Self-harm and self-poisoning in southern India: choice of poisoning agents and treatment
Corresponding Author Anuradha Bose, Department of Community Health, Christian Medical College, Vellore, India 632002. Tel.: +91 416 2284207; Fax: +91 416 2262268; E-mail: firstname.lastname@example.org
Objective To record cases of suicide and attempted suicide among a population of 108 000 people living in a primarily rural area of southern India, with the aim of guiding policies and strategies to restrict access to poisonous compounds at community level.
Method Community-based surveillance over a period of 2 years.
Results and conclusion The overall suicide rate was 71.4 per 100 000 population; the highest burden was among men. Most people died through hanging (81, 54%) and self-poisoning (46, 31%). Of the 46 who died from self-poisoning, 78.3% had taken pesticides and 19.7% had eaten poisonous plants. Eighty per cent of the self-poisoning cases obtained the poisonous substance in or in close proximity to the home, highlighting the importance of safe storage in the domestic environment. Of the 110 fatal and non-fatal self-poisoning cases, 87 (57.5%) were taken for treatment; 50 (57.4%) went to government hospitals and 37 (42.5%) to private facilities. This indicates the importance of including the private sector in the efforts to improve case management. Furthermore, the fact that 31 (67%) of the self-poisoning patients, who eventually died, were alive after 4 h provides an incentive to focus on improved case management and access to health services.
Automutilation et auto-intoxication dans le sud de l’Inde: choix des agents d’intoxication et traitement
Objectif: Enregistrer les cas de suicide et de tentative de suicide au sein d’une population de 108.000 personnes vivant principalement dans une zone rurale du sud de l’Inde, dans le but de guider les politiques et stratégies visant à restreindre l’accès à des composés toxiques au niveau de la communauté.
Méthode: Surveillance basée sur la communauté sur une période de deux ans.
Résultats et conclusions: Le taux global des suicides était de 71,4/100.000 habitants; la charge la plus élevée était chez les hommes. La plupart des gens sont morts par pendaison (n = 81; 54%) et par auto-intoxication (n = 46; 31%). Sur les 46 morts par auto-intoxication, 78,3% avaient pris des pesticides et 19,7% avaient consommé des plantes vénéneuses. 80% des cas d’auto-intoxication avaient obtenu la substance toxique à l’intérieur ou à proximité de la maison, ce qui souligne l’importance de la sécurité du stockage dans l’environnement domestique. Sur les 110 cas d’auto-intoxication fatale et non fatale, 87 (57,5%) ont été amenés pour traitement, 50 (57,4%) sont allés dans des hôpitaux publics et 37 (42,5%) dans des établissements privés. Cela montre l’importance d’inclure le secteur privé dans les efforts visant à améliorer la prise en charge des cas. En outre, le fait que 31 (67%) patients morts par auto-intoxication étaient encore en vie 4 heures après, incite à se concentrer sur l’amélioration de la prise en charge des cas et sur l’accès aux services de santé.
Auto-agresión y auto-envenenamiento en India del sur : escogencia de los agentes venenosos y tratamiento.
Objetivo: Registrar los casos de suicidio e intento de suicidio en una población de 108,000 personas viviendo en un área rural primaria del sur de la India, con el fin de guiar las políticas de acceso restringido a compuestos venenosos a nivel comunitario.
Método: Vigilancia basada en la comunidad a lo largo de un periodo de dos años.
Resultados y conclusiones: La tasa total de suicidio fue de 71.4 por 100,000 habitantes; la mayor carga fue entre hombres. La mayoría de las personas murieron por ahorcamiento (81, 54%) y auto-envenenamiento (46, 31%). De los 46 que murieron por auto-envenenamiento, un 78.3% habían tomado pesticidas y 19.7% habían comido plantas venenosas. Un 80% de los casos de auto-envenenamiento habían obtenido las sustancias venenosas en o cerca de sus hogares, subrayando la importancia de un almacenamiento seguro dentro del ambiente doméstico. De los 110 casos fatales y no fatales de auto-envenenamiento, 87 (57.5%) fueron tomados como tratamiento; 50 (57.4%) fueron a hospitales gubernamentales y 37 (42.5%) a centros privados. Esto indica la importancia de incluir el sector privado en los esfuerzos para mejorar el manejo de casos. Más aún, el hecho de que 31 (67%) de los pacientes con auto-envenenamiento, que eventualmente murieron, estaban vivos tras 4 horas, provee un incentivo para enfocarse en mejorar el manejo de casos y el acceso a centros sanitarios.
