Multiple Cause and Remedy Models of Depression
In its underlying ideas and assumptions, the Anatomy departs from standard contemporary analyses of depression, suggesting instead a model of disease causation with parallels to the network models very recently proposed as alternatives to more traditional assumptions. Two of these traditional assumptions are distinguishable. Most commonly accepted and familiar is an “etiological” model of disease whereby depressive disorder names and depressive symptoms are attributed to, an underlying organic state of disorder or dysfunction. This way of thinking has been associated with medical psychiatric orthodoxy, including the “harmful dysfunction” definition of mental disorder found in the 4th and recent 5th editions of the DSM (APA 2013). The following, typical definition of depression provides another example: “Depressive disorder is a long term, relapsing condition associated with high levels of disability and mortality. It has a neurobiological basis and is associated with functional and structural brain abnormalities” (Palazidou 2012: 1, emphasis added). Beneath the potentially misleading appearances of mere symptoms, it is here presupposed, real disorders involve objective dysfunctions that account for why the symptoms occur, as well as the way they cluster as they do to produce a characteristic profile. A related assumption, known as the common cause hypothesis, is that a single cause is so implicated (Cramer et al. 2011). On this model, all symptoms arise from, and are explained by, the same underlying dysfunction, imbalance, disorder, or diathesis condition (Kramer 2005; APA 2013). Depression symptoms are the product of underlying, endogenous causes, albeit that those causes remain, as yet, incompletely identified. The etiological and common cause hypotheses are separable, but in descriptions of depressive disorder, they are customarily combined. Contemporary conceptions of prevention, noted earlier, also reflect the common cause model, as the language of “relapse” and “remission” indicates. Although not continuous, episodes of depression are united by their common source: they are one chronic condition, rather than several.
Alternatives to the common cause etiological model hypothesize that depression symptoms are not the result of any single or underlying, endogenous disease process or dysfunction, but instead comprise a causally interrelated cluster, or network, of features, attributable to diverse causes, including the stressful life events with which depression has been observed to be correlated (Kendler et al. 2011; Borsboom and Cramer 2013; Wichers 2014; Zachar 2014; Cramer et al. 2011). While they may each be realized in biological states, the symptoms within any network are not each traceable to underlying biological dysfunction or disease. The totality of symptoms making up a disorder may not result from underlying causes at all, single or multiple. It is rather that, together, and over time, symptoms will arise and trigger one another to form a circuitry of causal relationships. Feedback loops result within these symptom clusters, and particular features will enter into a large number of mutual relationships with other symptoms in the network. Over time, a “vicious circle of experiences” will result; this is a severe disorder and a pattern that an individual is increasingly unable to escape (Wichers 2014: 1353).
The strongest argument in support of a network model, it has been asserted, is that it seems implausible to assume that symptoms like insomnia and fatigue, for example, are only correlated because both are the result of an underlying liability to develop depression (Cramer et al. 2011). It seems obvious that insomnia might directly cause fatigue and that a pattern of symptoms might emerge that goes from fatigue to concentration problems to self-reproach to depressed mood to insomnia and so on - in a loop that, with repetition, and through time, brings about a depressive episode.
A working presupposition of the Hypothesis of depression, network model is the presence of a finite set of symptoms forming a cluster that is relatively cohesive and stable. Such syndromes will vary in their degree of coherence, stability, and boundary fuzziness, leaving them dissimilar to the categorical kinds often associated with etiological models. And their status as distinct entities may shift with changes in the networks involved. The interrelated symptoms making up the network need not share any commonality, either. Some symptoms and symptom clusters in the domain overlap with others, but no shared feature need unite all of them (Zachar 2014: 238).
Burton’s account of the causes and symptoms of melancholy has much in common with the network model. First, it acknowledges that melancholy’s initiating causes include many that are the sort of stressful life events identified as causes in network models of depression. Thus, as we saw, the interaction Burton describes cannot be entirely captured in any set of uni-directional, body-to-mind, causal sequences. The rapid and incessant shifts and changes making up the course of melancholy in any individual anticipate the network model in the chaotic plurality of its multilevel, multicausal explanations, and also in the incessant feedback loops, where elements are transformed from cause to symptom to cause and back. And, contrary to much that has been asserted or assumed about causation in humoral medicine, underlying humoral states in no consistent way cause symptoms, as they might be supposed to do following the etiological model of disease. The arrangement of the humors is sometimes assigned a causal role; but no regular pattern of causation can be traced from bodily “temperature” to melancholy symptoms.
Also conforming to the network model is the temporally extended natural history of melancholy states at the center of Burton’s analysis. The examples used earlier, where symptoms (sadness, solitariness) become in their turn causes of further symptoms, illustrate a temporal sequence. By following the course of the disorder this way, it is possible to attribute its presence and extent. Severe melancholy is identified with the “chronic,” habituated condition. The transformation from brief and passing states of melancholy into the intractable condition that resists all efforts at amelioration is a process. Because entrenched and more severe forms of the disorder result from repetition, this is a final parallel between Burton’s melancholy and depression understood on a network model. The model relies on a cumulative effect. Although it will reflect other factors as well (e.g., temperamental tendencies and environmental triggers), severity is partly a function of the number of feedback repetitions linking symptoms together.
