Introduction

Prolonged Grief Disorder (PGD) was approved as a diagnosis by the American Psychiatric Association (APA) in 2020; it was included in their official guidelines, the DSM-5-TR, in 2022. Similarly, the World Health Organization (WHO) approved a new diagnosis of PGD in 2018; it was added to their official guidelines, the ICD-11-TR, which came into effect in January 2022. Although the DSM-5 and ICD-11-TR diagnostic guidelines for PGD differ in some minor respects, such as with regard to the specificity of their criteria for PGD, the two converge in their agreement that grief of a certain duration and intensity is a mental disorder. In other words: current official diagnostic guidelines assert that grief can be a mental disorder requiring treatment.

Mental disorder classifications are interactive kinds (^{Hacking 1999}). Interactive kinds are classifications in which the classification schema may interact with the thing classified and vice versa. So, to say that mental disorder classifications are interactive kinds is to say that our social understanding of any particular mental disorder interacts with, and changes, both the expression of that disorder, who counts as having it, and how people (including those diagnosed) perceive those with the diagnosis. These interactions are commonly referred to as “looping effects”. That looping effects are relevant for determining the costs and benefits of classifying a particular constellation of symptoms as a disorder (this process of classification is known as “medicalization”) has been noted in the literature. However, their significance with regard to the addition of PGD to the DSM has not been adequately explored.

When thinking about the looping effects of medicalizing grief, philosophers have tended to focus on potential negative ramifications, with many concluding that the DSM should not include PGD as a new diagnosis. One prominent concern is that medicalizing grief will alter the stories we tell about ourselves as grievers in such a way that our ability to authentically engage with the loss(es) in question is hampered. For example, ^{Michael Cholbi (2021, p. 183)} argues that medicalizing grief may alienate grievers from their experiences of loss.¹ I think that this concern is well-founded: the medicalization of grief has the potential to alter the stories we tell about ourselves as grievers in a way that hinders our ability to authentically engage with loss. However, it is also crucial to note that the nature and impact of the looping effects associated with including PGD in the DSM depend on two underappreciated factors:

[1] Whether the looping effects of being diagnosed with PGD are better or worse than those associated with the alternatives. We cannot draw conclusions about the value of including PGD in the DSM until we compare the potential looping effects of both its inclusion in and its exclusion from the DSM.

[2] How grievers (and society more generally) conceive of the domain of healthcare.

I explore both [1] and [2]. With regard to [1], I argue that if a consequence of removing PGD from the DSM is that people experiencing complex grief are more likely to be diagnosed with a disorder such as depression or Post-Traumatic Stress Disorder (PTSD) instead of PGD then this might be of greater detriment to their ability to engage with the relevant loss.² Broader diagnostic categories, in virtue of their breadth, have the potential to alter the stories we tell about ourselves even more significantly than fine-grained categories such as PGD.³ In other words, the looping effects resulting from broadly defined diagnostic categories may be more harmful than those associated with specific diagnoses such as PGD. Thus, there may well be greater costs incurred by opposing the DSM’s inclusion of highly specific conditions (such as PGD) than those incurred by its inclusion.⁴

With regard to [2], it is crucial to note that the concept of medicalization, though not an interactive kind in Hacking’s sense, is interactive in a similar way. When we medicalize something, we treat it as a condition best managed with the expertise of medical or healthcare professionals. The looping effects of medicalizing a condition, such as complex grief, will thus depend on how we conceptualize the domain of medical and healthcare professionals. Similarly, what we end up medicalizing will shape our conception of the domain of medical and healthcare professionals. So, what we choose to medicalize and our conception of the domain of medical and healthcare professionals interact with each other. As our conception of the domain of medical and healthcare professionals changes which it can and does- so too will the looping effects associated with medical diagnoses. This raises the following question: how should we conceptualize the domain of medical and healthcare professionals if our goal is to minimize looping effects that alienate people from their experiences in detrimental ways?⁵

