Beyond medicalization: Self-injuring acts revisited

Terminology in research on self-injuring acts

A plethora of terminologies, definitions, and inclusion criteria are used in research on self-injuring acts. The two most common terms are Non-Suicidal Self-Injury (NSSI) (e.g. Klonsky, 2007; Nock et al., 2006; Swannell et al., 2014; Zetterqvist et al., 2013) and Deliberate Self-Harm (DSH) (e.g. Bjärehed and Lundh, 2008; Hawton et al., 2002; Madge et al., 2008; Tsai et al., 2011; Ystgaard et al., 2009). NSSI is defined as intentional self-inflicted damage to body tissue without suicidal intent, while DSH also includes other self-destructive acts such as jumping from a height or ingesting a recreational or illicit drug, as well as suicide attempts. DSH is mainly used in the United Kingdom and some other European countries, but this term, especially when used in the United States, can hold the same definition and inclusion criteria as NSSI. Other terms are self-mutilation (e.g. Favazza, 1998; Rissanen et al., 2011), parasuicide (e.g. Crowell et al., 2005), self-inflicted injury (SII) (Crowell et al., 2005), self-injurious behavior (SIB) (e.g. Kleiman et al., 2014), direct self-injurious behavior (D-SIB) (e.g. Brunner et al., 2014), and non-suicidal deliberate self-harm (e.g. Brown, 2009).

The common denominator in these different terms is that they describe self-injuring as an intentional act. In this article, the term self-injuring acts is used when addressing the phenomenon and the literature on the subject.

Medicalization, mental suffering, and context

Medicalization consists of defining a problem in medical terms, using medical language to describe a problem, adopting a medical framework to understand a problem, and using medical intervention to handle the problem (Conrad, 1992). The medicalization process is bidirectional in the sense that a phenomenon can be both medicalized and demedicalized (Conrad, 2007). Conrad emphasizes medicalization to be a value-free process, but when non-medical issues come under medical influence or dominance, and people’s reactions to everyday situations are described with a medical vocabulary, there is a risk that ordinary reactions to stressful events and life situations are referred to the medical profession and regarded as too difficult and too complicated to be handled by the individual’s social network. Another consequence could be that medicalization undermines people seeing themselves as subjects (Parens, 2013). A diagnostic vocabulary and a focus on diagnoses and medical aspects on human reactions risk to objectify the individual and explain away human suffering (Wilkinson, 2005). An example of this is how the psychiatric profession has reclassified normal human sorrow and sadness for example due to the loss of a loved one, into largely an abnormal experience by not taking into account the context in which the symptoms occur (Horwitz and Wakefield, 2007).

Emotional responses to stressful and painful life events are affected by a myriad of factors, and Smith (2011) highlights how the individual’s interpretation of mental pain is relative to a range of interpersonal and social factors. Mental suffering, Smith concludes, is too complicated to be captured in a diagnostic system of medical or psychiatric diseases, as the individual is situated in a social world and mutually affected by the micro- as well as the macro level of society. Describing how mental suffering came to be treated as a medical problem, and what barriers these psychiatric constructs have produced, Smith calls attention to the importance of a multidimensional analysis. These perspectives emphasize the necessity to address the relationship between mental suffering and the larger social context, as it is impossible to capture the essence of suffering if it is only considered as an intrapersonal problem.

In contrast to such perspectives, the present tendency in society is to move toward a medical understanding and interpretation of a range of human conditions (Conrad, 1992, 2005, 2007). An example of this is how attention deficit hyperactivity disorder (ADHD) has been made a psychiatric diagnosis within a few decades. Conrad and Bergey (2014) describe this process as a striking example of how complex social forces, including pharmaceutical innovations and marketing, play a significant role in the medicalization process. According to Conrad, the search for biological factors, possible to manage with drugs, is turning the medicalization process into a process of biomedicalization. A biomedical perspective emphasizes presumed individual deficiencies and pathologies and aims at correcting or changing the individual. Such a perspective on human suffering obfuscates the impact of social, cultural, and gendered structures and expectations.

The female body, with its physiological and hormonal processes, as well as women’s traditional social roles, have made women more vulnerable than men to the medicalization process (Riessman, 1983), which indicates that the medicalization process is influenced by gender (Russell, 1995). Discussing how psychiatry historically has defined women’s emotions and reactions as signs of hysteria, personality disorders, and depression, Chesler (1972), Russell (1995), and Ussher (2011) point to how medicalization of women’s minds and bodies has been an ongoing process from the onset of psychiatric research, and how women who opposed or challenged social norms faced the risk of being diagnosed as mentally ill.

