Psychological autopsy: Overview of Indian evidence, best practice elements, and a semi-structured interview guide.

BACKGROUND: No review has been attempted, so far, on Indian psychological autopsy (PA) literature. There is also a dearth of interview guides which is at the heart of a PA procedure.
MATERIALS AND METHODS: Electronic searches of MEDLINE through PubMed, PsycINFO, and Google scholar databases were carried out from inception till February 2020 to identify relevant English language peer-reviewed articles from India, as well as global literature that provided information on best practice elements in PA. Abstracts generated were systematically screened for eligibility. Relevant data were extracted using a predesigned structured proforma, and a semi-structured interview guide was developed.
RESULTS: A total of 18 original articles, one case report, and three reviews/expert opinion articles which tried to give a description of PA procedure were found from India. Most Indian studies are of suicide PA (SPA), done to assess risk factors associated with suicide. There was a wide variation in reported rates of psychiatric morbidity among suicide decedents, while the other major risk factor for suicide in the Indian setting was stressful life events. An optimal approach to PA involves systematically collecting information from key informants and other sources using a narrative interviewing method, supplemented with psychological measures, and is probably best carried out within 1-6 months after the death.
CONCLUSION: There have been limited attempts to standardize PA. Most Indian studies use SPA. We propose a semi-structured PA interview guide, suitable for both research and investigational purposes. Copyright:

INTRODUCTION

In its essence, psychological autopsy (PA) refers to a postmortem investigative procedure that seeks to uncover the intention of the decedent through a thorough retrospective evaluation through structured interviews of informants as well as a perusal of relevant records.[1] Originally devised for investigating, clarifying, and assisting police inquiries on the mode of death in equivocal deaths,[2] psychological autopsies have, in recent times, been more commonly used as a research tool for investigating risk factors for completed suicides. To a large extent, this has been driven by the accepted “truism” in suicide research that roughly 90% of suicide decedents have one or more diagnosable mental disorders.[3]

Although several PA studies have been published from India, thus far, there has been no attempt to systematically review the available literature. Further, currently accepted PA interview practices suffer from several limitations. Lack of standardized instruments or methods, informant bias, lack of interviewer training, recall bias (due to time elapsed between the event and the interview), and issues with selection of controls are some of the key methodological drawbacks[4] which have also, predictably, led to questions about the admissibility of PA evidence in courts.[5]

Against this background, the present review was carried out with three objectives; to review the Indian PA literature with a dual focus on methodological aspects as well as to summarize their key findings; to synthesize best practice elements of PA interview procedure and finally; and to outline a semi-structured PA interview guide to assist practice and research in this area.

MATERIALS AND METHODS

Search strategy and study selection

We performed an electronic search of databases such as MEDLINE through PubMed, PsycINFO, and Google Scholar databases from inception till February 2020 to identify relevant English language peer-reviewed articles on PA. For PubMed, we used the following combinations of MeSH and free text terms; ((((autopsies)) OR (autopsies[MeSH Terms])) AND (((((“autopsy”[MeSH Terms] OR “autopsy/instrumentation”[MeSH Terms] OR “autopsy/methods”[MeSH Terms] OR “autopsy/psychology”[MeSH Terms]) OR (autopsy)) OR (autopsy/instrumentation)) OR (autopsy/verbal)) OR (autopsy/methods)) OR (autopsy/psychology))) AND ((((((“interview”[Publication Type] OR “interviews as topic”[MeSH Terms]) OR “interview”[All Fields]) OR ((((“interview, psychological”[MeSH Terms] OR (“interview”[All Fields] AND “psychological”[All Fields])) OR “psychological interview”[All Fields]) OR (“interviews”[All Fields] AND “psychologic”[All Fields])) OR “interviews psychologic”[All Fields])) OR ((((“interview, psychological”[MeSH Terms] OR (“interview”[All Fields] AND “psychological”[All Fields])) OR “psychological interview”[All Fields]) OR (“interviews”[All Fields] AND “psychological”[All Fields])) AND (“suicide”[MeSH Terms] OR “suicide”[All Fields]) AND forensic[All Fields] AND (“autopsy”[MeSH Terms] OR “autopsy”[All Fields])).