Just as an illimitable number of causes for melancholy are permitted in the Anatomy account, so it seems to indicate that an almost infinite range of treatments, preventatives, and remedies will also likely be needed in any given case of melancholy symptoms. No one of these might alone be sufficient, Burton indicates, but together they may succeed. “If not alone, yet certainely conjoyned [they] may do much” (III,2,5,2: 207). Associated with today’s common cause, etiological disease models, where a single causal factor is assigned to account for the symptom cluster relied on for the diagnosis of depression, there is a tendency to think in terms of a single, isolated treatment response: with a sufficiently targeted treatment, a single cause can be expected to require only a single remedy. This “magic bullet” attitude has been most evident in the claims made on behalf of antidepressant drugs (Healy 1997: 21). The misleading specificity associated with magic bullet thinking about depression was central to David Healy’s critique of these assumptions. Until the last decades of the nineteenth century, he asserts, the idea that there might be a specific remedy for a specific ill was “tantamount to quackery” (Healy 1997: 10). On the version of the common cause, etiological disease model, Healy refers to as the “bacteriological model” of depression, a specific, underlying vulnerability or dysfunction is necessary, if not sufficient, to explain all symptoms. That is the “disease specificity,” to which “therapeutic specificity” was an obvious response (Healy 1997: 10). The mistaken idea of therapeutic specificity, Healy thus attributes to the misapplication of models from one disease (the infections successfully treated by antibiotics) to another (depression).
In the case of depression, Healy has argued, neither disease nor therapeutic specificities are appropriate. Depression is itself nonspecific: the category is better seen to refer to a generalized neurasthenia that, when it responds to antidepressant drugs, does so because their effect is also nonspecific. Appealing to humoral medicine, with its emphasis on illness as imbalance, he asserts that today’s antidepressant drugs probably work as “humoral” remedies, in the manner of old-fashioned tonics affecting sleep and appetite (Healy 1997: 259). In spite of these allusions to humoral medicine, and the holism common to each account, Healy’s analysis apparently differs from Burton’s inasmuch as for Burton expectation may in some considerable part contribute to healing. If placebo effects are necessary though not sufficient for therapeutic effectiveness, then they will interfere with the intended effects. No specific remedy can be effective for all melancholy symptoms because of the vagaries of placebo effects that, as Burton says, might allow one individual to be entirely healed merely by a strongly held idea (I,2,3,2). What is importantly shared between the Anatomy account and Healy’s, however, is the presupposition that neither the causes nor the treatments of these disorders are specific in their effects. No single cause engenders melancholy; moreover, no single remedy will alone be likely to be sufficient to treat it. And, if Healy is correct, the same is true for depression.
Should a multicausal account of depression require a multi-remedial approach, we need to understand why. So how is Healy’s allegiance to nonspecific remedies to be understood? Part of his discussion, it must be pointed out, is directed at the incomplete neuroscience that attributed spurious specificity to antidepressant substances whose effects are, or were at the time of his writing, incompletely understood. The regulations instituted to protect consumers are the source of some of this problem. By the middle of the twentieth century in the USA, Healy emphasizes, the FDA designated certain drugs as prescription only, insisted on randomized control trials, and encouraged the pharmaceutical industry to develop specific treatments for what were judged to be specific diseases (Healy 1997).
In light of network suppositions, two further explanations for this link between multiple causes and multiple treatments deserve attention. Integrated and holistic medical approaches emphasize the synergistic effects arising out of combining remedies, as we saw. Such enhanced effects might occur and if so might serve to diminish, disrupt, or sever intra-network bonds. On a different (although compatible) hypothesis, it can be argued that no specific remedy will be effective, since the networks comprising the totality of depression symptoms are insufficiently coherent and/or stable. This is a position that we might fairly attribute to Burton with regard to melancholy, but it can also be found in the work of sociologist Alain Ehrenberg (2010), who similarly rejects any common cause or essentialist account of depression. There has been a transformation of the meaning of depression, Ehrenberg asserts, prompted in part by changes in the goal, refinements, and effects of the newer antidepressant medicines. Stimulants to action replaced the older idea of a cure for pathology. And no longer a curable illness, depression had become, and was by our contemporary psychopharmacology treated as, a chronic, incurable personality trait of inhibited action. Insufficient or excess serotonin is not the cause of the several disorders SSRIs treat, but rather the “neurochemical vector of a person’s equilibrium,” systemically regulating the balance between the opposing dimensions of inhibition and impulsiveness (Ehrenberg 2010: 169). Whether depression symptoms have so wholeheartedly changed their nature, and in this particular way, we might debate. But such changes seem entirely compatible with the network model and with Healy’s concerns over magic bullet reasoning. Many remedies are required to address the range and variability of symptoms included in the loose category of depression.
-  Empirical research on the networks model of depression is thus far limited. But one study ofsymptoms caused by stressful life events has shown the network model to have significant predictive advantages over common cause hypotheses (Cramer et al. 2011).
-  The relation between more and less severe forms of depression is disputed, those distinguishingmelancholic and non-melancholic depression asserting a categorical separation. Moreover antidepressant medicines have long been known to outperform placebo more robustly when severe disorder is involved (Klein 1974).
-  Healy traces the initial magic bullet to 1930s work resulting in the first antibiotics (sulfa drugs).
-  That a multifactorial etiology for depression argues for a multipronged approach to interventionis voiced in some recent analyses. See Jacob (2012).