Apr 8, 2015 - Tony Salvatore .... determined to be mentally ill at the time of death (Robbins et al., 1959). ..... unshakeable commitment enjoyed by the model.
DISCUSSION DRAFT
8 April 2015
Do 90% of Suicide Victims Really have Serious Mental Illness? Psychological Autopsy Studies, Psychopathology, and Suicide
Tony Salvatore Montgomery County Emergency Service Norristown, PA Abstract For more than fifty years the psychological autopsy has been used in suicide research. The most common application is to determine the incidence of diagnosable psychiatric disorders in suicide victims. These studies have consistently found that about ninety percent of the decedents who were the subjects of such inquiries had signs of mental illness based on interviews of those who knew them in life. This finding is cited to such an extent that it almost seems to be axiomatic. As a result, mental illness appears to overshadow other suicide risk factors. Individuals with mental illness have a high risk of suicide but this may be because of exposure to many other risk factors, some of which are precipitated by mental illness or which interact with it. Psychological autopsy-based studies of suicide victims have received little critical attention. Here they are examined in terms of the reliability of informants, the methodology, feasibility of diagnosis by proxy, other data, and new theories of suicide. --The World Health Organization (WHO) (2014) has estimated that there were over eighthundred thousand suicides worldwide in 2012, one such death every forty seconds, with an estimated age-standardized rate of 11.4 per one-hundred thousand people. WHO reports that suicides are under-reported in many countries because of cultural sensitivity and that public health surveillance systems to record such deaths are nonexistent or limited in many Third World nations (particularly in Africa and parts of Asia), and that suicides may be misclassified even in nations with strong vital statistics registration. In the United States in 2013, there were 41,149 reported suicides, making it the tenth leading cause of death in the nation (American Foundation for Suicide Prevention, 2015). WHO (2014) cautions that despite a strong link between mental illness and suicide, no single cause has been found and many factors play a role in a suicide. This caveat notwithstanding, turning to the suicide research and prevention literature for insight on the topic of suicide and mental illness will quickly produce this statement in some form: Ninety percent of suicide victims had a psychiatric disorder (e.g., Goldsmith et al., 2002; 1
DISCUSSION DRAFT
8 April 2015
Litts et al., 2008). Mental illness has indeed been shown to be a factor in suicide, but “the relative importance of mental disorders compared to social strains is not fully clear” (Stack, 2014). Be this as it may, the prominence of the association between psychiatric disorders and suicide is pervasive in suicide prevention. It shapes the conversation about suicide. For these reasons and others, more attention should be given to how mental illness came to be seen as being so prevalent among suicide victims and basically being elevated to the status of an explanatory model of suicide. Where did the 90% figure come from? The conclusion that ninety percent of suicide victims suffered from some form of mental illness is the product of research based on “psychological autopsies” (AKA “follow-back studies”) of suicide victims involving inquiries among their survivors. This methodology was originally developed in the US as a procedure to resolve equivocal deaths. These arise when the cause of death is known but the mode is not immediately clear and it is uncertain if it should be classified as a suicide (Shneidman, 1981). A psychological autopsy (PA) attempts to develop a profile of a victim from accounts gathered through interviews with family members, friends, co-workers, and other contacts (Litman, 1989). Medical examiner’s and coroner’s reports and other documentation may also be included in psychological autopsies, but there is no standard protocol. There are two applications of the PA methodology. The first is as a forensic tool used to discern the nature of a specific decedent’s death. This involves a very detailed analysis of an extensive array of data generated by interviews and culled from official reports to develop “a rich psychological biography” (Knoll & Hazelwood, 2009). Such through-going analyses take considerable time and resources and tend to “flexible” in how they are conducted (Knoll & Hazelwood, 2009). Few research-oriented PAs focus on single cases. One exception is Autopsy of Suicidal Mind in which eight suicide researchers attempted to make inferences about the suicide of adult male based on his suicide note, other materials, and interviews with several family members and friends (Shneidman, 2004).