For other databases, the search terms were adapted as appropriate. Two independent psychiatrists performed the literature search. A manual search of the reference lists of generated articles was also done to locate relevant articles.

A total of 2387 articles were generated initially. We included all articles involving PA-based research from India. We also included PA studies in literature that provided relevant information to synthesize best practices for the following key steps of PA procedure; how and whom to select as informants, contact/approach methods, timing, setting, and termination of the interview. Gray literature (such as conference proceedings) was not searched.

Based on these inclusion criteria and after removing duplicates, 184 articles were shortlisted for inclusion based on their title and abstracts. After further filtering, 156 articles were shortlisted for full-text examination, and finally, 39 articles were included for the present review. Two independent authors participated in the study selection and all authors reached a consensus on the studies to be included. Being a narrative review, we neither attempted computation of effect sizes nor performed a risk of bias assessment for included papers.

Data extraction

Selected studies were categorized under two headings; first, Indian studies using PA method and second, evidence for best practice elements in PA interview. Accordingly, the results of the current review are structured under these headings.

We extracted relevant data from the Indian studies and tabulated them under the following headings; author and year, sample size and participant characteristics, methodology used, main results, and any other special features. Data extraction was done independently by two authors.

RESULTS

Suicide psychological autopsy: Indian scenario

Original/case studies using psychological autopsy as method (n = 19)

Verbal Autopsy only to ascertain suicide rate (n = 8)

A series of papers from Vellore[678910] used verbal autopsy to ascertain age-specific suicide rates in the same locality during different time periods. All these studies followed a similar methodology. Community health workers, health aides, and nurses independently visited the home, relatives, and neighbors of the deceased, traditional healers, and village leaders. Circumstances of the death were discussed with doctor and sometimes independently verified to check the accuracy of reporting.

Interestingly, they all reported a higher average annual suicide rate compared to the national average and conclude that suicide is largely underreported in the existing system. One of these studies[8] also found that stressful life events were reported by nearly all the decedents.

Two studies from Kerala[11] and Tamil Nadu[12] used verbal autopsy method to determine suicide rate and contribution of suicide to overall mortality. Trained health workers interviewed surviving spouses, close associates, or neighbors and wrote their verbal autopsy report. These were independently reviewed by one[11] or two[12] physicians to verify cause of death. Results from these studies showed that suicide underreporting varied between states.

A nationally representative survey[13] using nonmedical trained field workers and employing an augmented version of a verbal autopsy, called routine, reliable, representative, re-sampled household investigation of mortality with medical evaluation method (RHIME) found that suicide rates in India were among the highest in the world and suggested interventions such as restrictions on access to pesticides to control suicide.

Psychological autopsy with restricted information (n = 2)

Two studies[1415] used cross-sectional design to assess risk factors for completed suicide. Only relatives of the deceased were approached. Police records or additional information was not sought. A self-designed questionnaire with a special focus on recent stressful life events was used in one study,[14] while in the other,[15] relatives were approached 3-month postsuicide completion to enable the respondents to be out of the grief period and minimize recall bias while providing information.

Psychological autopsy involving extended information (n = 1)

When further evidence was collected through various sources such as suicide notes and circumstantial evidence from the police and the magistrate, the process of PA yielded valuable information such as gender-specific risk factors for completed suicide by hanging in Cuttack, Orissa.[16]

Psychological autopsy involving “key informant” relatives (n= 4)

In the earliest such PA study from India, authors used a detailed 178-item questionnaire with structured assessment of life events and psychiatric morbidity to interview key informants within 1–3 months after the death.[17] All interviews were done by a psychiatrist. A similar methodology, with structured instruments for psychiatric morbidity and life events and psychiatrist-led interviews, was followed in Kerala.[18]

Elsewhere, trained medical social workers and a counselor, who approached “key informant relatives” (defined as those who lived with the deceased for the past 2 years) used a self-designed semi-structured pro forma to conduct interviews.[19]

A study from Goa,[20] which involved “key informant relatives,” conducted the PA in the presence of 2–4 family members, against norm of one person at a time. Interviews were conducted using a semi-structured questionnaire in local language, evaluated stressful life events, and made diagnoses using the International Classification of Diseases-10thEdition.[21] A qualified psychiatrist performed the interviews after a gap of 1–3 months.

