شدت قصد خودکشی، روش و رفتار پیشین برای یک عمل خودآسیبی: مطالعه مقطعی بازماندگان خودآسیبی مراجعه کننده به بیمارستان سوم در میسور، جنوب هند

Abstract Background Rates of self harm are high in south India, but little is known about the relationship between antecedent behaviour, suicidal intent and method. Aims To identify clinical, social and behavioural antecedents preceding an act of self-harm. Methods 200 participants, consecutively presenting with deliberate self harm to a hospital in south India, were interviewed. Socio-demographic and clinical characteristics were recorded, together with behaviours preceding self-harm. The Pierce Suicidal Intent Scale and Mini International Neuropsychiatric Inventory were administered.

. Introduction WHO estimates that about 170,000 deaths by suicide occur in India every year with especially high risks in the young and in south India, where about half of suicide deaths are due to poisoning (principally ingestion of pesticides) (World Health Organization, in press and Aaron et al., 2004). Suicide rates recorded in the south India region are much higher than the national estimate and this is a major public health concern (Aaron et al., 2004 and Prasad et al., 2006) A wide range of clinical, social and behavioural factors are known to influence the risk of deliberate self-harm (DSH). In India, younger age, rural residence, use of alcohol, previous self-harm, presence of psychological distress, socio-economic adversity and access to poisons are known risk factors (Aaron et al., 2004, Prasad et al., 2006 and Radhakrishnan and Andrade, 2012). There are known to be differences in suicide and DSH behaviour in India compared to western countries. These include a high incidence of organophosphate insecticide (OP) poisoning, a larger proportion of married women and a low incidence amongst elderly people. Interpersonal relationship problems, life events and psychological distress (as opposed to psychiatric disorders) are common antecedents (Prasad et al., 2006, Radhakrishnan and Andrade, 2012 and Khan, 2002). Studies in western populations show a robust association between suicidal intent at the time of DSH and eventual suicide (Harriss and Hawton, 2005 and Skegg, 2005). In contrast, little is known about the role of suicidal intent and its interaction with other risk factors in India. Factors that are believed to be protective (e.g. religion and faith) or predictive of higher risk (e.g. use of alcohol, expression of intent and writing a suicide note) in the west have not been systematically studied in India. Examining cultural and region specific antecedent behaviours and their relationship with suicidal intent and method of self-harm is a step towards more accurate identification of individuals at high risk of suicide. It is also important in understanding the specifically Indian phenomenon of DSH and suicide by OP poisoning

4. Results 4.1. Demographic characteristics Of the 200 participants, 115 were interviewed within 24 h, 162 within 3 days and the remainder within a week of carrying out the act of DSH. The socio demographic, educational and occupational characteristics of the sample are set out in Table 1. The participants were aged between 18 and 70 yrs with a mean age of 29.1 yrs (SD 10.5). 101 (50.5%) were male and 99 (49.5%) were women. The sample broadly matched the characteristics of the local population according to the Karnataka 2011 census (2011 census: Karnataka population, 34% urban, 76% rural). There were slightly fewer Muslims and Christians in the sample than in the local population (2011 census: Karnataka population 12.2% Muslim, 3.1% Christian) (Diector of Census Karnataka 2011). Table 1. The characteristics of the study sample. Categories n (%) Mean age (SD) 29.1 (10.5) Gender Male: Female 101: 99 (50.5%: 49.5%) Occupation Professional 23 (11.6) Skilled workers 13 (6.5) Semiskilled worker 17 (8.5) Un skilled worker 69 (34.7) Unemployed 8 (4.0) Student 18 (9.0) Home maker 51 (25.6) Education Graduate/post graduate 11 (5.5) Intermediate/post high school 27 (13.5) High school 58 (29.0) Middle school 39 (19.5) Primary school 22 (11.0) Illiterate 43 (21.5) Religion Hindu 179 (89.5) Muslim 18 (9.0) Christian 3 (1.5) Family type Nuclear 131(65.5) Joint/multigenerational 68 (34.0) Living alone 1 (0.5) Locality Urban 54 (27.0) Rural 146 (73.0) Marital status Single 67 (33.5) Married 127 (63.5) Widowed 5 (2.5) Divorced 1 (0.5) Table options There was no statistically significant relationship between marital status and urban/rural location nor between marital status and OP poisoning. Over all, the participants were most commonly Hindu, married, living in nuclear families and from the rural area. 4.2. The act of self harm In 74 cases (37%), the method of self harm was OP poisoning, in 43 (21.5%) non OP poisoning, in 75 (37.5%) tablet consumption and in 8 (4%) other (drowning, hanging and self cutting). Non-OP poisoning refers to ingestion of household chemicals other than organo-phosphate pesticide. Pesticide poisoning was the most common method of self-harm in men and tablet consumption in women. There were no cases of self-immolation. In 44 cases (22%), the act of self harm occurred with concurrent alcohol consumption. 