2
DISCUSSION DRAFT
8 April 2015
The second type of PA is as a tool in survey research with large numbers of decedents. These take the form of a “structured interview by a trained investigator” to “produce a psychological biography that creates a pathway to death” (Berman, 2013). The purpose of these “assessments in absentia” is to try to reconstruct a victim’s thoughts, feelings, and behaviors before the suicide (Clark & Horton-Deutsch, 1992). This methodology was first used with suicide cases in a 1959 study in which more than ninety percent of the victims were determined to be mentally ill at the time of death (Robbins et al., 1959). Similar subsequent research “established that more than 90% of completed suicides have suffered from usually comorbid mental disorders” (Isometsa, 2001). Consequently it is “now generally accepted that 90%...of suicides have one or more psychiatric disorders….” (Roy, 2001) The PA’s contribution to this breakthrough has led to it being referred to as the “gold standard” for suicide research (Berman, 2013). What about the other ten percent? PAs seem to have so strongly joined suicide to pre-existing mental illness that the absence of psychiatric morbidity in in even a small number of victims compels an explanation. One study took note that since so many PA studies found that ninety percent of suicide victims are afflicted with mental illness it was necessary to determine why mental illness had not been found in the other ten percent (Ernst et al., 2004). The authors concluded it was because researchers had simply “failed to detect” the signs of mental illness that must have been present. In other words, it apparently may be assumed that fully one-hundred percent of suicide victims would be found to manifest signs and symptoms of a psychiatric disorder if the researchers were more careful in gathering their data. Others saw the explanation for this apparent discrepancy lying in the how mental illness manifested itself in some of the decedents who may have merely been sub-syndromal. For example, it has been proposed that all suicide victims are indeed mentally ill but in some symptoms are simply not sufficiently strong to support a formal psychiatric diagnosis (Joiner, 2005). A similar explanation for the absence of clearly discernible psychiatric symptoms in suicide victims is that some victims may have “more subtle psychopathological alternations that are unidentifiable using the PA method” (Millner et al., 2013). 3
DISCUSSION DRAFT
8 April 2015
Issues with Informants A key issue with retrospective studies is the scope and depth of the informants’ recollections as well as the emotional effects of the inquiries. In PA studies the decedent’s next of kin are considered to be “knowledgeable proxies” (Milner et al., 2012). However, there may be limits to what such sources can contribute. Shaffer (1988) points out that psychological autopsies are inherently limited to what the informant had opportunity to directly observe. Lichter (1981) saw psychological autopsies as open to bias and hearsay information. The effect of traumatic loss and bereavement on informants who were close relatives or friends of a suicide victim as well as the amount of time that may have elapsed between the death and the interview may be factors that affect the accuracy of the responses (Jacobs & Klein, 1993). A study focusing specifically on the reactions of informants noted that “interviews are likely to be emotive and potentially stressful” and that almost one-quarter reported being upset during the questioning (Hawton et al., 2003). Concern has been expressed on how participation in a PA may affect informants who are “suicide survivors” (Beskow et al., 1990). Those closest to the suicide victim in life may also be severely traumatized by the loss and the circumstances in which it may have occurred (Cvinar, 2005). Behavioral science has not defined the timeframe for recovering from a suicide. Nonetheless, PA researchers do not hesitate to try to enlist the recently bereaved as informants. A European study (Gustofsson & Jacobsson, 2000) concluded that the optimal time to reach out to family members is two to three weeks after the suicide. Another European report a decade earlier recommended that interviews be done two to six months after the suicide (Beskow et al., 1990). A study of family members during the first month after a suicide found all to be acutely bereaved and extremely impaired emotionally (Mitchell et al., 2009). This should raise concern about the suitability of such individuals as informants in the first weeks or even months after their loss. A Korean study cited complaints of stress by family members who had been interviewed “within 3 months follow the suicide” (Sea et al., 2013). The severe bereavement of prospective informants is reflected in the contention that being a PA interviewee may be beneficial because it is an opportunity to talk about the loss, relieve any guilt, and receive referrals to support groups (Henry & Greenfield, 2009). Less attention has 4
DISCUSSION DRAFT
8 April 2015
been given in the literature to how interacting with suicide survivors affects the interviewers. An examination of the ethics of psychological autopsies noted that “interviewing a survivor is a delicate matter” and the “psychological strain” on the interviewer must be taken into consideration (Beskow et al., 1990). The very nature of suicide bereavement raises questions about the reliability of suicide survivors in identifying signs and symptoms of possible mental illness in someone they cared about. Three potentially complicating aspects of suicide loss identified in the literature are: (i) the struggle to understand why the suicide happened; (ii) feelings of responsibility for not preventing the suicide; and (iii) negative sentiment towards the victims (Jordan, 2001). Any of these could color the survivor-informant’s recall. Suicide survivors may seize upon mental illness as an explanation for a loss that may be otherwise unexplainable. Attributing mental illness to the victim may also lighten the burden of guilt that some may feel. Suicide survivors are reported to experience self-stigmatization because of their relationship to the victims and to even question their own mental health (Jordan, 2001; Dunn & Morrish-Vidners, 1987). If many suicide survivors agonize over missing the warning signs of suicide in their family member or friend it may be asked how they can help in making a psychiatric diagnosis, which relies on recalling similar, and perhaps even more obscure, behavioral signs on the part of their loved one in the weeks or months before their suicides. There is also the possibility of projection. Studies of suicide survivors have found that they were more likely than controls to have histories of psychiatric disorders and to have close relatives with mental illness (McIntosh, 1993). Methodological Concerns The prevailing view of the PA in suicide research seems to be that it has few methodological issues or only minor ones that are readily correctable (Cavanaugh et al., 2003). This research modality has been described as “highly rigorous, population-based studies…invaluable…in verifying causal hypotheses” that are “generalizable to the population” (Mosciciki, 1997). Assessments such as this gave impetus to this growing stream of research, enhanced the credibility of its results, and possibly discouraged even constructive criticism. 5