Psychological autopsy in special populations

Psychological autopsy of farmer suicides (n = 3)

Three PA studies have been carried out on this population. Two of them attempted to quantify risk factors for completed suicide among farmers in Wardha, Maharashtra. Structured pro formas were used in both. Results of one study were compiled as a monograph with little details on the methodology or results of the field study.[22] The other study[23] employed a case–control design and included the assessment of stressful life events for preceding 3 years, psychological antecedents, economic factors, CAGE questionnaire to screen for alcohol use disorder, and used Diagnostic and Statistical Manual of Mental Disorders (DSM)-IV[24] for making psychiatric diagnoses. Interviews were carried out after a minimum gap of 3-month postsuicide.

A study on farmers from Vellore[25] assessed presumptive life events, generated diagnoses using Structured Clinical Interview for DSM-III-R,[26] and collected collateral information from traditional healers, village leaders, and health records. Relatives were approached after a minimum gap of 2-month postsuicide completion.

Psychological autopsy in emergency (n = 1)

A recent case report[27] of PA performed in the emergency room following death due to self-poisoning was characterized by a thorough clinical history including the context of attempt, past psychiatric history, substance abuse history, premorbid personality assessment, assimilating various caregivers' accounts of the deceased, and using DSM-5 for psychiatric diagnoses.

In the above section, we focused on the methodology adopted by Indian PA studies. Table 1 summarizes their main findings.

