Suicide in Sri Lanka 1975–2012: age, period and cohort analysis of police and hospital data
© Knipe et al.; licensee BioMed Central Ltd. 2014
Received: 25 February 2014
Accepted: 5 August 2014
Published: 13 August 2014
Sri Lanka has experienced major changes in its suicide rates since the 1970s, and in 1995 it had one of the highest rates in the world. Subsequent reductions in Sri Lanka’s suicide rates have been attributed to the introduction of restrictions on the availability of highly toxic pesticides. We investigate these changes in suicide rates in relation to age, gender, method specific trends and birth-cohort and period effects, with the aim of informing preventative strategies.
Secular trends of suicide in relation to age, sex, method, birth-cohort and period effects were investigated graphically using police data (1975–2012). Poisoning case-fatality was investigated using national hospital admission data (2004–2010).
There were marked changes to the age-, gender- and method-specific incidence of suicide over the study period. Year on year declines in rates began in 17–25 year olds in the early 1980s. Reduction in older age groups followed and falls in all age groups occurred after all class I (the most toxic) pesticides were banned. Distinct changes in the age/gender pattern of suicide are observed: in the 1980s suicide rates were highest in 21–35 year old men; by the 2000s, this pattern had reversed with a stepwise increase in male rates with increasing age. Throughout the study period female rates were highest in 17–25 year olds. There has been a rise in suicide by hanging, though this rise is relatively small in relation to the marked decline in self-poisoning deaths. The patterns of suicides are more consistent with a period rather than birth-cohort effect.
The epidemiology of suicide in Sri Lanka has changed noticeably in the last 30 years. The introduction of pesticide regulations in Sri Lanka coincides with a reduction in suicide rates, with evidence of limited method substitution.
KeywordsSri Lanka Pesticide poisoning Suicide Period effects Birth-cohort effects
The epidemiology of suicide in low and middle income countries (LAMIC) in Asia differs from the west. In most western countries the incidence of suicide increases with age; in rural Asia there is also an earlier peak in incidence in 15–30 year olds [1–4]. A suggested reason for the high incidence in young people is that self-harm with low suicide intent is highest in 15–34 year olds worldwide and is often impulsive. In countries such as the UK, the methods that impulsive self-harmers have readily available, such as paracetamol, have low case-fatality (often less than 1%) . In contrast, in rural Asia with the ready availability of highly toxic pesticides, some with case fatalities over 20% , impulsive acts more often result in completed suicide.
If the high rate of suicide in young people in rural Asia is a consequence of high lethality methods being used in impulsive acts of self-harm, reducing the toxicity of the pesticides available should reduce the case fatality associated with intentional poisoning and alter the epidemiology (age/sex patterns) of suicide. Consequently there should be a lowering of suicide incidence in young people, as well as a reduction in the overall suicide rate so long as equally lethal alternative methods are not adopted (method substitution).
Sri Lanka was notorious for having one of the highest suicide rates in the world in the early 1990s . In subsequent years its suicide rate declined dramatically, a fall in rates which coincided with a series of bans on WHO class I pesticides starting in 1984 [8, 9]. In a previous analysis of suicides in Sri Lanka (1975–2005), we found evidence that reductions in suicide in both men and women coincided with the ban of the most toxic pesticides . This decline in suicide rates did not appear to be associated with secular trends in other risk factors for suicide such as unemployment, alcohol misuse, divorce or Sri Lanka’s civil war (1983–2009). Previous investigations of the decline did not assess age, birth-cohort or period specific effects on suicide trends; such an analysis would help clarify the impact of regulations on particular age groups and possible long term effects on a generation of individuals exposed during their early life to a high incidence of suicide in their community. Furthermore, since our earlier analysis, a series of company trials and further regulations, starting in 2004 and culminating in 2008, have occurred in Sri Lanka, these resulted in reductions in the toxicity of highly toxic paraquat formulations. These interventions were followed by a complete ban on paraquat, dimethoate, and fenthion, from 2011 [9, 10].
Using police and hospital data we investigated: i) changes to age/gender and method specific pattern of suicide in Sri Lanka between 1975 and 2012; ii) changes in case fatality following pesticide regulations; and c) investigated the consequence of pesticide availability/toxicity on period and birth-cohort effects.