DISCUSSION DRAFT
8 April 2015
Pouliot and De Leo (2006) offered a critique that identified a number of shortcomings with PA suicide research, the most serious of which was an absence of standardization. Nonetheless these authors did not question this approach or the findings produced to date. A test of the validity of proxy reports concluded that they best captured actual suicidal behavior, but they were not as useful in regard to data such as emotional support (Connor et al., 2001). The victim’s ideation, threats, or past attempts were found to stand out more vividly than the underlying emotions that may have drove them. This may be the proper area of inquiry for this type of study. A comprehensive 2012 review asked: “Is it really possible to assign psychiatric diagnoses to someone who is dead by interviewing someone else?” (Hjelmeland et al., 2012) Concerns about interviewer and respondent bias, timing of the inquiry, and how mental illness was identified were raised. The authors concluded that it is impossible to diagnose the deceased by proxy and that psychological autopsy studies cannot “serve as an evidence base for the claim that most people who die by suicide are mentally ill.” This objection seems to have been overlooked as the number of mutually supportive PA reports continues to mount. Psychological autopsies may be vulnerable to “confirmatory bias.” If the purpose is to try to make a psychiatric diagnosis, researchers may give more attention to responses indicating signs of mental illness than those that do not (Rogers & Lester, 2010). Perhaps the high incidence of mental illness found in suicide victims by psychological autopsies is related to their inherent focus on a specific suicide risk factor? A review of the PA as a forensic tool to determine “proximate cause” in regard to malpractice suits and other legal cases suggests that such studies do not have to be limited to psychological .variables. The method has utility in identifying “the role of a variety of factors in bringing about a suicidal death” (Jacobs & KleinBenheim, 1996). If its use is not constricted it is “the best method available to study the detailed characteristics of suicide victims” (Jacobs & Klein-Benheim, 1996). The “notable finding” of PA suicide research yielding “similarity of results over four decades of investigation” has been pointed out (Maris et al., 2000), but this remarkable track record is not questioned because “it does not appear that the…method significantly 6
DISCUSSION DRAFT
8 April 2015
overestimates the presence of disorder.” The “notable finding” of PA studies is the bulwark of the mental illness model1 of suicide that is confirmed by the ongoing replication of this line of research. As a result of this record, and regardless of any lingering methodological concerns, this finding is now an “established ‘truth’ among suicidologists” (Hjelmeland et al., 2012). The mental illness model of suicide seems to be seen as incontrovertible and inexorable among a large number of researchers and practitioners. The high incidence of mental illness found in PA suicide studies seems to almost have the status of a law in suicide prevention. Issues with Psychiatric Evaluation and Diagnosis PAs purport to be able to establish a psychiatric diagnosis ex post facto. There are many studies that have validated this capability (e.g., Brent et al., 2007). These studies generally have not taken the nature of a psychiatric evaluation and the diagnostic process into account. Psychiatric diagnoses are customarily performed in the course of communication between two living beings (even when telepsychiatry is used). In practice, psychiatric diagnoses are the product of a direct individual clinical interview by a psychiatrist or other clinician with the necessary skills. The clinical interview is how the personal problems and symptoms used to formulate a psychiatric diagnosis are identified. It begins with a “doctor-patient relationship” that is the context for (i) eliciting personal information about emotions, feelings, attitudes, and relationships, (ii) sorting, testing, and prioritizing of this information, and (iii) arriving at a diagnosis (McCready, 1986). A psychiatric diagnosis is not, however, the result of simply taking a clinical history; it involves determining how the individual is affected by her or his history. This does not seem derivable from someone else’s memories however strong or by means of an interviewer’s skills or tools. A face-to-face interview assures inclusion of the patient’s perspective and direct observation of interpersonal behavior. Neither of these can be supplied by third parties.
1
This term is basically equivalent to, but more explicit than, the more commonly used “medical model.” 7
DISCUSSION DRAFT
8 April 2015
Psychiatric interviews are intended to be comprehensive, multidimensional, and empathetic. American Psychiatric Association (2013) guidelines outline the scope of a psychiatric evaluation. One facet of the evaluation is the “Mental Status Examination” that looks at personal appearance, speech, affect, behavior, stream of thought, and other items that require first-hand observation. A psychiatric diagnosis is the outcome of a highly structured assessment process that even when done “by the book” is still a clinical judgment call. While “real time” psychiatric evaluations may incorporate collateral sources they are never solely based on such sources as are psychological autopsies. A further concern is the number of steps that information gathering and filtering may take in a PA before it reaches the psychiatric diagnostician. In most such studies the process starts with trained interviewers who meet with the informants. The interview data is then passed to a psychiatrist who makes a diagnosis. In some studies the diagnoses are determined in case conferences involving a board-certified psychiatrist and a psychologist (Kelly, 1997). This approach is dictated by research considerations, but it distances the diagnostician from the primary source of the information on which the diagnosis is made and precludes and opportunity to evaluate the source as would happen in a conventional psychiatric interview of collateral sources. The term “autopsy” literally means “to see for one’s self.” However, in a PA the experiences, behaviors, feelings, beliefs, and motivation of a suicide victim are seen by the researcher and diagnostician through the recall of others.