Table 1

Summary of Indian studies using psychological autopsy method

Author, year	Sample size and characteristics	Method used	Main findings	Special remarks
PA studies to determine suicide rates (n=8)
Joseph et al., 2003[6]	108,873	Verbal autopsy	Mean suicide rate for 6-year period was 95.2/100,000 (range: 83.7-106.3/100,000) and was stable during the period	These figures are much higher than the national average
Aaron et al., 2004[9]	108,000 (young people aged 10-19 years were included)	Verbal autopsy	The average annual suicide rate for young men and women was 58 and 148/100,000, respectively. Suicides were the leading cause of death in this age group	The very high rates noted call for urgent intervention in this group
Abraham et al., 2005[10]	108,873 (elderly i.e., more than 55 years were included)	Verbal autopsy	The average annual suicide rate was 189/100,000 for people over 55 years. Hanging and organophosphorous poisoning were the most common methods	These figures are very high and call for concerted efforts in this group
Prasad et al., 2006[8]	108,873	Verbal autopsy	Average suicide rate was 92.1/100,000. Hanging and organophosphorous poisoning were the most common methods. Acute or chronic stressful life events noted in nearly all subjects	These figures are 8-10 times higher than the national average. Study recommends setting up sentinel centers for suicide monitoring
Bose et al., 2006[7]	108,000	Verbal autopsy	Suicide constituted 11.3% of all deaths across age groups. Hanging and self-poisoning (with pesticides) were the preferred means of suicide	High burden of suicide was noted particularly in the 15-29-year age group
Gajalakshmi and Peto, 2007[12]	38,836 cases of suicide	Verbal autopsy	The average annual suicide rate for men and women were 71 and 53/100,000, respectively. Suicide contributed 9% of overall deaths. The most common mode was self-poisoning	Figures were higher than national average
Soman et al., 2009[11]	647	Verbal autopsy	Suicide constituted 6.6% of all deaths. Male-to-female suicide ratio was 1.7. Among females aged between 15 and 24, suicides constituted more than 50% of all deaths. Hanging was by far the most frequently used method (64%), followed by poisoning (10%)	Suicide underreporting in Kerala found to be less than in other states
Patel et al., 2012[13]	1,100,000	Enhanced verbal autopsy using routine, reliable, representative, RHIME method	About 3% of deaths in individuals aged 15 years or older were due to suicide. 40% of suicide deaths in men and 56% of suicide deaths in women occurred at ages 15-29 years. About half of suicide deaths were due to poisoning (mainly ingestions of pesticides)	Suicide death rates were higher in rural than in urban India Southern states have nearly ten-fold higher age-standardized suicide death rates than some of the northern states
PA studies to determine risk factors and cause of suicide a. Single group cross sectional studies (n=5)
Khan et al., 2005[14]	50 suicide cases, all aged between 15 and 35 years	Semi-structured, self-designed questionnaire. Demographic, clinical and psychiatric risk factors evaluated	Majority did not have psychiatric disorder (76%) or substance use (82%). The presence of precipitating factors and stressful life events are two important reasons for suicide	Most of the families of deceased (68%) had knowledge of their suicidal tendencies but could not prevent it
Behere and Behere, 2008[22]	Not mentioned. All farmers who committed suicide between January 2005 and March 2006 were included	Verbal autopsy method. Structured enquiry form with 52 questions (both open- and closed-ended questions) was used. Interview was done by trained doctor	Farmer suicides were found to be multifactorial. Social reasons and economic loss leading to family conflicts, depression, and substance use were found to be major drivers	Dedicated local communities for suicide prevention and farmer self-help groups proposed as possible interventions
Chavan et al., 2008[19]	101 suicide cases assessed	Semi-structured pro forma used to record sociodemographic profile, psychosocial variables, and treatment details. Interviews done by social worker and a qualified counselor (clinical psychologist)	High rates of suicide among migrant workers noted. Psychosocial stressors found in 60.3% of victims. Psychiatric illness diagnosed only among 33.6%	Prevention efforts in Indian context need to focus on migrants and psychosocial factors
Bastia and Kar, 2009[16]	104 cases of suicidal death by hanging (other modes of death not studied)	No details of proforma given. Information regarding suicide note and circumstantial evidence collected from police and magistrate records	Dowry stress, unemployment, financial and interpersonal conflicts were common reasons for suicide. Mental illness seen in only 4.8% of subjects	Social practices and perceptions are highlighted as priority areas for intervention
Srivastava et al., 2013[20]	100 cases of suicide	Semi-structured questionnaire based on ICD-10, including Life events scale to assess relevant life events (author used standardized Konkani version of the scale)	Psychiatric morbidity present in 94% of decedents. Contact with specialist mental health services and general health services was present in 40% and 50%, respectively	High rates of mental illness among suicide cases noted. Majority do not have a history of prior contact with health services
b. Case control studies (n=5)
Vijayakumar and Rajkumar, 1999[17]	100 completed suicide versus 100 neighborhood controls	Predesigned questionnaire with 178 items including Paykel Scale for Life Events. Psychiatric diagnosis made using SCID. Psychiatrist conducted all interviews	Presence of psychiatric morbidity, positive family psychiatric history, and recent life events were the risk factors identified. Psychiatric morbidity identified in 88% of suicide decedents	Conclusions were that risk factors for suicide are universal and not culture specific
Gururaj et al., 2004[15]	269 completed suicides versus 269 living controls	Semi-structured interview schedule was developed for the study. Four trained research officers with an educational background in sociology/social work/rural development involved in data collection	Previous suicide attempts, interpersonal conflicts, mental illness, economic loss, substance use and unemployment were observed risk factors for suicide. Psychiatric morbidity observed in 43% of cases. Personality disorder noted among 20%	Protective factors noted were good coping, problem solving and positive outlook to life
Manoranjitham et al., 2010[25]	100 completed suicides versus 100 living controls	Semi-structured interview schedule used. Psychiatric diagnosis made using SCID. Trained nurse practitioner and health-care worker conducted the interviews	Psychiatric diagnosis was present in 37% of deceased. Psychosocial stress and social isolation were bigger contributors to suicide than psychiatric morbidity	Psychosocial factors need greater focus in suicide prevention efforts in Indian context than psychiatric morbidity
Kumar et al., 2011[18]	166 completed suicides versus 166 living controls	Semi-structured interview schedule used. Psychiatric diagnosis made using SCID. PSLE used for life events. Two psychiatrists conducted the interview	Psychiatric morbidity seen in 66.7% of cases. Life events and social issues (migration and loneliness) emerged as significant risk factors for suicide
Bhise and Behere, 2016[23]	98 farmer suicide cases versus 98 living controls	Semi-structured interview schedule was used. CAGE questionnaire used to screen for alcohol-use disorder. Psychiatric diagnosis was made using DSM-5. Stressful life events in prior 3 years were enquired into	Economic issues, psychiatric illness, and stressful life events were found to be major drivers of suicide among farmers. Psychiatric morbidity present in 60% of cases, with most receiving no treatment	Socioeconomic and psychological issues are key risk factors among farmers
c. Case report (n=1)
Kulkarni et al., 2015[27]	Single case study	Thorough clinical history including context of attempt, past psychiatric history, substance abuse history, premorbid personality assessment, assimilating various caregivers’ accounts of the deceased, and using DSM-5 for psychiatric diagnoses	Major depressive disorder and inhalant abuse with marital strife was the precipitating factor.