I will argue that a promising response to this is to adopt of a modified institutional framework⁶ for conceptualizing the domain of healthcare. Moreover, a case can be made that if an institutional conception of health is adopted then the inclusion of PGD as an official diagnosis in the DSM would do more than simply minimize adverse looping effects on self-conception: it would actually help to orient those diagnosed with PGD towards their grief in a positive way. This is because institutional frameworks can help grievers to identify with their grief and to see it as something that needs engaging with. In other words, with regard to healthy engagement with grief, institutional frameworks promote authentic engagement of exactly the kind that is needed. The exploration of [2] will thus have two important takeaways:

This paper develops these points as follows: Section 1 discusses the DSM-5 diagnostic criteria for PGD and the rationale behind adding a grief-specific disorder to the DSM; section 2 gives a more detailed account of Hacking’s picture of mental disorders as interactive kinds. Section 3 briefly articulates a few things about the nature of grief, explaining the distinction between state and process views. Section 4 explores the suggestion that medicalizing some forms of grief will lead to detrimental looping effects, in particular that it will lead those who are grieving the loss of a loved one to experience this grief less authentically, i.e. this section will explore concerns, raised by others, that the medicalization of grief might alienate grievers from their experiences by insulting the dignity of their lost relationships, reframing their experiences in impersonal medical terminology, and leading to feelings of “passive victimhood”; section 5 considers the flip side of this: will refraining from medicalizing prolonged or complicated grief, and consequently subsuming this highly specific category of mental distress under a broader conceptual category result in an even more significant negative impact on authenticity of experience?; section 6 engages with the question of how we might minimize the adverse looping effects inherent in medical diagnoses of psychiatric conditions, introducing Kukla’s institutional account of health conditions and drawing out the novel benefits of institutional frameworks for thinking about health conditions like PGD.

1. The DSM’s Decision to Identify Prolonged Grief Disorder as a Mental Disorder

As previously noted, the two predominant classificatory systems for delineating mental disorders and guiding clinicians in their diagnosis (the DSM and the ICD) have both recently incorporated a novel diagnostic category: Prolonged Grief Disorder. Although the two systems carve up the symptomatology of the disorder slightly differently, they both converge on the underlying idea that grief of a certain duration and intensity is a mental disorder in its own right and not merely a contributing causal factor in the development of other, more conventionally established, mental disorders. Given this underlying similarity, I will not lay out both sets of diagnostic criteria in depth. I shall instead focus on the diagnostic criteria of the DSM: their definition of “mental disorder”, how their approach to classifying symptoms of grief has changed in recent years, their description of the symptomatology of PGD and the motivations behind its inclusion in the DSM-5-TR.

The DSM, now in its fifth edition, serves as a guide to aide in the diagnosis of mental disorders. More specifically, the DSM-5 describes its purpose as: “fulfill[ing] the need of clinicians, patients, families, and researchers for a clear and concise description of each mental disorder”. In other words, the classifications outlined in the DSM are designed to be utilized by both medical professionals and the wider public. The DSM defines a mental disorder as:

a syndrome characterized by clinically significant disturbance in an individual’s cognition, emotion regulation, or behavior that reflects a dysfunction in the psychological, biological, or developmental processes underlying mental functioning. Mental disorders are usually associated with significant distress or disability in social, occupational, or other important activities. An expectable or culturally approved response to a common stressor or loss, such as the death of a loved one, is not a mental disorder. (^{American Psychiatric Association 2013a})

It is worth noting that whether one’s grief is considered a disorder will thus depend, in part, on whether one’s pattern of grieving matches up with the expectations and norms of their wider social context.