Some Countervailing Data The Centers for Disease Control and Prevention’s (2014a) National Violent Death Reporting System (NVDRS), has collected data on suicide from multiple sources from sixteen states in the US. There are four principal sources of data: death certificates, medical examiner and coroner reports, law enforcement reports, and vital statistics records. There are no interviews of family or friends. NVDRS data may include information on mental health problems and treatment, recent employment problems, finances, interpersonal relationships; physical health problems; and the means and circumstances of death.
8
DISCUSSION DRAFT
8 April 2015
In 2008, of 9276 suicides on which data was compiled, approximately 45.4% had a diagnosed mental illness at death. The NVDRS has found that “mental health problems were the most common circumstances among suicide decedents” in the US. However, the incidence in a very large population of victims is half that routinely reported by PA studies. The NVDRS culls data from a larger and far more consistent source than any PA study. NVDRS data has been available for several years and its usefulness in suicide prevention planning has been highlighted, but PA studies far overshadow it as a primary source for exploring the suicidemental illness link. This may change after NVDRS data includes 32 states (Centers for Disease Control, 2014b). The National Survey on Drug Use and Health (NSDUH) is a large scale population surveys that asks about suicidal behavior. In 2004-05, the NSDUH asked a national sample of adults age 18 and older about depressive episodes and suicidal behavior (Substance Abuse and Mental Health Services Administration, 2006). During their worst or most recent depressive episode just over one-half thought it would be better if they were dead, two-fifths had suicidal ideation, and one-fourth made a suicide plan or an attempt. The latter group must be considered at very high risk, but does not come close to the proportion of suicide completers found to have a psychiatric disorder in psychological autopsies. Depression does not represent the totality of psychiatric disorders, but in some form it is the diagnosis most given in psychological autopsies (Hjelmeland, 2012). It has been suggested that suicide attempters and completers are two different populations as males typically account for more of the latter while females comprise most of the former (Maris et al., 2000). Perhaps one further difference is that mental illness has a more deleterious effect on those individuals going on to make fatal attempts. A study of mental illness and subsequent suicidal ideation, suicide planning, and suicide attempts using data from the National Comorbidity Survey Replication (NCS-R), a survey of US households (Nock et al., 2010), found that approximately eighty percent of US suicide attempters have some form of mental illness, which the authors note is generally consistent with PA findings. However, the results also suggest that different disorders may drive different suicidal behaviors. Depression may bring on suicidal ideation while conditions linked to poor impulse control such as bipolar disorder may have more to do with suicide attempts. Suicidal 9
DISCUSSION DRAFT
8 April 2015
behavior clearly increases suicide risk but in most cases it does not lead to death. As a contributor to non-fatal suicidal behavior mental illness warrants serious attention but not the pre-eminent role accorded it by PA studies. What do the new suicide theories have to say? The “Interpersonal Psychological Theory of Suicide” proposes that a potentially fatal suicide attempt requires: (1) a strong belief that one is a liability and not fulfilling expectations, (2) a deep sense of loneliness and isolation; and (3) a sense of fearlessness about lethal selfharm (Joiner, 2005). These beliefs may commonly occur in those with serious mental illness, but the theory does not assign primacy to mental illness as the reason why people die by suicide. According to this theory, the capability for making a suicide attempt emerges, via inurement to repeated physically painful or fear-inducing behaviors or experiences, which reduce the innate drive for self-preservation. Joiner explicitly observes that “mental illness alone does not provide a satisfying explanation for suicide because mental illness is much more common than suicide.” Life experiences common to those with serious mental illness, such as loss of self-esteem, weakened or failed social supports, self-injury, exposure to violence, disability, among others, contribute to a sense of burdensomeness, a lack of social connectiveness, and an acquired capability for lethal self-harm. These are the elements that place an individual at grave risk of suicide. The “Integrated Motivational-Volitional Model” sees a suicide attempt as the outcome of a three-phased process — premotivational, motivational, and volitional (O’Connor, 2011). An individual may progress from low to high danger of suicide through the interplay of fixed risk factors, triggers, ideation, intent, formation of a plan, and attempting suicide. An individual may progress from low suicide risk to high risk through the interplay of fixed background factors (e.g., gender or mental illness), triggers (e.g., a recent loss), ideation, intent, formation of a plan, and completion of the plan to commit suicide. At the premotivational phase, an individual has serious risk factors for suicide but is not suicidal. These may include financial problems, legal issues, divorce or other interpersonal conflicts, substance abuse, and mental illness. Strong self-criticism, negative self-appraisal, and a sense of failure in meeting one’s 10
DISCUSSION DRAFT
8 April 2015