The commercialization of agriculture taking place in most developing countries increases the farmers’ dependency on agrochemicals and may increase availability of highly toxic pesticides in rural households (Ecobichon 2001). As a result, self-poisoning with pesticides has become a significant medium for self-harm, where many such acts, at moments of acute distress, have fatal and sometimes unintended consequences (Conner et al. 2005; Konradsen et al. 2006). It is now estimated that, death by self-poisoning, especially with pesticides, accounts for 30% of all suicides globally (Gunnell et al. 2007). A range of underlying factors such as domestic problems, alcohol or drug addiction, emotional distress, depression, physical illness, failed exams, social isolation or financial hardship have been identified, that make individuals at risk for self-harm (van der Hoek & Konradsen 2005; Konradsen et al. 2006). Furthermore, the restricted access to a well-equipped health facility, where pesticide poisoning can be managed, results in high rates of mortality (Eddleston et al. 2002).
In India, as in the rest of the region, efforts are made to improve the clinical management of poisoning patients and to use regulatory policies to restrict access to the most toxic compounds. There have been very few community-based studies published on the kind of pesticides that are used, and the healthcare access in cases of suicides and attempted suicides.
The results presented here complement earlier studies among the same community aimed at establishing the burden of deaths following intentional and unintentional injuries (Joseph et al. 2003; Aaron et al. 2004; Bose et al. 2006). In this study, the focus is on suicide and suicide attempts, especially after self-poisoning. In addition to providing an update on the incidence of deaths following self-harm in this South Indian study community, this report aims to provide community-based information on broad categories of poisonous agents most important for self-poisoning; how individuals obtain access to poisonous agents; and it describes treatment seeking behaviour.
Identification of self-harm cases
The findings presented here are the results of an ongoing surveillance system established in Kaniyambadi Block by the Christian Medical College (CMC), Vellore, India, covering a population of approximately 108 000 where all births and deaths within the community are recorded prospectively. The mortality information is based on the use of the verbal autopsy method and has been described earlier by Aaron et al. (2004). The surveillance system is tiered, with information being gathered by health workers living in the villages, the information then being quality-assessed by a team including the village health nurse, the public health nurse and a medical doctor. A village health worker covers a population of approximately 5000 individuals and obtains information from health facilities, healthcare outreach programmes, community organizations, observations and systematic household visits. Annually, 10% of all obtained data are validated by independent observers from the Department of Community Health, CMC.
For this study, the community-based surveillance system in place was used to identify all deaths from self-harm and, through further training and supervision of the health workers living in the villages, efforts were made to also identify non-fatal suicide attempts. Information was obtained from the individual self-harming or by persons closely associated with the person. However, based upon years of research among the study population, it was acknowledged that identifying all cases of non-fatal self-harm attempts would not be possible and many cases of self-harm and self-poisoning, especially the less serious, will have been missed; but the study design is likely to have captured all fatal cases of self-harm.
Self-poisoning was defined as the deliberate ingestion of any toxic compound, and included death after self-poisoning and non-fatal self-poisoning attempts. A questionnaire was devised, with demographic details, mode of self-harm and details of treatment sought, if any. In case of self-harm with poisoning, questions were asked to find out the type of poison, and how the poison was obtained and stored. Interviews were conducted in or near the homes, with the individual concerned or with family members, within 7 days of the episode being reported by the village health worker, and were independently verified by a senior field worker. Wherever available, the containers of the poisonous substance used for self-harm were examined to obtain more specific details on the nature of the poison.