The previous iteration of the DSM, the DSM-IV, contained a “bereavement exclusion” for the diagnosis of Major Depressive Disorder (MDD) (^{American Psychiatric Association 2000}). This exclusion was removed with the publication of the DSM-5 in 2013.⁷ In 2022, the DSM-5 then underwent textual revisions which took the additional step of adding PGD as a diagnostic classification. The bereavement exclusion in the DSM-IV stated that those who have experienced a bereavement within the last 12 months are not eligible for a diagnosis of MDD. The rationale for adding the bereavement exclusion to MDD was based on empirical research observing significant overlap between symptoms of normal grief (“normal” in the sense that symptoms are typical of human distress reactions and typically subside on their own without psychiatric treatment) and symptoms of MDD (^{Clayton, Desmarais, and Winokur 1968}). The bereavement exclusion was added, in part, to address the concern that grieving individuals would otherwise meet the threshold for a diagnosis of MDD and be liable to misdiagnosis and inappropriate treatment for MDD.⁸ Typically, a diagnosis of MDD requires that an individual presents with five or more symptoms of the disorder. These symptoms include (among others) sadness, insomnia, fatigue, difficulty concentrating, decreased appetite, and loss of interest in usual activities all six of which are also common symptoms of normal grief. The overlap between symptoms of “normal” grief and MDD is so wide that in studies by Clayton et al., 42% of participants with normal grief reached the DSM’s symptom threshold for MDD (^{Hensley, and Clayton 2013}). Consequently, the bereavement exclusion precluded the diagnosis of MDD in individuals with depressive symptoms whose symptoms are “better accounted for by bereavement”. More precisely, a diagnosis of MDD was only indicated if the individual had not experienced a bereavement within the last two months or if their depressive symptoms were characterized by a subset of MDD symptoms that are not typically shared by grievers: marked functional impairment, morbid preoccupation with worthlessness, suicidal ideation, psychotic symptoms, or psychomotor retardation (^{American Psychiatric Association 2000, p. 356}).

Despite the significant overlap in symptoms between grief and MDD, the bereavement exclusion was not carried over from the DSM-IV to the DSM-5. The stated rationale behind this was that its inclusion problematically suggested that individuals in the midst of acute grief were immune to simultaneously experiencing MDD when in fact the two can co-occur. For instance, the APA noted that empirical and clinical evidence suggests that bereavement may precipitate major depression in people who are “especially vulnerable” and that MDD may lead to grief that is more severe and prolonged. Crucially, the removal of the bereavement exclusion was not intended to collapse the distinction between grief and depression, but, rather, to emphasize the possibility of their comorbidity. Instead of a bereavement exclusion, the DSM-IV contained notes cautioning clinicians to differentiate between normal grieving associated with a significant loss and a diagnosis of a mental disorder. Although the intent was not to collapse the distinction, opponents of the move were concerned (i) that it would lead to diagnostic inflation (ii) that grieving patients would be misdiagnosed with MDD and (iii) that it was part of an increasing trend towards an over-pathologization of the human experience (^{Cacciatore and Frances 2022}).

In 2022 the DSM’s stance on grief changed again, with the introduction of PGD in the DSM-5-TR. The decision to introduce PGD as a novel diagnosis was supported by research indicating that people struggling with symptoms of depression in addition to a bereavement do not respond to standard depression treatments as well as those experiencing depression in the absence of bereavement do. Additional research indicates that grief-specific treatments are significantly more effective than standard depression treatments for those who fall into this grouping,⁹ which suggests that there would be benefits associated with creating a specific diagnostic category of PGD.

The inclusion of PGD in official diagnostic guidelines is controversial. One reason for this is that grief is a normal part of the human experience. Consequently, there are concerns that medicalizing certain forms of grief risks pathologizing a normal, and perhaps fundamental, aspect of our lives as human beings. The move to extend the duration of symptoms necessary for diagnosis from six months to a year is a response to this concern. There are at least two substantive underlying assumptions in play here:

These assumptions deserve highlighting as they should not slide by unquestioned. I will later argue that assumption (i) arises from a misguided, and unhelpful, understanding of the concept of a health condition (a category of which mental disorders are a subset). If we shift to a more apt institutional- conceptualization of health conditions, then these concerns about medicalizing normal aspects of the human experience have no teeth.¹¹

2. Mental Disorders as Interactive Kinds

Ian Hacking introduces the notion of an interactive kind: classifications in which the classification schema may interact with the thing classified and vice versa. A classification is interactive in the relevant sense when the act of classification impacts how those who are classified experience themselves and, in turn, how others perceive of them and treat them.