RHIME – Re-sampled household investigation of mortality with medical evaluation; ICD-10 – International Classification of Diseases-10; SCID – Structured Clinical Interview for DSM-III-R; PSLE: Presumptive stressful life events scale; DSM – Diagnostic and statistical manual of mental disorders; PA – Psychological autopsy

Indian reviews/expert opinions on psychological autopsy as an investigative tool (n = 3)

An open review on the PA methods of equivocal death revealed that PA is usually required when the cause and mode of death is uncertain.[28] Usual indications for a PA were suspected suicide, impending criminal investigations, to acquire insurance claims (which are void if death occurred due to suicide within a certain period of policy issuance) and to appeal for malpractice suits. The recommended PA team included magistrate/coroner, medical officer, psychologist, psychiatrist, psychiatric social worker, police investigating officer, and law enforcement authorities. While the interviewer is usually any mental health professional, the interviewee can be persons from the family, friends, co-workers, neighbors, physicians, eyewitnesses, priests or religious figures, and other acquaintances, as appropriate.

Additional sources of information may include suicide notes, medical records, school records, military records, employment records, coroner's report, forensic medical results, police reports, and crime scene analyst reports. Focus areas for assessment include personal views about the deceased person, any behavioral or emotional distress noticed in the deceased, and perceived contextual reasons behind the death. Crucially, the process and outcome of PA are viewed as an expert opinion with probabilities.

A selective review on methods, ethics, and standardization of PA in India[29] identified two types of PA: the suicide PA (SPA) and the equivocal death PA (EDPA). While SPA (to ascertain reasons for suicide) is used in necessary circumstances, the EDPA (to clarify ambiguous deaths) is rarely used in India. The PA can also be used to determine the mode of death (suicide, natural, homicide, and accidental), means of death (head trauma, gunshot injury, heart attack, or suffocation, etc.) as well as the motivation behind death (intentional [on purpose], sub-intentional [an act was meant to harm but not kill], and unintentional [an accident]).

Recommendations included areas of inquiry based on Shneidman's 16 criteria, to conduct interviews within 2–6 months after the death, and to acquire informed consent while collecting information. Starting with information from the crime scene followed by review of police records, interviewing third parties, checking decedent's records (letters, e-mails, journal entries, cell phone records, audio or video recordings, bank accounts, student, or employee records), analyzing relationships, support systems, and the deceased person's occupation was the suggested workflow.

Pertinently, the review also reiterates the nonlegitimacy of PA in Indian courts and its importance in life insurance claims. Ongoing life insurance benefits could affect the quality and quantity of data provided by the informant. The absence of standard methodology, presence of recall bias in relatives, distorted versions from various relatives, and limited collateral records are discussed as further limitations of PA. In an attempt to address the lack of standardization of PA interview procedure,[30] a forensic expert has proposed a brief outline for PA.

Overview of best practice elements in psychological autopsy

Number and selection of informants

The primary choice of informants is always the next of kin; in other words, the spouses/parents/ first-degree relatives. Other informants may include family, friends, co-workers, neighbors, and family physician/mental health-care provider, the last two in light of the large proportion of adults who visit a health-care provider within a month of their suicide.[313233] In case of adolescent suicides, close friends should also be interviewed as they may confide their suicidal thoughts only to peers.[3435] Selection of informants often varies from case-to-case basis and becomes challenging when the individual is living alone.[36] Furthermore, when suitable medical records are available, relevant information from this also should be included.

Approaching informants and timing of interview

Different ways of approaching informants for a PA interview have been described. These include approaching at funeral home,[37] at the home “unannounced,”[31] by letter[35383940] and by telephone.[41] In most of these, the compliance has been generally good (>80%). Telephone calls followed by a letter which contains the detailed information for the study and then visiting the home was suggested by Beskow et al.[42] This may ensure physical availability on the proposed day and has lower rejection rates.