For cases by reason of poisoning by pesticide, an attempt was made to identify the pesticide. Where a container could be found or traced, the chemical name was identified. The two most commonly used pesticides for each plant were identified by detailed questioning of farmers, pesticide shop owners and the men who spray the fields. In cases where the container could not be traced, the plant for which the pesticide was used was identified, and the pesticide assumed to be one of the two most popular brands.
Health services in the study area
The health facilities in the study area, all within a distance of 30 km by road, range from private practitioners who have no or limited admission facilities, to a tertiary government medical college facility, where treatment is free. There are three primary health centres, each serving a population of approximately 30 000. Not all primary health centres operate 24 h a day. In addition, there are the secondary- and tertiary-level hospitals of CMC, a non-profit organization providing medical services for free or at subsidized rates.
Over the 24-month study period, a total of 149 suicides were recorded with 66 females and 83 males. Hundred and three people who had attempted suicide but survived were identified. The overall suicide rate in the 2 years between January 2006 and December 2007 was 120.3/100 000. The rate of suicides was higher in men (130.9/100 000) than in women (109.7/100 000) and in keeping with previously reported figures from this group (Bose et al. 2006; Aaron et al. 2004). Only among the 15- to 24-year age group were females significantly more likely to commit suicide, with a rate of 148.5 vs. 82.7 for men, a 2.53 (1.47–4.35; χ2 = 12.92, P = 0.00**) times higher chance. In all other age groups, males had higher rates, with a peak among those older than 65 years, with a rate of 302.4, a 5.4 (2.22–13.51; χ2 = 18.53, P = 0.00**) times higher risk of committing suicide.
Mode of suicide
Poisoning and hanging are the most common modes of suicide; 110 used poisoning (43.7%) and 107 hanged themselves (42.5), followed by 20 burning (7.9) and 14 drowning (5.6%; Table 1). There was an association between mode of suicide and survival, the lethality being greater in hanging, drowning and burning, and least in self-poisoning (χ2 = 26.71, df = 3, P = 0.000).
Table 1. Mode of self-harm and survival status among 252 cases registered over a 24-month period in Kaniyambadi Block, India
|Hanging||26 (24.3)||81 (75.7)||107 (42.5)|
|Poisoning||64 (58.2)||46 (41.8)||110 (43.7)|
|Drowning||4 (28.6)||10 (71.4)||14 (5.6)|
|Burning||8 (40.0)||12 (60.0)||20 (7.9)|
|Fall from height||1 (100.0)||0||1 (0.4)|
|Total||103 (40.9)||149 (59.1)||252|
In addition to the 46 cases of death from self-poisonings, 64 cases of non-fatal self-poisonings were recorded. Table 2 shows the various types of poisons used. Pesticides were the preferred agents, 68 (61.8%) both for suicides and attempts, followed by poisonous plants in 21 cases (19.1%); these were primarily yellow oleander (Thevetia peruviana) and Cleistanthus collinus, known as oduvanthalai (Nillipalai) in India. Pharmaceutical tablets were used in 18 cases (16.4%) and other poisons including household chemicals were used in 2 cases; for one case the agent remained unknown (2.7%). Of those who died after consuming poisons, 36 (78.3%) did so after ingesting pesticides and 9 (19.5%) after ingesting plant poisons. Interestingly, all of those who used pharmaceutical tablets to attempt suicide survived, which makes it worthwhile to seek treatment in the event of anyone attempting suicide using this method.
Table 2. Types of poisons used among fatal and non-fatal self-poisoning cases at Kaniyambadi Block, India
Identification of the pesticide used was performed by visiting a local shop selling pesticide, getting a list of the first and second choices of pesticide among the farmers in the area, for the crops under cultivation. When direct information was not available from the next of kin, it was determined what plant the pesticide that was consumed was used for. It was then determined that it was likely to be the first or the second choice that had been listed from the dealer. Details of the pesticides that are in use in the area are given in Table 3. Almost all the pesticides mentioned were class Ia, Ib or II after the WHOs classification 1 Only one pesticide was class III and only one was class U (Table 2).