Sri Lanka is an island state with a population of 20.3 million (2011 census) situated off the south east coast of India. 77% of Sri Lanka’s population lives in rural areas (2011 census) and nearly a third of those employed work in the agricultural sector .
Age, gender and method specific suicide data were obtained, where available, for 1975–2012 from the Department of Police, Division of Statistics, Sri Lanka (http://www.police.lk/index.php/crime-trends). These publicly available data cover the majority of the island, but due to the civil war (1983–2009), data from three districts, Mullaitivu, Kilinochchi and Mannar, are missing from 1998–2004. We do not have the data broken down by district for the other years but assume they exclude these areas, because the years we have district level data confirm their exclusion during the civil war and we have no reason to suspect they would have been included for any other reason in the other years (post 1983). Based on the recent census (2011) these districts account for 1.5% of the total population and therefore their omission is unlikely to distort observations of the overall trends.
Categories of suicide deaths coded during different periods
Number of suicide methods/categories
Jumping in front of train
Jumping in front of train
Jumping in front of train
With sharp cutting instrument
With sharp cutting instrument
With sharp cutting instrument
Acetic Acid Poisoning
Ingestion of acids
Medicinal drug overdose
Jumping from a height
Ingestion/injection of addictive drugs
Publicly available hospital admission and hospital death data for poisoning (ICD10 codes: T36-T65), by age and gender, were obtained from the Ministry of Health for Sri Lanka. Age and gender data was only available for 2004–2010. Data were available for the following age categories: 5–16, 17–49, 50–69 and 70 +.
Censuses have been carried out in 1971, 1981, 2001 and 2011 in Sri Lanka. Whilst the final 2011 census results had not been released at the time of this analysis, data from a 5% sample of the census were available and we used this for relevant rate estimates. Mid-year population estimates for 1975–2010 were obtained from the Registrar General for Sri Lanka (publicly available). With regard to the estimates, there were three abrupt changes to the population estimates for different age groups centred on the three census years (1981, 2001 and 2011) (see Additional file 1), suggesting the mid-year population figures had been imperfectly estimated for the different age groups. To adjust for this, we recalculated the mid-year populations between the census years (1971–2011), for further details see Additional file 1. We used these estimates to calculate age and gender specific suicide rates (5 year age bands). The population estimate for 2012 has not been released; therefore the 2011 population was used as the denominator. Three year moving averages were used to smooth the estimates of trends in rates for graphical presentation, with averages centred on the middle year, e.g. the average rate for 1999–2001 would be plotted against 2000.
Graphical presentations were used to investigate secular trends of suicide in relation to age, sex, method, period and birth cohort. Age groups were collapsed into four broad age categories for ease of interpretation (17–25, 26–35, 36–55 and 55+). The hospital data was used to investigate trends in poisoning admissions, and case-fatality was calculated using information on live discharges and in-hospital deaths. Case fatality is reported as number of deaths per 100 admissions.
In order to investigate birth-cohort effects, we identified the birth years of the population represented in the suicide data available (1975–2012). We then tracked forward the age/gender specific suicide rates for each cohort. For example to calculate the 21–25 age-specific rate for the 1960 birth cohort, the mean suicide rate of the 21–25 age group over the years of 1981–1985 was used. Gender specific rates were plotted against 5-year age groups for each birth cohort.
We also investigated period effects by calculating the rates of suicide for each age group spanning five year periods starting from 1975. We did this by averaging the age/gender specific rate for each 5-year time period and plotted this against 5-year age groups for each calendar period.
In order to assess whether a period or cohort effect model was a better fit for our data we employed the Akaike information criterion (AIC) to help us assess the quality of the model for our data . A smaller AIC score reflects a better fit. For each gender, two Poisson regression models were fitted. The first model regressed the number of suicide events against age and cohort group; this model had 25 parameters. The second regressed the number of suicide events against age and period group; this model had 16 parameters. The resulting log likelihood function for each model was used to calculate the AIC score (AIC = 2*parameters-2*log likelihood). We acknowledge that age, period and cohort effects are confounded, and that what appears to be an effect of one of these factors on rates, can always be explained as being the result of some combination of the other two . However we are assuming that age and period (younger more impulsive; changing availability of highly toxic pesticides) are more parsimonious explanations of changing rates than cohort effects (coping strategies set by exposure to high rates at a sensitive age leads to higher risk of suicide later in life). We would only consider the changing rates as cohort effects if this model described the data in a more straightforward way. That is we would need an interaction between age and period to achieve the same fit to the data as a cohort effect model.