The timing of PA interview has varied between studies, from between a week[3743] to 6–12 weeks[3544] after death. Whenever possible, one should avoid approaching informants and interviewing them close to the anniversary of death, birthday of the deceased, or at family occasions such as impending weddings/religious festivals.

In most Indian PA studies, the timing for interview was between 1 and 3 months after death,[151720] while global literature suggests that the optimum time may lie between 2 and 6 months following death.[42]

Setting of interviews

Both setting and timing of interview needs to be flexible. Informed written consent should be obtained before the interview. The interviews usually take place at the informant's home or a neutral location, depending on mutual convenience. Interviews can last between 2 and 5 h; a single session with breaks or multiple sessions over subsequent days is ideal to minimize interview fatigue.[36] Sometimes, based on initial information, it may emerge that more people need to be interviewed. Accordingly, a second round of interviews may be planned.

Termination of the interview

The interview is usually terminated appreciating the informants for participation as the long interview process is taxing and their queries if anything after completion of the interview needs to be addressed. At times, they might go through grief or suffer from a mental disorder and help should be offered for the same.[345]

Proposed semi-structured format for psychological autopsy interview

Lack of a standard format for interviewing informants in PA has raised legitimate concerns about the reliability/validity of the process. To our knowledge, only one previous study has outlined a detailed semi-structured interview format for PA.[46] A previous Indian study also attempted to draw up a preliminary outline for an interview,[47] whereas a few others[141719202325] used self-designed questionnaires with minimal elaboration. Drawing upon these, as well as the literature discussed in this article, we propose a comprehensive semi-structured format for the PA interview [Appendix 1]. The interview is divided into six parts and 13 items as follows and is designed to be used for each informant separately. A list of potential informants and supplemental measures are listed in Tables 2 and 3, respectively.

Table 2

Potential sources of information for psychological autopsy interview

Individuals who can be selected for interviews	Additional sources of information/records to be verified
Close relatives who have been staying with the deceased during the time preceding death (spouse, children, brother, sister, and parents)	Medical records
	Physical autopsy report
Close relatives who may be residing nearby or is in regular physical or telephonic contact with the deceased	Suicide note (if any)
	Personal diary/log/accounts book
Neighbours	Updated bank passbook
Friends/close acquaintances	Social media accounts (Facebook/Twitter/Instagram/Whatsapp)
Family physician/treating specialist (including mental health professional (if applicable)/health care worker caring for the person/traditional healer	Phone logs/text message history
Religious leaders of groups with whom the deceased had affiliations	Email logs
Village/local leaders	CCTV footage from the area of death
Postman
Investigating police officer

Table 3

Supplementary measures/tools used in psychological autopsy studies

Domain of assessment	Measures/tools	Number of psychological autopsy studies that have used the measure
Assessment of mental disorders	For adults - SCID-I and SCID-II for DSM-IIIR[26,48]	15
	ICD-10[21]	10
	For adolescents-K-SADS-EP[49]	4
Assessment of life events	IRLE[50]	11
	LEDS[51]	2
	LESE[52]	1
	LES[53]	1
	SRRS[54]	2
	List of threatening experiences[55]	1
Suicidal intent/hopelessness	Beck’s suicide intent scale[56]	10
	Beck’s hopelessness scale[57]	6
Depression	HDRS[58]	9
	GDS[59]	2
Psychological constructs of personality	Personality assessment schedule[60]	6
	NEO-FFI[61]	4
	SCID-II for DSM IV[48]	3
	Standardized assessment of personality[62]	1
Social support	DSSI[63]	11
	SPSI[64]	3
	Social networks and social support scale[65]	1
	Bille-brahe social support scale[66]	1
Impulsivity	Dickman impulsivity inventory[67]	7
	Barratt impulsiveness scale[68]	4
	Impulsivity rating scale[69]	1
Coping response	CRI[70]	3
	Brief COPE-28[71]	1
Childhood adversity	CECA scale[72]	1
Spiritual and religious beliefs	The royal free interview for religious and spiritual beliefs[73]	1
Loneliness	University of California Los Angeles loneliness scale[74]	1
Activities of daily living (ADL)	Instrumental activities of daily living scale and the physical self-maintenance scale[75]	4
Physical health status	Cumulative illness rating scale[76]	1
Family function	Family APGAR[77]	1
Aggression	Overt aggression scale[78]	1
Anxiety	State trait anxiety inventory[79]	3
Work demands	COPSOQ II[80]	1
	Swedish demand-control-support questionnaire[81]	1