Consequently, this kind of interaction is possible when and only when the entities being classified are self-aware in the sense that they are capable of acting under descriptions. In other words, interactive kinds are classifications that apply to people, and not inanimate objects, because people are the only entities capable of conceptualizing themselves under different descriptive classifications. Another way of putting this is to say that we experience ourselves in the world as being persons of various kinds (^{Hacking 1999, p. 103}). For example, I act under the description “daughter” when I give my mother a card on Mother’s Day. If I did not think of myself as a daughter (or, more generally, as someone with a mother) then I would not perform this action. My action depends on my conceptualizing myself in a certain way and on my having a particular conception of how daughters act. Hacking summarizes the idea thus:

We are especially concerned with classifications that, when known by people or by those around them, and put to work in institutions, change the ways in which individuals experience themselves-and may even lead people to evolve their feelings and behavior in part because they are so classified. (1999 p. 104; my emphasis.)

The interactive nature of interactive kinds is not just a function of individuals changing their self-conception in response to learning that they have been classified as belonging to a particular kind. Classification also impacts (i) how others perceive those who have been classified and (ii) how institutions and practices pertaining to these classifications develop and change over time interactions occur in “the larger matrix of institutions and practices” built around our classifications (1999). For instance: I would not buy my mother a card on May 14 ^th if we did not have the publicly recognized occasion that is Mother’s Day. My understanding of myself as a daughter and my beliefs about the actions that a good daughter should perform are shaped by both my understanding of the classification “daughter” and the institutions and practices built around it (and vice versa).

To say that mental disorders are interactive kinds is to say that our understanding of these disorders interacts with, and changes, both the expression of the disorders diagnosed, who counts as having them, and how people (including the diagnosed) perceive those diagnosed. These interactions are referred to as “looping effects”.

In a recent New York Times article, adapted from parts of her book Strangers to Ourselves: Unsettled Minds and the Stories That Make Us, Rachel Aviv illustrates several looping effects associated with the diagnosis of mental disorders ⁽²⁰²²⁾. Taking a brief look at one of these examples provides some initial motivation for the claim that psychiatric disorders are interactive kinds and clarifies the phenomenon of looping effects. This example is not intended to constitute an exhaustive or comprehensive illustration of all looping effects associated with psychiatric disorder classifications. Rather, it is intended to bolster the plausibility of the claim that psychiatric disorders are interactive kinds.

Aviv recounts that when new patients at a psychosis clinic began to learn about the definition of their diagnosis, the language they used to describe their experiences changed: “expert explanations replaced their idiosyncratic attempts to make sense of the world”. This process may have some benefits in that it furnishes patients with concepts that may help them to make their experiences intelligible to others, but it also has drawbacks. Aviv notes the following experience of a patient at the clinic who had been newly diagnosed with schizophrenia:

[She] studied the definition of schizophrenia in the Diagnostic and Statistical Manual of Mental Disorders and, over time, worried that she was inadvertently bending her own behavior to better fit the bounds of that category. She became self-conscious about the experience of thinking and began to wonder if she heard a soft voice behind each thought, until she felt she was hearing voices, a symptom of the disorder. She was no longer sure what was her authentic experience and what had been suggested to her by experts. (2022; my emphasis.)

In this example, the looping effects appear to change not just the patient’s self-conception but also the expression of her disorder (in terms of the symptoms she experiences changing).

More precisely: being given a psychiatric diagnosis and learning about how that diagnosis is understood by the medical profession can shape how patients describe the symptoms they are experiencing. This then alters how they conceptualize and relate to their experiences.

Importantly, the claim that mental disorder classifications are interactive kinds is, in one sense, neutral with regard to the ontology of the mental disorders.¹² For example, saying that the classification “depression” is an interactive kind is compatible with both pure social constructionist and biological “natural kind” accounts of the underlying mechanisms that occur with depression. This is because it is merely our concept of depression that is an interactive kind any underlying biological state/s that the classification picks out need not be similarly interactive. Consequently, my subsequent appeals to the potential looping effects of the classifications of depression and PGD do not commit me to a social constructivist position regarding the underlying causal mechanisms that correspond with these classifications.

3. Grief as a Process That Unfolds Over Time

Before investigating the potential looping effects of medicalizing grief, we first need a basic picture of what helpful grief narratives, or healthy engagement with grief, might involve.