Age-gender specific trends
Method specific trends
Additional file 2 shows age, gender and method-specific suicide rates in two time periods (1982–84 and 2010–12). The age and gender distribution of poisoning suicides in 1982–84 closely resembles the patterning seen for overall suicides (Figure 2). In 2010–12 poisoning suicides still closely mirror the overall suicide pattern for the same period; however, in men the pattern by age of hanging related suicides has changed to strongly resemble poisoning suicides.
Hospital admissions and case fatality
Hospital admission data stratified by age and gender were only available from 2004 to 2010. Age groupings were too broad to observe any meaningful patterns and therefore are not presented. Poisoning admissions have increased; male rates rose from 328 per 100,000 in 2004 to 463 per 100,000 in 2010 and in females rates rose from 270 to 421 per 100,000. Pesticide poisoning was the most common method in men until 2007, when other poisoning admissions were the most common. In women prescription/medicinal drug poisoning was the most common method (see Additional file 3). Organophosphorus and carbamate insecticide poisoning accounted for 72-75% of all pesticide admissions. Pesticide related case fatality dropped from 9.7 per 100 admissions in 2004 to 4.6 in 2010 in men and from 4.1 to 2.0 in women. Case fatality for drugs and other poisoning remained fairly constant throughout the period (1.1 per 100 admissions in men and 0.4 in women).
By calculating AIC for each birth-cohort and period effect model in men and women, we found that in men the AIC was 1698 for a birth-cohort model vs. 1620 for a period model, suggesting a slightly better fit for the latter models (smaller AIC score reflects a better fit). In women the difference was more pronounced, with an AIC score of 957 for the birth-cohort model vs. 672 for a period effect model.
Sri Lanka has seen dramatic changes to the age, gender and method specific pattern of suicide over the last 30 years; there are five main features to these changes. First, there have been year on year declines in all age groups since 1995; these declines began in the 17–25 year olds (in the early 1980s), followed by the 26–35 and ≥36 age groups in the late 1980s and mid 1990s respectively. Second, there have been marked changes in the age pattern of suicide in males; in the 1980s the highest rate was in 21–25 year olds and rates declined with increasing age, whereas the opposite pattern was seen in the 2000s with rates increasing with age. Third, the reductions in rates appear to be driven by a decline in self-poisoning suicides. There has been a rise in suicide by hanging, but this has been small in comparison with the fall in self-poisoning suicide. Fourth, there was a rise in drug related poisoning admissions, especially in women, and some evidence of reductions in case fatality associated with all poisoning admissions. Last, there was evidence of possible birth-cohort and period effects, though the analysis suggests a stronger period effect.
Strengths and limitations
A particular strength of this study is the wealth of detailed age and method specific national suicide mortality data spanning over 35 years; this is unusual for a LAMIC. We were able to identify groups with high suicide incidence and emerging patterns of method specific suicide. We are the first to report on age and gender trends of suicide in Sri Lanka having adjusted for the imperfect estimations of different age and gender population figures (see Additional file 1). This adjustment explains why the trends reported in this study differ to those previously reported [1, 3]. To our knowledge, this is also the first study to investigate birth-cohort and period effects in a LAMIC in Asia.
This evidence, however, should be considered in light of several limitations. One limitation is the incomplete coverage of suicide statistics from the Northern part of the country which was heavily affected by the civil war. Only 1.5% of the population, however, live in this area, so the overall findings of this study are unlikely to be greatly affected by their exclusion. A common problem of suicide statistics originating from LAMIC are their reliability in capturing true incidence. Systematic investigation of the quality and reliability of Sri Lanka’s suicide data has not been carried out. Anecdotally, from our fieldwork studies, we are aware that data quality and completeness may vary between police departments, but we have no evidence that data quality has changed over time in such a way as to bias our estimates of temporal trends.