expectations often emerge at this stage. All are preconditions for the possible onset of suicidality. The motivational phase begins with suicidal ideation and ends with a specific plan. Intent to die appears in this phase, and the person communicates this by voice, text, or some other method. The individual is suicidal, at high risk, and possibly moving toward a suicide attempt, but has not set a plan in motion. In the volitional phase, the individual’s suicide plan is underway and death may be imminent. There is a resolute commitment to terminating one’s life per a specific plan. The person has established means and the capability for lethal self-harm is operational. The individual is at or nearing the point of no return. In this theory, mental illness may be a risk factor at the premotivational phase, but it may be only one of many present and any or all may be preconditions for suicidality. These two theories acknowledge that mental illness plays a role in the onset of suicidality, but do not position it as a principal driver. Suicide is seen as the outcome of a complex, multifactorial process involving the interplay of many variables rather than just a byproduct of mental illness. These theories would seem to readily lend themselves to research using the PA method to search out evidence of burdensomeness, social disconnectiveness, and acquired capability for lethal self-harm in a suicide victim’s life or the passage from the premotivational to the motivational and volitional stages. How are mental illness and suicide related? Mental illness is a strong risk factor for suicide and the most recognized (Oquendo & Baca-Garcia, 2014). This was confirmed by a 1997 meta-analysis of well over two hundred studies of suicides of patients with a known diagnosis that found that almost all psychiatric disorders conferred an increased risk of suicide (Harris & Barraclough, 1997). These findings were confirmed by a more recent meta-review (Chesney et al., 2014). A review of studies published between 1959-2001 of suicide victims with and without a psychiatric hospitalization “showed that 98% of those who committed suicide had a diagnosable mental disorder” (Bertolete et al., 2002). A related study found that mood disorder diagnoses predominated and were attributed to just less than one-third of the suicide victims (Bertolete et al., 2004). These findings do not mean that almost every suicide victim had a diagnosable mental illness. 11
DISCUSSION DRAFT
8 April 2015
Many factors can add to suicide risk, and those for one person may not be so for another. Each may have an array of risk factors, and they may act and interact differently from person to person. This is the case with psychiatric disorders, which may add to suicide risk, but have “little predictive power and…do not account for why people try to kill themselves” (O’Connor & Nock, 2014). Familial and twin studies of suicidal behavior indicate that it occurs among close relatives independent of mental illness (Qin et al., 2002). Some family studies have found that the presence of suicidal ideation among relatives is correlated with mental illness but this was not found to be related to completed suicides (Brent & Mann, 2005). The proper perspective to take seems to be that “a psychiatric disorder is generally a necessary but insufficient condition for suicide” (Mann et al., 1999). The overstating of the role of mental illness in suicides has skewed the perception of its place as a stressor. This can be corrected by turning to the Stress-Diathesis Model that attempted to specifically explain suicidal behavior in individuals with psychiatric disorder and situates mental illness as the source of stress which interacts individual characteristics (Mann et al., 1999). Is there a suicide research paradigm? The paradigm concept is the idea that in any field there is a set of almost universally shared assumptions. The paradigm defines the field and shapes the rules of the game. The concept was introduced by Thomas Kuhn in The Structure of Scientific Revolutions (1962). A paradigm determines what gets taught and studied, the methods, and how the findings are interpreted. It defines the needs addressed. A paradigm supplies "all the answers" to researchers' and practitioners' questions. A paradigm becomes self-sustaining. It provides continuity and stability. Its tenets are defended and change is resisted. It becomes more complex and encompassing, but does not evolve. In suicide research, PA studies and their findings seem to have attained “paradigm” status. As a prevailing paradigm the mental illness model of suicide driven by PA studies binds researchers to a compelling body of empirical findings which may in turn deter taking a wider look at other suicide risk factors. Adherence to the mental illness model, despite criticism and 12
DISCUSSION DRAFT
8 April 2015
countervailing findings and theories, has been noted, but these may not sufficiently convey the unshakeable commitment enjoyed by the model. When two PA studies in China found levels of mental illness significantly below the customary ninety percent level, one reaction was that “the conventional wisdom that suicide is almost always the outcome of mental illness will not be altered by one or two studies” (Phillips, 2010). The answer is again to refine the methodology rather than to consider reconceptualizing the problem. Similar results from a Korean study led to the conclusion that this has more to do with Asian culture and the inclination to not discuss the suicides of relatives than the methodology (Sea et al., 2013). According to Kuhn, paradigms only change through radical and sudden shifts. These occur when new discoveries, knowledge, or concepts arise which cannot be rejected or assimilated by the old paradigm. Hopefully, these developments will mount and eventually break the hold that the mental illness model has on suicide research and suicide prevention. There may be hopeful but subtle signs the fixation on mental illness in suicide prevention may be loosening. For example, in its recent well-received publication, Preventing Suicide: A Global Imperative the World Health Organization (2014) includes “only people with mental disorders are suicidal” among the “myths of suicide.” WHO explicitly notes that “not all people who take their own lives have a mental disorder.” However, even WHO cannot seem to make a total break. It seems to go on to perpetuate this very myth (with some socio-economic limitation) by stating: “In high-income countries, mental disorders are present in up to 90% of people who die by suicide, and among the 10% without clear diagnoses, psychiatric symptoms resemble those of people who die by suicide.”