Philosophical accounts of grief are divided in their understanding of the phenomenon. On one hand, we have accounts claiming that grief is a particular kind of mental state or mental event (either a non-cognitive feeling, or a cognitive state of some kind). On the other hand, there are views which describe grief as a multifaceted process that unfolds over time.¹³ I will follow in Peter Goldie’s footsteps and appeal to the following extract from C.S. Lewis’s A Grief Observed, in which Lewis characterizes what it is like to grieve, to motivate and support the intuitive plausibility of process views:

In so far as this record was a defence against total collapse, a safety-valve, it has done some good. The other end I had in view turns out to have been based on a misunderstanding. I thought I could describe a state; make a map of sorrow. Sorrow, however, turns out to be not a state but a process. It needs not a map but a history, and if I don’t stop writing that history at some arbitrary point, there’s no reason why I should ever stop. There is something new to be chronicled every day. Grief is like a long valley, a winding valley where any bend may reveal a totally new landscape. As I’ve already noted, not every bend does. Sometimes the surprise is the opposite one; you are presented with exactly the same sort of country you thought you had left behind miles ago. That is when you wonder whether the valley isn’t a circular trench. But it isn’t. There are partial recurrences, but the sequence doesn’t repeat. ^{(1961, p. 50)}

While I think that process views of grief are more plausible than mental state/event views, my goal is not to argue decisively for this here (nor do the points I make depend on taking a particular stance). That said, the points I raise in this paper do align better with, and lend support to, process views.¹⁴ If we view grief as a process, then a diagnosis of PGD suggests that grievers are struggling with aspects of the complex interplay of emotions, behaviors and actions that occur in the wake of a bereavement. On the other hand, if we view grief as a mental state, then this suggests that those with PGD are struggling in virtue of the state that they are in. Therapeutic interventions designed to support those struggling with PGD (such as PGDT, discussed in section 6) are often designed to help grievers engage differently with their experiences of loss. This suggests that those experiencing PGD may be struggling, in part, due to features of the way that their grief is unfolding over time (in particular, due to the way/s that they have been engaging with their experiences of loss). This insight aligns better with process views which, unlike state views, prioritize the dynamics of grief. Relatedly, whether one views grief as a process or a state may impact the looping effects associated with PGD. For example, if someone views grief as a state rather than a process then a diagnosis of PGD may render her more vulnerable to experiencing the harmful feelings of passivity discussed in section 4.

The key takeaway here is that, regardless of whether one ultimately adopts a process or state view, once we start thinking about the relationship between classifications and their looping effects then we can start to see that there are important consequences associated with how we conceptualize grief.

4. Medicalizing Prolonged Grief and the Potential for Detrimental Looping Effects

Michael Cholbi delineates at least three potential adverse looping effects that may arise from the medicalization of grief ^{(2021, p. 182)}:

These potential looping effects share a common feature: each raises the concern that medicalizing prolonged grief will negatively impact grievers’ understanding of the relevant loss. In other words, the worry is that medicalizing prolonged grief will problematically distort the narratives that those diagnosed with PGD construct around their experiences. Or, for those who may be opposed to framing the worry in narrative terms: medicalizing grief may impact grievers understanding of what has been lost in a way that adversely affects their ability to engage with its value. If someone’s attempts to understand the significance and value of her losses are impeded, then this can make it harder for her to engage with these losses appropriately. By way of analogy: if I think that the reason my car won’t start is that it has a dead battery, when in fact it is out of gas, then directing my energy towards replacing the battery won’t be helpful.

Cholbi concludes that these potential looping effects give us reason to oppose the inclusion of a grief specific mental disorder in the DSM ^{(2021, p. 184)}. Others opposed to the inclusion of PGD in the DSM have suggested another potential adverse looping effect on patient self-conception:

1. Diagnosing grievers with a mental disorder “insult[s] the dignity of loving relationships”. (^{Cacciatore and Frances 2022})

This criticism echoes Cholbi’s concern that being diagnosed with PGD might distort narratives surrounding grief in an undesirable way. In this case, the purported distortion seems to be that viewing prolonged grief as a medical condition might lead to grief narratives that inaccurately portray the significance of a griever’s relationship with their lost loved one.¹⁵ I think that this concern is just one aspect of a larger worry: it would not just be the significance of the lost relationship that is jeopardized, but, more importantly, the significance of the loss of the person themselves.