ICD-10 – International Classification of Diseases-10; SCID – Structured Clinical Interview for DSM-III-R; PSLE: Presumptive stressful life events scale; DSM – Diagnostic and statistical manual of mental disorders; PA – Psychological autopsy; K-SADS-EP – Schedule for Affective Disorders and Schizophrenia for School-Age Children, Forms E and P; IRLE – Interview for Recent Life Events; LEDS – Life events and difficulties schedule; LESE – Life event scale for the elderly; LES – Life Experiences Scale; SRRS – Social Readjustment Rating Scale; HDRS – Hamilton depression rating scale; GDS – Geriatric depression scale; NEO-FFI – NEO five-factor inventory; DSSI – Duke social support index; SPSI – Social problem solving inventory; CRI – Coping response inventory; CECA – Childhood experience of care and abuse; APGAR – Adaptive partnership growth affection and resolve; COPSOQ – Copenhagen psychosocial questionnaire- long version

Any conclusion about a PA should be restricted to opinions on whether the death was due to suicide or not, if the autopsy was conducted to classify an equivocal death, or probable cause/reasons for suicide if the autopsy was conducted to investigate the reasons behind suicide. Investigators should refrain from commenting on other possibilities such as accidental death or homicide.

DISCUSSION

Overview of Indian research using psychological autopsy

More than two decades after the earliest Indian PA study[17] was published, the volume of PA literature from India still remains very low. Further, all the available PA literature is related to suicide psychological autopsies, done on cases where the modality of death is not in question. No study or report so far has focused on EDPA.

Of the available 22 PA-related research articles from India, 18 were original research, three were reviews or expert opinions, and one was a case report. Of the 18 original PA studies, five studies[1517182325] used a case–control design, while five were single group studies[1416192022] on suicide deaths. Eight studies[678910111213] used the verbal autopsy method in selected villages to ascertain the average annual suicide rate while correcting for underreporting. One of these was a nation-wide study.[13]

Nearly all the verbal autopsy studies estimated that the average annual suicide rate is higher than the reported national average. As most of them were conducted many years ago, it may be worthwhile repeating these studies now, in light of changes in suicide-related legislation and practice.

Common risk factors noted for suicide in the five case–control studies[1517182325] were the presence of mental illness, prior suicide attempts, interpersonal conflicts, substance use, financial loss, stressful life events, and solitary living arrangement. The demographic risk factor for suicides was not consistently observed.

Interestingly, there was a wide variation in observed prevalence rates of mental illness among the deceased; figures ranged from 37% to 88%.[1725] In general, studies that employed a structured diagnostic instrument noted higher prevalence rates than those without. Nonetheless, stressful life events were also a significant contributor to suicides in all these studies.

Five studies used single group (suicide) cross-sectional design.[1416192022] Common findings were strikingly similar and included the occurrence of stressful life events among majority of attempters.[14161920] Here too, the prevalence of mental illness was disparate; figures ranged from as low as 4.8%[16] to as high as 94%.[20]

Together, no clear conclusions can be drawn but average rates of mental illness among Indian suicide decedents appear to be comparable to pooled estimates from low- and middle-income countries.[82] It is also unclear whether the observed high rates of stressful life events among the deceased are a cause or effect of psychiatric morbidity, and further research is needed to answer this question. Nonetheless, it appears safe to conclude that suicide prevention efforts in the Indian setting must focus on both control of psychiatric morbidity and psychosocial risk factors.

Evidence informed methodological best practices for psychological autopsy

Regarding methodological aspects, available evidence suggests that selection of informants must be decided on a case-to-case basis based on proximity to deceased, especially in the days leading up to the death. All attempts must be made to collect information from as many sources as possible.

The optimal time for interviewing appears to be between 2 and 6 months after death to strike the right balance between grief resolution and recall bias. A single interview may last anywhere between 2 and 5 h but may extend longer depending on the informant's productivity. As interview fatigue may set in during this long period, it is advisable to pace it comfortably, give adequate breaks in between, or have multiple sittings. At all times, it is important to respect the integrity of the deceased, allow processing of emotions, and give informants adequate time to respond to questions.