Concluding Comments Despite the issues raised here, psychological autopsies remain a practical and viable research strategy for use in gathering information about the background, actions, and experiences of suicide victims. However, their use should be untethered from the mental illness model of suicide and not be restricted to trying to define the mental states of the deceased who took their lives. Post-mortem inquiries into the backgrounds of suicide victims 13
DISCUSSION DRAFT
8 April 2015
might contribute more to our understanding if they took the form of “psychosocial autopsies” that attempted to also reconstruct the victim’s social behavior and interactions. PA studies may address a wider range of factors, but mental illness seems to over-ride everything else. It has been suggested that such post-mortem inquiries give more attention to social factors in the victim’s life and be conducted as “sociological autopsies” (Scourfield et al., 2012). Widening the purview of such inquiries would better fit with the multifactorial nature of suicide and capture a wider range of possible risk factors. The social and interpersonal dimensions of a suicide victim’s life may be more readily and clearly recalled by informants than intra-psychic features that are not as easily observed. Adding or expanding the psychosocial aspect of after-suicide studies of victims would better relate this research to the new theories of suicide, such as the Interpersonal Psychological Theory, which see the lack of social connectiveness and similar variables as raising the risk of suicide. The purpose of this discussion is not to diminish use of the PA in suicide research or to challenge the significance of mental illness as a risk factor for suicide and a contributory factor in many suicides. The intent is to take a critical look at a sector of suicide research that, given its influence in the field, should get much more critical attention. Kuhn (1962) remarked that paradigms emerge as a scientific field matures. Suicidology emerged comparatively recently as a sub-discipline and is still in the process of establishing its “empirical foundations” (Maris et al., 2000). The scientific study of suicide must be based on systematic observation. Suicide presents some obstacles to meeting this requirement given its nature and low rate of incidence. The PA is one way to overcome these hurdles through retrospective observation and a nonexperimental research design (Maris et al., 2000). This may at least in part explain the largely unquestioning faith in the PA and its usual findings, which may be seen as helping to shore up the field’s “empirical foundations.” Be that as it may, it can at least be said that suicidology may have prematurely locked on a paradigm that may be inordinately influencing both suicide research and suicide prevention. Every field of scientific inquiry has at least two major constituencies. There are the researchers and the primary direct consumers of research who understand and acknowledge the limitations of research and generally tend to be circumspect in making generalizations. 14
DISCUSSION DRAFT
8 April 2015
Then there are those who draw their understanding of research from secondary and tertiary sources in which the provisos of the primary sources have not carried over and one or more generalizations may have crept in. Perhaps the majority of individual and organizational suicide prevention advocates fall into this category and this accounts in large part for the popularity of the mental illness model. Some part of the responsibility for adherence to this model also lies with the researchers, academics, and policymakers who seem to be as fascinated by the mental illness model as the grassroots level proponents of suicide prevention. One group repeatedly cites studies supporting “Ninety percent of suicide victims had a psychiatric disorder.” The other group repeatedly cites the phrase without the citations. The effect is the same. The paradigm endures.