Production of scientific knowledge

To understand the medicalization of self-injuring acts, the context in which the scientific knowledge has been produced has to be examined. The production of knowledge is irrefutably intertwined with ideological aspects, theoretical foundations, academic disciplines, and cultural, societal, and political structures (Bacchi, 2009; Haraway, 1991; Harding, 1986; Mannheim, 1976). Science, Harding (1986) declares, whether natural or social, is produced in a context not only motivated by scientific incitements but also governed by political and economic interests.

Attention to a problem is also intertwined with power relations that allow particular problem representations to emerge and gain status. Presuppositions and assumptions behind a problem formulation include some aspects of the problem and exclude other aspects to reduce complexity. Consequently, science and its methods not only describe but also create and produce realities, and scientific investigations interfere with the world, and always make a difference, politically or otherwise (Law, 2004). It is therefore critically important to examine the problem representations to see where and how they have been produced and how they have been perpetuated, in order to detect other important aspects (Bacchi, 2009).

Science production must be understood as situated in the local context where it is created (Haraway, 1991). Within a particular context, a doctrine or a canon is often formed with respect to acceptable and correct research methods (Mannheim, 1976). When a scientific paradigm has been established, further research within the paradigm is undertaken within what Kuhn (1962) calls “normal science.” Knowledge created within the paradigm is continuously reproduced, while problem formulations that are not within the borders of the paradigm are rejected.

The theories, concepts, and vocabulary central to a specific field of research hamper the possibility to see beyond the borders of the researcher’s own scientific area (Longhofer and Floersch, 2012). A field of research can therefore be perceived as a self-referring, autopoiesic system (Luhmann, 2003 [1985]), where theories and concepts can be described as communicative elements that are based upon earlier elements. In such an autopoiesic system, stable ideas, structures, and processes are continually reinforced once they have been created. New empirical observations are both based upon and related to earlier observations. Within an autopoiesic system, ideas, structures, and processes may resist change, as the borders of the system can be rigid and difficult to surpass. The status of a system and its members influence the openness towards new inputs and experiences outside of its borders (Luhmann, 2003 [1985]).

The early stages

In the early clinical research, the phenomenon was predominantly studied among women within psychiatric settings, and these women were ascribed certain mental and emotional characteristics. The typical self-injurer was described as a young, attractive, intelligent, and on the surface socially adapted woman who was periodically overwhelmed by inner emotional tensions (Grunebaum and Klerman, 1967). Graff and Mallin (1967) interpreted self-injuring acts among women at a psychiatric ward as an expression of psychopathology and argued that women who self-injure were either promiscuous or overly afraid of sex and unable to relate successfully to others.

Self-injuring acts were also associated with female bodily functions. Rosenthal et al. (1972) related wrist cutting to menstruation, as 60 percent of the women in their study reported that they cut themselves during this period. “Even more strikingly, almost half of the cutters had always had irregular menstrual periods and frequent amenorrhea” (Rosenthal et al., 1972: 1363). Wrist cutting, they concluded, could be understood as a means of dealing with genital trauma and conflict centering on menstruation. Pao (1969), in discussing the characteristics of psychiatric patients who only made superficial, delicate, and careful incisions, assumed that some of the women, though strikingly feminine looking, had lived through a prolonged time as tomboys. The few male “delicate cutters” were described as “pretty boys” and quite effeminate.

As self-injuring acts were associated with psychopathological issues, a discussion on the feasibility of making it a separate diagnosis was initiated. Based on a meta-analysis of 56 different case studies, Pattison and Kahan (1983) argued that self-injuring acts met the three essential features of disorders of impulse control in the Diagnostic and Statistical Manual (DSM) III: (1) failure to resist an impulse, (2) increasing tension before committing the act, and (3) experience of pleasure, gratification, or release at the time of committing the act. They argued that self-injuring acts should be considered for inclusion in the DSM IV, and that this diagnosis should be defined as characterized by severe, uncontrollable impulses, major self-mutilation, a low level of lethality, onset in late adolescence, and repetitive episodes over the years. This discussion on diagnostic classification still continues, as described later in this article.

