3. Barlow's 3-factor Model 4. Barlow's hierarchical model of anxiety disorders, aka triple vulnerability model
Explanations of comorbidity
Mineka, Watson & Clark, 1998: 1. Genotypic Model - the underlying liability are genotypic 2. Phenotypic Models > Two Factor > Tri-partite Model > Barlow's Three-Factor Model > Barlow's Hierarchical Model of Anx Ds 3. Cognitive/Affective Theories > Hopeless/Helplessness > Beck's Depression 4. Attachment Theory (Bowlby, 1980) 5. Information-Processing Models
Multivariate Models of Causality
Multivariate models combine a number of univariate and bivariate models to capture more complex models. They are often too complex and not parsimonious. Example: Synthesizes associated liabilities and multiformity models. Liabilities A and B can be expressed as disorders 1, 2, and 3 and 3, 4, 5, and 6 respectively. Disorder 3 can result from the expression of liability A or B. Disorder 3 is explained by multiformity! The liabilities are also correlated.
Why are comorbidity rates a problem?
Patients with comorbid conditions have significantly different and more negative outcomes than patients with just one diagnosis. Zimmerman & Chelminski, 2003: compared demographic, clinical, family history and psychosocial characteristics of three independent groups: MDD without GAD, MDD with GAD, and pure GAD. Found that those with comorbid GAD/MDD had higher suicidal ideation, poorer social functioning, and higher rates of multimorbidity. Drabick & Kendall, 2010: comorbidity is common among youth and adults and is associated with increased symptom severity, worse impairment, more negative correlates, differential treatment response, and distinct courses. Jensen 2003: Although we know that comorbidity exists, we have little research about it, or about treating it, particularly as most treatment studies exclude individuals with comorbid disorders.
Why do we have such high comorbidity rates?
The DSM may be too precise in terms of demarcating boundaries between disorders. The DSM may be forcing dimensional disorders into a categorical framework, creating artificial separations of broader symptoms. >Comorbidity could occur due to chance, particularly as the number of disorders in the DSM increases. {In order to determine what 'chance' would be for comorbidity of two disorders, multiply the base rates of the disorder (like you would for any probability). This explanation does not cover most comorbidities, which occur at a much greater rate than chance.} >Symptom co-occurrence: there is an unnecessarily large overlap between symptoms required for diagnosis (consider MDD vs. GAD, ODD vs. ADHD, ADHD vs. Bipolar). >Diagnostic thresholds may be too low. >Sampling Bias: cases may be selected from "biased" samples because individuals with more disorders have more opportunities to be selected for various studies >Population Stratification: we sample from inpatient clinics, low SES groups, etc. and generally get samples with many risk factors (ex. childhood maltreatment, low SES, financial stress, single-parent households) that do not generalize to the greater population - associated liabilities model!
How might we reduce the comorbidity problem?
Widiger and Clark, 2000: suggest we rearrange diagnostic structure so that the very overlapping disorders end up in the same set or described with similar symptoms - consider sequential co-occurrence as a meaningful pattern. Hierarchical clusters of symptoms that are more or less specific to certain disorders. When sets of symptoms co-occur, it might help to label/diagnose for communication, research, and treatment purposes. Widiger and Sankis, 2000: propose getting rid of axis I/II distinction and making chronic or early onset subtypes of the Axis I disorders. They also suggest adding an interpersonal relatedness section. Hyman, 2010: have a classifications system that is empirically driven and focused on brain dysfunction. Krueger & Markon,2006; Kessler et al, 2011: assume correlated liabilities model and reorganize based on latent underlying factors.