15
DISCUSSION DRAFT
8 April 2015
References American Foundation for Suicide Prevention, Facts and Figures at http://www.afsp.org/understanding-suicide/facts-and-figures retrieved March 6, 2015. American Psychiatric Association (2013) Practice Guidelines: Psychiatric Evaluation of Adults, Second Edition, Washington. DC: American Psychiatric Association. Berman, L. (2013) Clinical lessons: The psychological autopsy, Washington, DC: The American Association of Suicidology. Bertolete, J. & Fleischmann, A. (2002) Suicide and psychiatric diagnosis: A worldwide perspective, World Psychiatry, 1(3), 181-185. Bertolete, J., Fleischmann, A., DeLeo, D. & Wasserman, D. (2004) Psychiatric diagnoses and suicide: Revisiting the evidence. Crisis, 25(4), 147-155. Beskow, J., Runeson, B. & Asgard, U. (1990) Psychological autopsies: Methods and Ethics, Suicide and Life-Threatening Behavior, 20(4) 307-323. Brent, D., Perper, J., Moritz, G., Allman, C., Roth, Schweers, J., & Balach, L. (2007) The validity of diagnoses obtained through the psychological autopsy in adolescent suicide: Use of family history, Acta Psychiatrica Scandinavica, 87(2), 118-122. Brent, D. & Mann, J. (2005) Family genetic studies, suicide, and suicidal behavior, American Journal of Medical Genetics Part C: Seminars in Medical Genetics, 133, 13-24. Cavanaugh, J., Carson, A., Sharpe, M. & Lawrie, S. (2003) Psychological autopsy studies of suicide: A systematic review, Psychological Medicine, 33:1, 395-405. Centers for Disease Control and Prevention (2014a) Injury Center: Violence Prevention. http://origin.glb.cdc.gov/violenceprevention/pub/mmwr.html, retrieved March 6, 2015. Centers for Disease Control and Prevention (2014b) CDC awards $7.5 million to expand the National Violent Death Reporting System to 32 states, CDC Media Relations, September 8, 2014, http://www.cdc.gov/media/releases/2014/p0908-NVDRS.html, retrieved March 6, 2015. Chesney, E., Goodwin, G. & Fazel, S. (2014) Risks of all-cause suicide mortality in mental disorders: A meta-review, World Psychiatry, 13, 153-160. Clark, D. & Horton-Deutsch, S. (1992) Assessment in absentia: The value of the psychological autopsy method in studying the antecedents of suicide and predicting future suicides in Maris. 16
DISCUSSION DRAFT
8 April 2015
R., Berman, Maltsberger, J. & Yufit, R. (Eds.), Assessment and prediction of suicide, New York, NY: The Guilford Press. Conner, R., Conwell, Y. & Duberstein, P. (2001) The validity of proxy-based data in suicide research: A study of patients 50 years of age and older who attempted suicide, Acta Psychiatrica Scandinavica, 104:6, 454-457. Cvinar, J. (2005) Do suicide survivors suffer social stigma? A review of the literature. Perspectives in Psychiatric Care, 41(1), 14-21. Dunn, R., & Morrish-Vidners, D. (1987) The psychological and social experience of suicide survivors. Journal of Death and Dying, 18(3) 175-215. Ernst, C., Lalovic, A., Lesage, A., Seguin, M., Tousignant, M. & Turecki, G. (2004) Suicide and no axis I psychopathology, BMC Psychiatry, 4:1, 7. Goldsmith, S., Pellmar, T., Kleinman, A., & Bunney, W. (Eds.) (2002) Reducing suicide: A national imperative, Washington, DC: The National Academies Press. Harris, E. & Barraclough, B. (1997) Suicide as an outcome for mental disorders. A meta-analysis, British Journal of Psychiatry, 170, 205-28. Hawton, K., Houston, K., Malmberg, A. & Simkin, S. (2003) Psychological autopsy interviews in suicide research: The reactions of informants, Archives of Suicide Research, 7, 73-82. Henry, M. & Greenfield, B. (2009) Therapeutic effects of psychological autopsies: The impact of investigating suicides on interviewees, Crisis, 30(1), 20-24. Hjelmeland, H., Dieserud, G., Dyregrov, K., Knizek, B. & Leenaars, A. (2012) Psychological autopsy studies as diagnostic tools: Are they methodologically flawed? Death Studies, 36, 605626. Isometsa, E., (2001) Psychological autopsy studies – A review, European Psychiatry, 16(7), 379385. Jacobs, D. & Klein, M. (1993) The expanding role of psychological autopsies in Leenaars, A. (Ed.) Essays in honor of Edwin Shneidman, Northvale, NJ: Jason Aronson. Jacobs, D. & Klein-Benheim, M. (1996) The psychological autopsy: A useful tool for determining proximate causation in suicide cases, Bulletin of the American Academy of Psychiatry and Law, 23(2) 165-182. Joiner, T. (2005) Why people die by suicide, Cambridge, MA: Harvard University Press. 17
DISCUSSION DRAFT
8 April 2015