Expansion of the field

By the end of the 1990s, the research field expanded from psychiatric settings to studies on prevalence in the community. Still, research remained within the realms of psychiatry and psychology, regarding self-injuring acts as an individual health problem within the community, using the traditional medical methodology of survey studies. One of the first studies investigating the magnitude of self-injuring acts among adolescents outside of clinical settings was published by Hawton et al. (2002). Since then, a large number of prevalence studies on non-clinical populations have been conducted in most Western countries (e.g. Baetens et al., 2011a; Bjärehed and Lundh, 2008; Brunner et al., 2007; Cerutti et al., 2011; Claes et al., 2014; Harriss and Hawton, 2011; Hawton et al., 2002; Klonsky, 2011; Laukkanen et al., 2009; Laye-Gindhu and Schonert-Reichel, 2005; Lundh et al., 2007; Mohl and Skandsen, 2011; Moran et al., 2012; Muehlenkamp et al., 2009; Nixon et al., 2008; Nock et al., 2006; O’Connor et al., 2009; Plener et al., 2009; Ystgaard et al., 2009; Zetterqvist et al., 2013). The prevalence of self-injuring acts has also been investigated in Asian countries (e.g. Hanania et al., 2014; Kharsati and Bhola, 2015; Matsumoto et al., 2008; Shin et al., 2009; Tsai et al., 2011; Wan et al., 2011; You et al., 2011).

As there is no consensus on what acts should be included in studies on self-injuring acts, the prevalence rates vary. In the Asian studies, the prevalence rates of self-injuring acts varied between 9.2 percent and 31.2 percent. In an overview on international prevalence of NSSI and DSH, Muehlenkamp et al. (2012) found rates ranging from approximately 5 percent to 37 percent in adolescent community samples.

In a comparative study on findings of regional studies on adolescent self-injury in 11 different European countries, Brunner et al. (2014) found that life-time prevalence of D-SIB ranged from 17.1 percent to 38 percent across the included countries. They also found that female gender was significantly associated with higher rates of self-injuring acts in most countries. A questionnaire study covering over 30,000 mainly 15- and 16-year-old Australian and European school pupils (Madge et al., 2008) found that DSH was twice as common among girls than among boys.

In a meta-analysis of gender differences of NSSI, Bresin and Schoenleber (2015) found that the rates of NSSI in clinical samples varied from 11.8 percent to 84 percent, while in community and college samples the rates varied from 14 percent to 24 percent. This meta-analysis found the literature to be inconsistent with respect to gender differences, as some studies reported significantly higher rates among women while others found no differences. Such differences might be due to different inclusion criteria in the studies but may also be related to differing self-injuring methods between men and women (e.g. Posick et al., 2013). As results are heterogeneous, arguments for developing common definitions and standardized methodologies and screening instruments in research on self-injuring acts have been raised by among others Borschmann et al. (2012) and Swannell et al. (2014).

Searching for medical and gendered explanations

Although the prevalence studies reveal self-injuring acts to be a common phenomenon in the community, only a small minority of those who self-injure are known in clinical settings (e.g. Baetens et al., 2011a; Fortune et al., 2008; Hawton et al., 2002, 2012; Michelmore and Hindley, 2012; Morey et al., 2008; Ystgaard et al., 2009). Still, the medical view on self-injuring acts is perpetuated in the substantial research that has been undertaken to explore an assumed intrapersonal basis for self-injuring acts (e.g. Baetens et al., 2011b; Bjärehed et al., 2012; Cerutti et al., 2011; Nock et al., 2006; Victor and Klonsky, 2013).

Prevalence studies on self-injuring acts are often designed to investigate gender differences. Such studies focus on if emotions are related to sex and gender stereotypes (e.g. Straiton et al., 2012), if character traits differ between men and women (e.g. Sornberger et al., 2012), whether temperament differs between men and women (e.g. Baetens et al., 2011b), and if such factors have a positive or negative effect on the risk for self-destructiveness (e.g. Kleiman et al., 2014). These studies generally conclude that women handle anxiety, depressed mood, despair, and other affective states diametrically different than men. Such studies have evoked reactions by some researchers who claim that the focus on girls and women have concealed the fact that boys and men also self-injure (e.g. Inckle, 2014; Taylor, 2003) but might do so for different reasons and motives and by using different self-injury methods than girls and women (e.g. Russell et al., 2010).

There is a robust consensus that self-injuring acts are used to handle mental suffering and to regulate negative and unwanted emotions (e.g. Bentley et al., 2015; Jenkins and Schmitz, 2012; Kleiman et al., 2014; Klonsky and Muehlenkamp, 2007; Nock et al., 2007; Victor and Klonsky, 2013; Voon et al., 2014), but some efforts have been undertaken to find biological factors, such as attenuated serotonin levels, differences in endogenous opioid functioning, and altered cortical neural patterns behind self-injuring acts (e.g. Crowell et al., 2005, 2008; Plener et al., 2012; Stanley et al., 2010).