Although Goldie does not explicitly address the issue of medicalizing grief, his narrative process account of grief is congenial to similar concerns. Goldie emphasizes that the way in which one narrates one’s grief matters. For instance, he cites empirical research indicating that impersonal narrating a mode of narration characteristic of those with post-traumatic stress- can prevent people from being able to evaluate and respond in an emotionally appropriate way to past events (^{Goldie 2012, pp. 70-72}, citing ^{Barclay 1995, p. 113}; ^{Eich et al. 2011}, and ^{Conway 2003, p. 218}). Cholbi’s concern that medicalizing grief may lead individuals to conceptualize and describe their experiences in the language of clinical psychiatry, rather than in their own words, and that this may stymie their ability to adapt over time, seems particularly salient here.

5. PGD And Looping Effects: The (Positive) Flipside

In this section I motivate the following argument:

P1. If the DSM removes the classification of a grief-specific disorder, then people experiencing complex grief will either (i) be diagnosed under a different, broader, classification (such as depression, attachment disorders, or PTSD) instead, or (ii) not be diagnosed at all.

P2. A diagnosis of depression, attachment disorder, or PTSD has greater potential to harmfully distort the self-conception of individuals primarily experiencing prolonged grief than a diagnosis of PGD does.

C1. If the removal of PGD from the DSM leads to an increase in the diagnosis of grievers under broader diagnostic classifications, then the removal of PGD from the DSM has greater potential to harmfully distort the self-conception of those experiencing prolonged grief than its inclusion.

P3. If someone is struggling with prolonged grief and is not diagnosed with PGD or a similar grief-related disorder, then this will deprive them of assistance that could be of benefit to them.

C2. If the removal of PGD from the DSM results in no diagnosis for people struggling with prolonged grief, then its removal will deprive these grievers of assistance that could be of benefit to them.

Via P1, C1, and C2:

C3. For those struggling with prolonged grief, the removal of PGD from the DSM would either lead to a greater potential for harmful distortions of self-conception or it would deprive them of access to assistance that could be of benefit to them.

If PGD is removed from the DSM, then there are only two possible alternative diagnostic scenarios: either grievers will be diagnosed under a different classification, or they will not be diagnosed at all. P1 is thus un-controversially true. Which of (i) or (ii) is the more likely outcome is an open question, but the suggestion that if PGD is removed from the DSM then people experiencing complex grief are more likely to be classified under broader diagnostic categories is plausible for several reasons. Firstly, as noted in section 1, there is a large overlap between symptoms of prolonged or complicated grief and MDD. There is also an overlap between symptoms of PGD and PTSD, “the main differential diagnostic considerations for complex grief include normal acute grief and major depression, and, if the death is violent, PTSD” (^{Shear 2012, p. 123}). Interestingly, proponents of removing PGD from the DSM have suggested that patients struggling with prolonged or distressing grief symptoms who would benefit from psychiatric care are better served by the broader diagnostic classification of “attachment disorder” (^{Cacciatore and Frances 2022}). Consequently, it is at least plausible to think that, in the absence of a grief-specific diagnostic category, some grievers will be more likely to be diagnosed under broader diagnostic classifications. But, even if one is not convinced of this, the alternative is that grievers will not be diagnosed at all and, as we shall see in subsection 5.1 and subsection 5.2, both options lead to the harmful looping effects suggested in P2 and P3.