Jordan, J. (2001) Is suicide bereavement different? A reassessment of the literature. Suicide and Life-Threatening Behavior, 31(1), 91-102. Kelly, T. (1997) A Comparison of Suicide Completers with Psychiatrically Disordered and NonDisordered Subjects Via the Psychological Autopsy Method, 9th National Symposium on Doctoral Research in Social Work, Columbus, OH: Ohio State University College of Social Work. Knoll, J. & Hazelwood, R. (2009) Psychological Autopsy, in Allan Jamieson & Andre Moenssens (Eds.), Wiley Encyclopedia of Forensic Science, Hoboken, NJ: John Wiley & Sons. Kuhn, T. (1962) The Structure of Scientific Revolutions, Chicago, IL: University of Chicago Press. Litts, D., Radke, A. & Silverman, M. (Eds.) (2008) Suicide prevention efforts for individuals with serious mental illness: Roles for the state mental health authority. Washington, D.C. National Association of State Mental Health Program Directors/Suicide Prevention Resource Center. Litman, R. (1989) 500 Psychological autopsies, Journal of Forensic Science, 34, 638-646. Mann, J., Waternaux, C., Haas, G. & Malone, K. (1999) Toward a clinical model of suicidal behavior in psychiatric patients, The American Journal of Psychiatry, 156, 181-189. Maris, R., Berman, A. & Silverman, M. (2000) Comprehensive Textbook of Suicidology, New York, NY: Guilford Press. McIntosh, J. (1993) Control group studies of suicide survivors: A review and critique, Suicide and Life-Threatening Behavior, 23(2), 146-161. Milner, A., Jerneja, S. & DeLeo, D. (2013) Suicide in the absence of mental disorder? A review of psychological autopsy studies across countries. International Journal of Social Psychiatry, 59(6), 545-554. Mitchell A., Sakraida T., Kim Y., Bullian L., & Chiappetta L. (2009) Depression, Anxiety and Quality of Life in Suicide Survivors: A Comparison of Close and Distant Relationships, Archives of Psychiatric Nursing, 23, 2-10. Mosciciki, E. (1997) Identification of suicide risk factors using epidemiologic studies, Psychiatric Clinics of North America, 20(3), 499-517. Nock, M., Hwang, I., Sampson, N. & Kessler, R. (2010) Mental disorders, comorbidity and suicidal behavior: Results from the National Comorbidity Survey Replication, Molecular Psychiatry, 15, 868-876.
18
DISCUSSION DRAFT
8 April 2015
O’Connor, R. (2011) The integrated motivational-volitional model of suicidal behavior, Crisis, 32:6, 295-298. O’Connor, R. & Nock, M. (2014) The psychology of suicidal behavior, The Lancet Psychiatry, 1:1, 73-85. Oquendo, M. & Baca-Garcia, E. (2014) Suicidal behavior disorder as a diagnostic entity in the DSM-5 classification system: Advantage outweigh limitations, World Psychiatry, 13:2, 128-130. Phillips, M. (2010) Rethinking the role of mental illness in suicide, American Journal of Psychiatry, 167(7) 731-733. Pouliot, L. & De Leo, D. (2006) Critical issues in psychological autopsy studies, Suicide and LifeThreatening Behavior, 36, 491–510. Qin, P., Agerbo, E. & Mortenson, P. (2002) Suicide risk in relation to family history of completed suicide and psychiatric disorders: A nested case-control study based on longitudinal registers, Lancet, 360, 1126-1130. Robbins, E., Gassner, S., Kayes, J., Wilkinson, R. & Murphy, G. (1959) Some clinical considerations in the prevention of suicide based on a study of 134 successful suicides, American Journal of Public Health, 49, 888-889. Rogers, R. & Lester, D. (2010) Understanding suicide: Why we don’t and how we might, Cambridge, MA: Hogrefe Publishing. Roy, A., (2001) Consumers of mental health services. Suicide and Life-Threatening Behavior, 31 (Supplement), 60-83. Sea, J., Lee, C., Kim, K. & Kim, S. (2013) Characteristics of Korean Suicide: A Case-control Psychological Autopsy study. Korean Social Sciences Review, 3(2) 129-155. Schneidman, E. (1981) The psychological autopsy, Suicide and Life-Threatening Behavior, 11, 325-340. Scourfield, J., Fincham, B., Langer, S. & Shiner, M. (2012) Sociological Autopsy: An Integrated Approach to the Study of Suicide in Men, Social Science and Medicine, 74 (4) 466-473. Shaffer, D. (1988) The epidemiology of teen suicide: An examination of risk factors, Journal of Clinical Psychiatry, 49, Supplement, 36-41. Stack, S. (2014) Mental illness and suicide, The Wiley Blackwell Encyclopedia of Health, Illness, Behavior, and Society, Hoboken, NJ: John Wiley & Sons, Ltd.
19
DISCUSSION DRAFT
8 April 2015
Substance Abuse and Mental Health Services Administration (2006), Suicidal thoughts, suicide attempts, major depressive episodes, and substance abuse among adults, The OAS Report, Issue 34, Rockville, MD: Substance Abuse and Mental Health Services Administration. Wong, P., Chan, W., Beh, P., Yau, F., Yip, P. & Hawton, K. (2010) Research participation experiences of informants of suicide and control cases: Taken from a case-control psychological autopsy study of people who died by suicide, Crisis, 31(5), 238-246. World Health Organization (2014) Prevent Suicide: A Global Imperative, Geneva, Switzerland: WHO. Zhang, J., Conwell, Y., Zhou, L. & Jiang, C. (2004) Culture, risk factors and suicide in rural China: A psychological autopsy case control study, Acta Psychiatrica Scandinavica, 110, 430-437.
20