Toward a psychiatric diagnosis

A large proportion of research on self-injuring acts has examined the correlation between self-injuring acts, gender, and psychopathology, for example borderline personality disorder (BPD) (e.g. Bjärehed and Lundh, 2008; Kleiman et al., 2014; Lyons-Ruth et al., 2013; Nock et al., 2006; Selby et al., 2012) or other emotional disorders (e.g. Bentley et al., 2015). Self-injuring acts is one of the nine criteria for BPD (DSM-IV-TR, 2000), and BPD is a diagnosis primarily given to women, with a gender preponderance of 70 percent to 77 percent (e.g. Bjorklund, 2006). As people who self-injure do not always fulfill the additional criteria required for a diagnosis of BPD, NSSI has been proposed as a separate diagnosis. In a large randomized community sample of 3060 adolescents (50.5% female), Zetterqvist et al. (2013) found that 83 percent of those who fulfilled the criteria for NSSI were girls.

In-Albon et al. (2013) and Wilkinson (2013) have argued that an NSSI diagnosis should facilitate the communication between clinicians and those who self-injure.

Examining the interaction between women and staff at a psychiatric ward, Lindgren et al. (2011) found the communication to be characterized by a contest on the preferential right of interpretation. People who self-injure often describe encounters with professionals as characterized by misunderstandings, as the medical and nursing professionals perceive their self-injuring acts as irrational and illogical (e.g. Harris, 2000; Lindgren et al., 2011; Taylor, 2003; Taylor et al., 2009). Studies based on interviews with health-care professionals indicate that they find it difficult to encounter and understand people who self-injure (e.g. McHale and Felton, 2010; Patterson et al., 2007; Sandy, 2013). This begs the question whether an NSSI diagnosis would facilitate or hamper the communication.

The medical prerogative—consequences for the research approach

The incapacity to investigate a phenomenon outside the limits of the researcher’s own scientific paradigm has been stated as a problem by among others Kuhn (1962), Mannheim (1976), Luhmann (2003 [1985]) and Longhofer and Floersch (2012), who argue that established and accepted theories, concepts and methods within a paradigm continue to both guide and limit further research. In the words of Bacchi (2009), assumptions and presuppositions behind a problem formulation include certain aspects and exclude others. The continued assumption of intrapersonal and psychopathologic factors as the basic foundation for self-injuring acts, and the absence of reflection on whether this assumption is compatible with the magnitude of self-injuring acts in the community, shows the characteristics of what Luhmann would call an autopoiesic system.

Research on self-injuring acts was originally limited to interviews with psychiatric patients, forming the early view of the typical self-injurer as a woman with psychiatric problems. The prerogative of the medical field to decide who is ill and who is healthy, what is normal and what is deviant makes it difficult to challenge the medical perspective with its high status and power. An example of this is that when psychological behavioral science entered the research field of self-injuring acts, the medical view was not challenged, but instead the intrapersonal and psychopathologic view was reinforced in the search for emotional deviation, different characteristics, and abnormal ways of functioning among those who self-injure compared to those who do not. The medical view has recently been further reinforced by attempts to find biological explanations for self-injuring acts. Such a movement toward a biomedicalization process further distances self-injuring acts from its social, cultural, and gendered context.

Science and scientific methods not only describe but also partly create the phenomenon under investigation, as emphasized by Bacchi (2009), as well as by Law (2004) and Haraway (1991). An example of this is the arguments for NSSI as a separate diagnosis, which fit well into the medical paradigm. The arguments are often based on results from self-report questionnaires on self-injuring acts and emotions. As the findings in the studies are subjected to interpretations, the theoretical approach of the researcher is decisive for the results and the outcome of the investigation, and it is the researcher who decides whether or not those self-reports indicate signs of psychopathology that could qualify for a diagnosis.

A medical vocabulary—consequences for understanding self-injuring acts

Searching for individual characteristics, using vocabulary as personality traits, emotional sensitivity, and psychopathology, to explain mental suffering is well established within the medical/psychiatric and psychological paradigm. The psychiatric habit of medicalizing women’s minds and bodies (Chesler, 1972; Graff and Mallin, 1967; Grunebaum and Klerman, 1967; Pao, 1969; Rosenthal et al., 1972; Russell, 1995; Ussher, 2011) has contributed to the readiness to classify self-injuring acts as an outcome of female biology and women’s fragile psychological structure, neglecting the context of such acts. The problem representations inherent in this vocabulary exclude the social context in which self-injuring acts occur and continues to define and articulate self-injuring acts as emanating from intrapersonal factors, although biological factors, emotional sensitivity, and psychopathology cannot sufficiently explain the high frequency of self-injuring acts in the community.