The decriminalisation of suicide in 1998 in Sri Lanka coupled with changing attitudes towards suicide may result in more deaths being attributed to suicide post 1998, so we may have underestimated the extent of the recent reductions in suicide. Equally, it is possible that the suicide rate in women, especially young women, is underestimated. The act of suicide may jeopardise the family’s honour and social standing within a community  and therefore families may hide details regarding the incident or persuade doctors and coroners not to document the suicide and report it as an accident. This is likely to have affected the rate of suicide throughout the study period and is therefore unlikely to have influenced secular trends observed. In addition we were unable to investigate the emergence of method specific suicides by urban and rural areas separately. This information would have been useful in identifying any differential impact of pesticide regulations in the areas of the country where they are most often used and so most easily accessible. We have not explicitly investigated the impact of changes in Sri Lanka’s political and economic environment on suicide trends; our previous analysis indicated that these were unlikely to have had a major influence on recent trends.
Data on hospital admissions for poisoning are collected from all peripheral (community centres) and tertiary (teaching and special) hospitals in Sri Lanka. This data is collected by the Ministry of Health using a standard returns form from all hospitals. Due to limited facilities in peripheral hospitals, poisoning patients are often transferred to tertiary hospitals (47.2% of all poisoning admissions) . With no linking of patient data for transfers, a degree of double counting of admissions occurs; which means we are likely to underestimate case fatality estimates, as some cases of self-poisoning as the denominator for our case fatality will be over-estimated as some individuals are included twice . This will not have affected our findings in relation to case fatality trends, as no national changes have been made to hospital data reporting during the study period. Lastly, age- and sex-specific hospital admission data was only available for a relatively short time period, meaning we could not evaluate long term trends.
Findings in context of literature
Asian suicide rates
In contrast to Sri Lanka’s declining suicide rate over the last 18 years, its neighbour India had fairly stable rates [2, 17], whilst in China rates have only declined markedly over the last 10 years [18, 19]. Both China and India can be considered to be similar to Sri Lanka as they both have a large proportion of their population engaged in agriculture. In other Asian nations such as Korea, Taiwan and Japan, pesticide poisoning accounts for a smaller proportion of total suicides (26% , 12%  and <4%  respectively). Furthermore, these countries are more economically advanced and have fewer agricultural worker (and therefore less easy access to, and familiarity with, pesticides) than is the case in Sri Lanka. Differences in the predominant religions and political situation may further compromise cross-country comparison in Asia.
As in Sri Lanka, the age and gender profile of suicide in China has changed [18, 22–24] in recent years. In the early 1980s in Sri Lanka, there were two peaks in the age-specific incidence of suicide in males, one in young people and the other in the elderly, similar to the pattern seen in China in the late 80s . In more recent years, there is a linear increase in male rates in Sri Lanka and China with age - a pattern typical of that seen in many western countries [2, 18]. This linear pattern is also now seen in female rates in China . Whilst female suicides in Sri Lanka have fallen over time, the age pattern – linear declines in rates with age are in marked contrast to those seen in the West (and China) where incidence is higher amongst older women. These changes to the age and gender profile in China and Sri Lanka, however, happened at different times. In China the changes started in the mid-late 90s, whereas in Sri Lanka these changes started in the early 80s. National data from India for the same time period are not available; therefore we are unable to make this comparison. Self-poisoning remains the commonest method in Sri Lanka, India and China, followed by hanging and drowning [2, 25].
Restriction of one method of suicide may lead to rises in the use of other methods for suicide (substitution). A review of suicide prevention through means restriction concluded that restriction leads to limited method substitution . The extent to which this occurs, and its impact on overall suicide rates will depend on the lethality of the method restricted, the lethality of the individual’s alternative choice of method and the individual’s suicidal intent [27, 28]. For example in England and Wales the detoxification of domestic gas supplies saw a reduction in domestic gas poisoning suicides and a rise in drug overdose deaths . This substitution effect, however, was not observed in all age and gender groups and was most apparent in women and younger men. This is consistent with our analysis of Sri Lankan data where restriction of pesticides appeared to result in a differential increase in hanging suicides by age and gender. Our data suggests that a degree of method substitution has occurred but because this rise in hanging was small in relation to the fall in self-poisoning suicides, the net effect was a marked reduction in overall suicide rates.