Table 3. Place of death after self-harm in Kaniyambadi Block, India
|Hospital||33 (22.1)||20 (52.6)||16 (44.4)|
|Home||73 (49.0)||3 (7.9)||8 (22.2)|
|Field||16 (10.7)||7 (18.4)||5 (13.9)|
|On way to health facility||10 (6.7)||8 (21.1)||7 (19.4)|
In approximately 80% of the cases of self-poisoning, the agent was obtained from within the home (52.7%) or the premises around, i.e. field or garden (28/2%). Only in 13.6% of all poisons and 17.6% of poisoning by pesticides were the poisons obtained from a shop just before the self-poisoning. In the remaining cases, it was not clear where it was obtained.
Treatment-seeking behaviour and place of death
Of the 149 cases of suicide, 104 were not taken for care, or died en route. They were either found dead at the site of the ingestion of the poison, in the premises around the home or in the fields or died on the way to hospital. Of the 110 cases of self-poisoning, 87 (79.1) were taken for treatment with 50 (57.5%) going to government hospitals and 37 (42.5%) to private facilities. This distribution was even between pesticides and the other self-poisonings.
Table 3 shows the place of death for the different modes of attempts and suicides and Table 4 shows the gap between attempt and death, as estimated by the family. When comparing death from pesticide with the use of other poisonous agents, a relatively higher proportion (16 cases, 44.4%) died within 4 h of the attempt. All the cases of death by poison who died within 4 h of the self-harm had ingested pesticides.
Table 4. Time gap between suicide and death Kaniyambadi Block, India
|0–4||84 (56.7%)||16 (34.0%)||16 (44.4%)|
|4–8||19 (12.7%)||7 (14.9%)||5 (13.9%)|
|8–12||14 (9.3%)||10 (21.3%)||6 (16.7%)|
|≥12||32 (21.3%)||14 (29.8%)||9 (25.0%)|
Eighty-four (56.7%) of those who died did so within 4 h of having self-harmed. This implies that 43.3% of those who subsequently died were alive for 4 h or more after the intentional injury. This figure is higher for all poisons (65%), and a little less (55%) among those who took pesticides. Those who were found alive and taken for treatment within 4 h had 3.06 times better chances of surviving than those who did not get treatment (Table 5). Intuitive as this sounds, it clarifies the need for attempting to treat, as early and as effectively as possible, those who are found alive.
Table 5. Treatment time and survival among poisoning cases: people who get early treatment within 4 h have a 3.06 times better chance of surviving than those who do not
|<4 h||49 (81.7)||16 (59.3)||65 (74.7)||4.94||0.026||3.06||1.11–8.40|
|≥5 h||11 (18.3)||11 (40.7)||22 (25.3)|| || || || |
|Total||60||27||87|| || || || |
Of the 149 who died from suicide, 84 (56.4%) did so within 4 h of having self-harmed. Among the 46 deaths after self-poisoning, 16 (34.8%) are estimated to have died between 0 and 4 h following the attempt, 7 (15.2%) between 4 and 8 h, 9 (19.6%) between 8 and 12 h and 14 (30.4%) after more than 12 h. When comparing death from pesticide with the use of other poisonous agents, a relatively higher proportion (16 cases, 44.4%) died within 4 h of the attempt. The fact that 31 (67%) of the self-poisoning patients who eventually died were alive after 4 h provides an incentive to focus on improved case management and access to health services. However, the design of the study and the information obtained did not allow for a specific analysis of impact of treatment choices and access to treatment, as detailed case-specific information was not available on specific poisons ingested and treatment provided. The case fatality is not calculated as we do not believe that we identified all cases of attempts at self-harm.