Encounters in medical settings—consequences for the individual

Horwitz and Wakefield’s (2007) description of how psychiatry has transformed normal sorrow into depressive disorders is applicable to the field of self-injuring acts. The ongoing proposals to make NSSI a separate diagnosis in the DSM runs the risk of directing the attention of psychiatric care workers toward searching for individual features that are in line with the proposed diagnostic criteria. To label someone’s emotional status, and to frame this status in a diagnostic concept, make other aspects of the individual invisible. Wilkinson (2005) and Parens (2013) argue that such a medical view in encounters objectifies the individuals and make them passive receivers of treatment decided on and delivered by psychiatric care workers. When a diagnostic vocabulary dominates the description of people’s suffering it is, according to Wilkinson (2005), impossible to attend properly to the traumatic ways in which mental suffering occurs in lived experiences.

Significant in the hitherto accomplished research on self-injuring acts is a focus on intrapersonal factors, which influences professional care workers in encounters with people who self-injure. Findings in research on medical professionals’ views on people who self-injure indicate that encounters often are characterized by confusion and misunderstandings. Sometimes, there is also a struggle on the interpretation of the causes and motives for the self-injuring acts. The ongoing focus on intrapersonal factors and the medical diagnostic culture thus hamper the possibilities to support those who self-injure.

A demedicalized perspective and vocabulary

Referring to a US national survey screening for mental illness which indicates that almost half of the US citizens should face diagnosable mental disorder during their life-time, Conrad (2007) asks where the line between a mild disorder and normal life difficulties should be drawn. The same question can be asked with respect to self-injuring acts—is it possible that, roughly speaking, a fifth or even more of the adolescent population suffer from intrapersonal sensitivity, a fragile personality structure, or a mental disorder? If not, what other elements than intrapersonal factors could generate the high figures of self-injuring acts? And how should such elements be conceptualized, articulated, and researched?

So far there have been very few attempts from other scientific fields to revisit and reconceptualize self-injuring acts, which possibly is a reflection of the power and status of the medical field. To push research on self-injuring acts forward, there is a need to investigate and explore the impact of social, cultural, structural, and gendered factors in modern society. As Smith (2011) puts it, there is a myriad of factors, on the micro- as well as on the macro level, influencing and surrounding the onset of mental pain and mental suffering. Such factors need to be addressed in research on self-injuring acts, and perspectives from social sciences and gender studies should complement the medical interpretation of self-injuring acts, in order to create a broader and more nuanced view on the phenomenon.

Conrad (2007) has described how processes of medicalization of social phenomena have expanded during the last decades. Noticing the frequency of self-injuring acts in the community, the view on self-injuring acts as a medical issue needs to be revisited to move the phenomenon out of the medical sphere. That is, research on self-injuring acts needs to move beyond the present medicalization, in a process of demedicalization, if we are to understand the rationale behind the widespread phenomenon of self-injuring acts.

To implement other perspectives on self-injuring acts in order to demedicalize the phenomenon, the vocabulary used in research on self-injuring acts must be supplemented with social science vocabulary and concepts that problematize gender. The above mentioned critical studies on psychiatric and psychological research on self-injuring acts offer some useful concepts and perspectives with a potential to alter the medical view, as they emphasize the importance of researching the context where self-injuring acts take place. For example, Chandler et al. (2011) puts forward the observation that the impact of social class and socioeconomic circumstances are rarely discussed in research on self-injuring acts. Hodgson (2004) offers a view on self-injuring acts as partly (on the micro level) a response to emotions evoked in interpersonal relations and partly (on the macro level) a strategy to handle the societal pressure to appear “normal.” Ekman and Söderberg (2009) discuss, on the macro level, the pressure modern society places on young people today, and on the micro level, how interaction with significant others may hamper or facilitate the well-being of the individual. They argue for a systemic approach, which encompasses a circular view instead of a linear view, as it is impossible to determine one specific factor as responsible for the onset of self-injuring acts. Kokaliari and Berzoff (2008) argue that an analysis of the concept of power is essential in gender research on self-injuring acts, to make visible social and cultural power structures which could have more impact on the onset of self-injuring acts than individual intrapersonal factors. Self-injuring acts can be interpreted as a voluntarily chosen strategy (Adler and Adler, 2007), which implies that concepts such as activity, intentionality, and meaning are important alternatives to the medically connoted concepts for describing self-injuring acts. Expanding and deepening the perspectives outlined above would push research forward by putting less weight on intrapersonal characteristics, psychopathology, and diagnostic criteria and put more weight on social, cultural, and gender factors and norms.