PA generates sensitive information, and hence, safeguarding the confidentiality of the information generated is of paramount importance. As a confidence-building measure among informants, PA investigators can brief them upfront on the steps taken to ensure confidentiality, data safety, and anonymity. Steps include destroying any written information, audio or videotapes in case of telephonic interviews as well as case notes following completion of the report and analysis, limiting access to rooms where sensitive information is stored, and blinding other staff to the identity of the cases.[83] At all times, the privacy and confidentiality of the interviewee and the integrity of the deceased must be respected.[84]

There could be several barriers to accessing supplemental sources of information outlined in Table 2. In many jurisdictions, the medical records of the deceased are considered to be property of designated next of kin. Previous investigators have recommended approaching informants with a letter explaining the scope and nature of request along with the necessary approvals (ethics approval, if it is a SPA) and obtaining the consent of the kin to obtain their cooperation.[45] Even then, some agencies or systems involved may be reluctant to release information due to concerns about legal liability. Similar concerns may apply to other sources of information in PA interview [Table 2]. We are of the opinion that a collaborative approach with the informants and working with them to surmount administrative hassles required to obtain necessary records will be the optimal approach to a PA procedure.

Findings of the PA procedure could have implications for life insurance as well as health insurance. As discussed earlier,[2829] life insurance claims are not usually honored if the suicide occurred within a certain period of policy commencement. At present, most health insurance policies also do not honor medical claims for injuries due to suicidal attempt or self-harm; this is even after the Mental Health Care Act 2017[85] has come into force. By placing suicide or attempted suicide on record, the family members of the decedent in case of suicide, or the concerned individual and family in case of attempted suicide, may lose reimbursement claims from a health insurance policy, which could potentially be a serious financial blow. Hence, the need for a meticulous approach to performing PA cannot be overstated.

The proposed semi-structured format for PA interview has been prepared after going through the extant literature. It can be used for both suicide psychological autopsies as well as equivocal death autopsy and may assist investigators in delineating the probable mode and cause of death. The interview covers both static, stable, and dynamic risk factors as well as protective factors.[86] Investigators would be well advised not to base their conclusions on the probable mode of death (suicide, accidental death, homicide, or natural death) on any one or two factors, however important they may seem. Instead, they must follow a systematic approach including close examination of the circumstances surrounding death and the decedent, results of the physical (forensic) autopsy, physical examination of site of death, and the relative balance of risk and protective factors, supplemented by tools wherever necessary, similar to a structured suicide risk assessment process.[87] This will facilitate the formulation of structured professional judgment, which is what the proposed interview format seeks to espouse.

The items in the suggested guide must be viewed as question leads. It is important to allow respondents to build their narrative account of their relationship to the deceased and their understanding of the deceased's mental health condition and behavior in the days preceding death. Then, these narratives may be analyzed further to assist clinical impressions and reporting. The advantages of such a semi-qualitative approach to PA interviews and the perils of excessive standardization have been pointed out earlier.[4] It is for this reason that we have called it a semi-structured (and not a structured) interview guide.

The present review has certain limitations. We did not attempt a meta-analysis due to the heterogeneity in outcomes evaluated and methods employed across studies. It is quite possible that studies published in the gray literature or not indexed in the included databases may have been missed out, but every effort was made to include relevant studies by searching cross-references of included studies.

CONCLUSION

PA is a useful investigative tool to ascertain the mode of death in equivocal deaths by examining the factors surrounding death and the mental health status of the deceased. However, in our country, it has more commonly been used as a research tool to examine risk factors associated with suicidal deaths and to determine the reason for committing suicide. Given the wide variations in rates of psychiatric morbidity among suicide decedents, the complex interplay between psychiatric morbidity and the other major risk factor for suicide in the Indian setting, namely, stressful life events, deserve more research attention.

From a reliability/validity perspective, it is hoped that the proposed semi-structured format will assist and encourage the use of PA as both an investigative and research tool. Future validation studies using this instrument will enhance its utility. This will also improve our understanding of the complex multidimensional behavior that is suicide and inform suicide prevention programs.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.