The overall suicide rate found in this study with the highest burden in men and with the age-specific suicide rate highest among the 15–24 and the above-65 age groups are in accordance with previous studies in the same study area, indicating a continuing high burden of deaths from suicide in the community (Aaron et al. 2004; Bose et al. 2006; Prasad et al. 2006). The findings are also in accordance with previous studies from elsewhere in rural Tamil Nadu (Gajalakshmi & Peto 2007).
In this study information about non-fatal self-poisoning, attempts was collected for the first time in Kaniyambadi Block. Unfortunately, it was difficult to get the stories about suicide attempts verified at the community level with the result that the number of non-fatal episodes may have been significantly under estimated. One of the reasons for the low identification of attempts is the fact that suicides are criminalized in India, with attendant police formalities. Anecdotally, it is known that many patients who have self-harmed may present to healthcare facilities with somatization of symptoms, most commonly abdominal pain, but we did not attempt to validate this in our study among cases representing to health facilities.
We do not have an estimate of the case fatality in this series, on account of the probable underreporting of non-fatal self-poisoning cases. In Sri Lanka, even among self-poisoning cases reporting to health facilities after intake of highly toxic pesticides reach case-fatality ratios of up to 20–30% and often much below this (Eddleston et al. 2006a; van der Hoek & Konradsen 2006). A district-wide study in Sri Lanka found an overall case fatality rate of 7.4% after self-poisoning (Eddleston et al. 2006b).
The finding that approximately 80% of all self-poisoning cases had obtained the poisonous agent in the home or just outside the home highlights the importance of safe storage of poisonous substances in the homes. The different types of pesticides that are sold in Kaniyambadi Block are mainly pesticides in class Ia, Ib or II (WHO classification). The pesticides in class Ia, Ib and II are very toxic and can cause serious damage to the person ingesting these types of pesticides. They are considered ‘restricted use pesticides, in many countries. One of the ways that death caused by consumption of pesticides could be reduced is to limit the toxicity of the pesticides that are available for sale in the market. Such a measure seeks to reduce the lethality of the attempt, and not attempt to cause a reduction in the incidence of self-harm. The lower toxicity increase the chances of the person being found alive, and being taken for treatment, with resultant greater chances of survival.
This study was not designed to obtain qualitative or systematic information to explain gender differences in treatment seeking behaviour or the reasons for choices of specific poisonous agents for self-harm. However, the finding from this that more females than males die at home needs further investigations to assess if certain logistical, cultural or health system related aspects are preventing women from seeking assistance.
Eddleston et al. (2006a, 2006b) have developed a Haddon matrix to identify factors that increase the risk of fatal rather than non-fatal pesticide self-poisoning using Sri Lanka as a case. In this paper, it is argued that many important host factors of influencing the rate of fatal case such as age, gender and genetics cannot be changed, whereas other factors that potentially can be changed, such as alcohol use and mental health, have proved difficult to change. High suicide rates in developing countries are a measure of the lack of choice in these societies, and it is argued that a reduction in suicide rates will occur with population-based approaches that focus on improving the general health of populations (e.g. macroeconomic policies that aim for social justice), schemes to meet basic human needs, organizing local support groups within vulnerable sections of society, developing and implementing an essential pesticide list) rather than focusing on methods at an individual’s level (Jacob 2008).
The factors that we wish to highlight in this article are that, prevention at the current time can focus on restriction on the types of poisons that are available and promoting access to better health care. By presenting information on the type of compounds currently being used, and the showing that treatment enhances survival, we strengthen the argument for reducing the lethality of pesticides and for facilitating access to health care.
The authors gratefully acknowledge the help and support of the entire team of faculty and staff of the Department of Community Health, without whom this study could not have been carried out. The authors particularly thank the Health Aides, and Mr N. Babu and Mr Pandiarajan, the senior field workers who independently verified the reports of suicides and attempts.
WHO classification is used: Ia, extremely hazardous; Ib, highly hazardous; II, moderate hazardous; III, slightly hazardous; U, unlikely to be hazardous (WHO 2004).