6. Kukla’s Institutional Account of Health Conditions

Kukla proposes that we understand “health” as a special sort of institutional concept ⁽²⁰¹⁴⁾. Institutional concepts are concepts constrained by both the world and our social practices. Kukla offers paradigmatic examples such as “paycheck”, “voting”, “convict” and “student”. These concepts are constrained by both the world and our social practices because whether something belongs to one of these conceptual categories depends on whether, and if so how, it is embedded in social institutions we cannot simply choose to classify things under a particular institutional concept. For example, if the sheet of paper I am holding is not endorsed by an employer and would not be accepted for deposit in a bank, then it is not a paycheck regardless of whether someone chooses to refer to it as such. As Kukla puts it:

The existence of such things [institutional concepts] is thoroughly dependent upon elaborate social institutions, and to be such a thing is to be embedded in these institutions in the right way. You can’t be a convict without a legislative, justice, and penal system. Nothing counts as a pay-check without elaborate labor and economic institutions. Things don’t become or cease to be convicts or paychecks just because we choose to classify or declassify them in that way. Being either one has definite empirical consequences and preconditions. We may slowly refine or shift these kinds in accordance with our social needs. But we cannot simply discover that we were totally wrong about what a convict or a paycheck is, since our practices carved these kinds out. (2014, p. 525)

Viewing health as an institutional concept is central to Kukla’s Institutional Definition of Health:

A condition or state counts as a health condition if and only if, given our resources and situation, it would be best for our “collective” wellbeing if it were medicalized-that is, if health professionals and institutions played a substantial role in understanding, identifying, managing and/or mitigating it. In turn, health is a relative absence of health conditions (and concomitantly a relative lack of dependence upon the institutions of medicine). (2014, p. 526)

Kukla’s account of health conditions differs to pure social constructionist accounts because it is normative rather than merely descriptive. Social constructionist accounts define health conditions in terms of what is, or has been, medicalized by health institutions. Kukla’s account, in contrast, defines health conditions in terms of what should be medicalized, where what should be medicalized is what is in fact conducive of collective wellbeing given the social and natural facts. What should be medicalized is thus something we discover and something that we can be wrong about.

In this paper I refer to “institutional frameworks” for conceptualizing health conditions rather than Kukla’s account per se because I do not want to endorse all of the specifics of their account. In particular, the idea that we should determine what counts as a health condition by appealing to whatever it would be in our best interests to medicalize is promising, but I think that we must first have a robust conception of the health of an individual in order to identify and understand what our best interests are. We should not understand health itself institutionally, or simply in terms of an absence of health conditions. I take up this point in greater depth elsewhere, but two important upshots of this kind of modified institutional account that are worth mentioning here.

First, by allowing for a robust understanding of health per se, we open up conceptual space for the recognition of an important category of well-being that gets overlooked on Kukla’s account. People can experience bodily or psychological difficulties that are not recognized as health conditions on institutional accounts (because we do not yet and perhaps may never- have the resources or knowledge to treat them), but it does not follow from this that these individuals should be deemed to be in a state of good health. The modified account, on the other hand, enables us to capture this tripartite distinction.

Second, even though grieving often involves tremendous suffering and can, in some cases, be considered a health condition, it does not follow from this that a life of good health is a life without grief or even a life without PGD. Grief, in all its manifestations, plays an important role in the narrative of our lives. It reflects our attempts to do justice to the value and enormity of what has been lost. “Prolonged” grief in response to a loss may well be, in some cases, a constitutive element of a person’s flourishing. Thus, although prima facie counter intuitive, it is plausible to propose that a proper understanding of being in good health should accommodate, and perhaps even require us to experience, health conditions (when health conditions are conceived of institutionally). Kukla’s account cannot accommodate this, whereas the modified account that I suggest can.

Additionally, I do not want to commit myself to Kukla’s claim that what we should medicalize is what is best for our collective wellbeing. The central idea of Kukla’s account, that “real health conditions are conditions for which the tools and methods and support of medicine and its institutional mechanisms are genuinely helpful, given both the natural and the social facts” (2014, p. 525; my emphasis) is what I’m interested in utilizing here. For the time being, I remain neutral as to what being “genuinely helpful” might mean.