21 participants (10.5%) reported previous deliberate self harm. 18 described a single previous episode of DSH, but 3 participants (1.5%) reported three or more previous episodes. 4.3. Psychiatric diagnosis Majority of the participants, 136 (68%) did not meet criteria for a MINI psychiatric diagnosis. Depression was the most common mental disorder in the study population (n = 31, 15.5%). Prevalence of other psychiatric disorders was: Alcohol dependence (n = 20, 10%), Dysthymia (n = 6, 3%), Panic disorder (n = 4, 2%), Bipolar disorder (n = 1, 0.5%), Post traumatic stress disorder (n = 1, 0.5%), and Generalised anxiety disorder (n = 1, 0.5%). 4.4. Intent The characteristics of the study sample and level of suicidal risk as determined by the Pierce Suicidal Intent Scale (PSIS) are provided in Table 2. 14 (7%), 88(44%) and 98(49%) of the subjects were classified into low, medium and high risk according to their PSIS scores. PSIS scale is constituted of three categories: circumstances, self-report and medical risk. The mean score for each of the category for the study group and low, medium and high risk groups are provided in Table 3. Table 3 illustrates that there is consistency across the sub-categories, which would tend to suggest that the PSIS is reliable instrument to measure the suicidal intent in this population, and is not being skewed by cultural differences. For e.g. very few subjects in the study left a suicide note, which might suggest a lack of applicability of the scale in this population. However, the categories of the PSIS scale show remarkable internal-consistency in this study sample. Table 2. The characteristics of the study sample and level of suicidal risk as determined by the Pierce Suicidal Intent Scale (PSIS). Variable PSIS score level p Low %/SD Medium %/SD High %/SD Age in yrs mean (SD) 27.2 13.2 29.2 10.7 29.5 9.7 0.6 Income/yr in Rs. 31,250 38,207 29,305 42,155 30,688 27,501 0.9 Gender M:F 6:8 5.9:8.1 44:44 43.6; 44.4 51:47 50.5:47.5 0.9 Occupation Professional 1 4.3 7 30.4 15 65.2 0.013* Clerical, shop-owner, farmer 1 7.7 10 76.9 2 15.4 Semiskilled worker 2 11.8 5 29.4 10 58.8 Un skilled worker 2 2.9 28 40.6 39 56.5 Unemployed 1 12.5 0 0.0 7 87.5 Student 2 11.1 9 50.0 7 38.9 Home maker 5 9.8 28 54.9 18 35.3 Education Graduate/post graduate 0 0.0 5 45.5 6 54.5 0.6 Intermediate/post high school 4 14.8 14 51.9 9 33.3 High school 4 6.9 27 46.6 27 46.6 Middle school 3 7.7 18 46.2 18 46.2 Primary school 1 4.5 10 45.5 11 50.0 Illiterate 2 4.7 14 32.6 27 62.8 Religion Hindu 12 6.7 81 45.3 86 48.0 0.3 Muslim 1 5.6 7 38.9 10 55.6 Christian 1 33.3 0 0.0 2 66.7 Family type Nuclear 10 7.6 61 46.6 60 45.8 0.6 Joint 4 5.9 26 38.2 38 55.9 Living alone 0 0.0 1 100.0 0 0.0 Locality Urban 8 14.8 23 42.6 23 42.6 0.009* Rural 6 4.1 65 44.5 75 51.4 Marital status Single 7 10.4 26 38.8 34 50.7 0.4 Married 6 4.7 61 48.0 60 47.2 Widowed 1 20.0 1 20.0 3 60.0 Divorced 0 0.0 0 0.0 1 100.0 PSIS scores categorised in to mild, moderate and severe. * Statistically significant after adjusting for age, gender and method of self harm. Table options Table 3. Consistency of severity scores across PSIS categories. Categories of the PSIS scale Level of the Risk N Mean (SD) 95% CI Min Max Circumstances Low 14 1.0 (1.30) 0.25–1.75 0.00 4.00 Medium 88 3.05 (1.72) 2.68–3.41 0.00 7.00 High 98 6.02 (1.92) 5.64–6.41 1.00 10.00 Total 200 4.36 (2.47) 4.02–4.71 0.00 10.00 Self r report Low 14 0.43 (0.65) 0.06–0.80 0.00 2.00 Medium 88 2.81 (1.60) 2.48–3.16 0.00 7.00 High 98 5.72 (1.72) 5.38–6.07 2.00 9.00 Total 200 4.08 (2.36) 3.75–4.41 0.00 9.00 Medical risk Low 14 0.57 (0.86) 0.08–1.06 0.00 2.00 Medium 88 1.69 (1.14) 1.45–1.93 0.00 4.00 High 98 2.30 (1.20) 2.06–2.54 0.00 4.00 Total 200 1.91 (1.24) 1.74–2.08 0.00 4.00 Table options Unskilled labourers and professionals had higher PSIS score (adjusted for age, gender and the method of self-harm) when compared to skilled labourers and the unemployed (p = 0.013). Those from rural areas had higher PSIS score when compared to urban dwellers, irrespective of the method of self-harm, after adjustment for age and gender (p = 0.009). Major Depressive Disorder (MDD) was associated with significantly higher levels of PSIS