An important benefit of institutional frameworks is that what we medicalize need not correspond to biological “dysfunctions” or abnormalities (conceived of as scientistic natural kinds demarcating diseases). Institutional frameworks make room for treating diagnostic classifications, such as PGD, merely as identifiers of categories that are determined by contingent empirical facts contingent facts about whether unifying particular groupings of symptoms under a diagnostic classification as a health condition will be helpful,¹⁸ rather than tying health conditions to natural kinds. Recall that looping effect 1 (seeing oneself as a passive victim) relies on an underlying view of health conditions, or disorders, as essentially external phenomena acting on us, or as primarily biologically determined pathologies or dysfunctions. In the previous section I suggested that a more nuanced picture of disorder may make room for us to think of a diagnosis like PGD as nothing more than an official recognition that we are struggling with our grief, that this struggle needs addressing, and that we are identifying it as a disorder precisely because this enables relevant professionals to help us engage with it in different ways. If we conceptualize disorders, or health conditions, using an institutional framework then we end up with exactly this kind of picture. This is a novel virtue of institutional frameworks.

Moreover, if one’s understanding of health conditions reflects the assumption that we should medicalize things that the tools of medicine are poised to help with, then our understanding of the treatment options will also shape our understanding of the disorder. Prolonged Grief Disorder Therapy (PGDT) is the most extensively tested treatment that exclusively targets PGD with demonstrated efficacy (^{The Center for Prolonged Grief}). PGDT is a talking therapy designed to help grievers “get to know grief, manage strong emotions, think about the future, rebuild strong relationships, think about the death, revisit reminders of the loss and access living memories” (^{The Center for Prolonged Grief}). In other words, PGDT is designed to help grievers engage differently with their grief. Consequently, if one understands that PGD is a health condition precisely because medical professionals can help grievers engage with their grief then a diagnosis does exactly the opposite of encouraging grievers to view themselves as passive victims. Understanding health conditions as institutional concepts thus has the potential to not only reduce concerns relating to various potential adverse looping effects but also to foster positive looping effects by encouraging active engagement of exactly the kind that opponents of this new diagnostic classification value.¹⁹

Considering looping effect 2 can help us to see yet another beneficial impact of institutional frameworks. Recall looping effect 2: medicalizing grief may lead individuals to identify themselves with their grief, labelling themselves as “grievers” in a way that assigns grief a lasting part in their identity. In section 5, I argued that this looping effect is a problem regardless of whether we include PGD in the DSM, but that it is less problematic when we do have a grief-specific diagnostic category. It is plausible that the impact of looping effect 2 may also be reduced by adopting an understanding of health conditions as institutional concepts rather than biological or natural kinds. Recall that the contingency of diagnostic classifications is a central feature of institutional frameworks. Diagnostic classifications identify categories that are determined by ever changing empirical facts about whether the tools of medicine would be beneficially employed in aiding the constellation of symptoms unified under the classification. This renders the classification something malleable: it is not something wholly determined by the biology or psychology of the person diagnosed. If this malleability is part of patients’ understanding of PGD when they are diagnosed, then it will inform the shift in self-conception that occurs with the diagnosis. Additionally, it seems plausible that an awareness of this malleability might make the shift in self-conception less likely to involve permanently labelling oneself as a “griever” in a way that assigns grief a lasting role in one’s identity. This is because it would influence their understanding of the reason for their diagnosis: namely, it would highlight that the reason why they have a diagnosis of PGD is simply that this diagnosis serves a useful functional role enabling them to access support to help them engage with their experience of loss in new ways.

7. Summary

Opponents of including PGD in the DSM have been concerned with the potential for harmful looping effects, suggesting that the diagnosis might adversely impact grievers’ self-understanding and ability to navigate their experience of grief. My project in this paper has been to call attention to beneficial looping effects that might be achieved by medicalizing (some) experiences of grief, which in turn provide heavyweight reasons in favor of medicalization that have been overlooked in recent discussions. The aforementioned concerns regarding harmful looping effects are attached to a background conceptualization of health conditions as pathologies, rather than to the fact of diagnostic classification itself. Shifting to an institutional framework for conceptualizing health conditions would be a potential mechanism for equipping patients to engage more authentically with medicalized conditions affording grievers access to the conceptual and practical resources attached to institutional recognition without the costs attached to a background conceptualization of health conditions as pathologies.