score when compared to other diagnoses and with no mental illness (p = 0.05). However, previous self-harm, concurrent use of alcohol and expression of an intent (verbal and written combined) were not significantly associated with higher PSIS score or any particular method of self-harm. 4.5. Antecedent behaviour Table 4 provides the distribution of the studied behaviours in the week and in the four weeks period preceding the act of self-harm with corresponding PSIS scores for the same. Absenteeism from work, visit to the family shrine, increase in the intake of alcohol, making a pilgrimage and increase in smoking were the most common of the reported behaviours (in that order) in the month preceding the act of self-harm. A trend was observed for these behaviours to be more frequent within the week preceding the act of self-harm, though the differences failed to reach statistical significance (Chi square test p = 0.5). Only 10 participants in the study had seen a doctor (including a psychiatrist) in the month before the act. Three participants had a mental health consultation with a doctor in the month before the act. Three participants had written a suicidal note and one participant with known cannabis use reported having increased the level of consumption prior to the act of self-harm. Table 4. Antecedent behaviour and distribution of PSIS score. Behaviour Number of subjects with the reported behaviour in the previous PSIS score in those with the behaviour in the preceding 4 weeks PSIS score in those without the behaviour in the preceding 4 weeks p 0–1 Week 1–4 Weeks Mean (SD) Mean (SD) Expression of intent to the family 9 3 9.75 (4.5) 10.35 (4.5) 0.7 Expression of intent to a friend 6 0 12.2 (1.7) 10.3 (5.6) 0.3 Expression of intent to a health professional 1 0 13.0- 10.3 (4.6) 0.6 Writing a suicide note 4 0 14.0(5.8) 10.2 (4.5) 0.1 Visit to the family shrine 21 5 10.7 (4.4) 10.2 (4.6) 0.6 Religious pilgrimage 9 7 10.5 (4.3) 10.3 (4.6) 0.9 Partition of assets 5 5 11.3 (3.2) 10.3 (4.6) 0.5 Absenteeism from work* 16 10 12.3(3.6) 10.2 (4.6) 0.04 Visited a doctor 3 4 10.0 (2.9) 10.3 (4.6) 0.8 Visited a psychiatrist 1 2 11.0 (3.6) 10.3 (4.5) 0.8 Increased use of alcohol 14 7 9.8 (4.4) 10.4 (4.5) 0.6 Increased level of smoking 11 4 9.5 (4.5) 10.4 (4.5) 0.4 Made a will 1 1 12.5 (2.1) 10.3 (4.5) 0.5 Pierce Suicide Intent Score (PSIS) is analysed as a continuous variable between those with and without the behaviour in the preceding 4 weeks by Independent t test. * The observed association of absenteeism with higher level of PSIS score was not significant after adjusting for age, gender and the method of self-harm. Table options 4.6. Intent and antecedent behaviour Mean PSIS scores in those with and without the corresponding behaviour in the preceding 4 weeks is provided in Table 4. The observed association of absenteeism with higher level of PSIS score was not significant after adjusting for age, gender and the method of self-harm. None of the other antecedent behaviours showed any significant association with the severity of PSIS scores (see Table 3). There were no significant differences in the level of PSIS score and psychiatric diagnoses (p = 0.2) and history of previous self-harm (p = 0.4). 4.7. Intent and method of self harm There were no statistically significant differences in the level of PSIS score (a measure of suicide intent) between different methods of self-harm (ANOVA p = 0.5). The mean PSIS score for hanging (n = 8) was 10 (SD 4.4, 95% CI 6.29–13.71); OP poisoning (n = 74) was 10.92 (SD 4.478, 95% CI 9.88–11.96); non OP poisoning (n = 43) was 9.79 (SD 4.673, 95% CI 8.35–11.23) and for the tablet consumption (n = 74) was 10.04 (SD 4.52, 95% CI 9.00–11.08). 4.8. OP and non-OP poisoning A sub-group analysis was carried out to compare those who had ingested OP pesticides (n = 74) with those using other methods of self-harm (n = 126). OP poisoning was significantly more common in rural residence (p = 0.02), men (p = 0.002) and professionals (p = 0.02). There were no other statistically significant differences between these groups in socio-demographics, level of PSIS score and any particular behaviour antecedent to self harm. 4.9. Antecedent behaviour and method of self harm The frequency of certain individual behaviours across different methods of self-harm was low, restricting behaviour versus method analysis. Therefore, methods of self-harm were grouped into pesticide poisoning (n = 117) and other methods of self-harm (n = 83). The latter group included self-cutting, self-immolation, tablet consumption, hanging and drowning. There were no statistically significant differences in frequency of occurrence of these behaviours between these two groups.