Birth-cohort and period effects
Investigations into birth-cohort and period effects have mainly been conducted in non-Asian countries [30–40], with the exception of studies conducted in Japan  and South Korea . The majority of these studies present evidence of both period and birth-cohort effects [35–41], which is consistent with the findings from this study. The Sri Lankan evidence, however, points to a slightly stronger period effect which appears to relate to the availability of the most toxic pesticides in the market place. This is similar to the findings of three other studies conducted in the United Kingdom , Australia  and New Zealand . These studies suggest a possible period effect as a result of changes in the availability of toxic domestic gas in the United Kingdom and barbiturates in New Zealand and Australia. In addition, whilst there will be combinations of age and cohort effects which will be equivalent this period effect is a more parsimonious explanation of the observed data.
Interpretation and implications
Our data suggests that the introduction of bans on all WHO Class I pesticides have contributed to the reduction of overall suicide rates, age specific rates and poisoning-related case fatality. Reductions to case fatality may also be a result of better management of poisoning cases in hospitals . These reasons, however, are unlikely to account for all the reduction in incidence. The time trends analysis suggests that there may be other important contributors. In 1970 and 1976, Endrin, toxaphene, chlordimeform, thallium and DDT were banned. These pesticide regulations had little effect on the overall and age-gender specific suicide rates. The limited effect may be because these pesticides were not commonly used/used for suicide; or that these pesticides were quickly replaced by other toxic formulations especially given the increase of pesticides available in the country post-open economy (1977).
In 17–25 men and 26–35 women, some plateauing of suicide rates occurred prior to the introduction of the bans on WHO class I pesticides. Also prior to pesticide regulations there were reductions in the incidence of suicide in 17–25 women. In 1980, when the plateauing and reductions started in these age groups, the ratification of the Control of Pesticides Act occurred, as did the establishment of both local and international knowledge and influence networks . Whilst the Office of the Registrar of Pesticides was not established until 1983 and the WHO class I pesticide bans were not enforced until 1984, the establishment of local and international networks is likely to have increased awareness of the problem of pesticide related suicides. This may have resulted in unofficial changes to the control, sale and storage of pesticides.
Sri Lanka also experienced a long civil war which started in 1983 and ended in 2009. It is possible that the war impacted on the suicide rate, but this is unlikely, especially as the suicide rate increases pre-1983 and starts its decline 14 years before the end of the civil war.
In women there may also be an alternative explanation to the plateauing and reductions. Changes to the country’s economic situation in the late 1970s to early 90s saw a growing number of young women migrating to foreign countries and urban centres for employment . Migration of these young women to foreign countries would have resulted in the removal of some at-risk individuals from the suicide statistics reported by the police, but would have been included in the census population figures. Migration of young women to urban centres where pesticide availability is less may also explain the earlier decline in the suicide rate in this group. Migration also allows these women to gain autonomy and control over their finances, which may in turn reduce the risk of self-harm.
Older men had the highest sex-specific rates of suicide in the post-regulation period. This pattern is in keeping with that seen in many western countries and may reflect increased levels of psychiatric morbidity and physical ill-health with age. It may also be that the years of poverty/stress accumulate as these men get older and increase suicide risk. It is also possible that as ageing is often associated with losses in income and occupation, especially in South Asian countries , older men experience a loss of status; this loss of status may then in turn increase their risk of suicide .
Pesticide regulations do not appear to have changed the pattern of suicide in women to more closely resemble western counterparts; a higher rate in young compared to older women prevails, whereas in the west older women have the higher rate. This may be viewed as contradicting the hypothesis that the high rate of suicide in young women is due to impulsive acts of self-harm being transformed into completed suicide due to the toxic pesticides employed. It would be expected therefore that the removal of these pesticides should see a reversal in the age profile to resemble a western profile. A possible reason to why we don’t see this reversal is that the higher suicide rate in these young women may reflect the high toxicity of the remaining pesticides. Young women, however, also have higher rates of suicide by other methods (e.g. hanging), which suggests that considerations to other sociocultural factors may also be important in this group. For example subsets of these young women are likely to be newly married and typically are newly residing with their husband’s family, sometimes at great distance from their own family. This environment gives rise to strained family relationships; in addition with economic and social changes, more young women are entering the work force, both locally and internationally. This group of women are threatening the conventional gender and generational hierarchy and this often leads to family arguments due to shifts in power [46, 47]. Attempts to address these issues may help to reduce the elevated rate seen in this group.
Sri Lanka’s secular trends suggest the existence of a period effect, especially in women, based on the availability of highly toxic pesticide. This is more parsimonious than a cohort effect, based on learning a style of coping due to being exposed to a high suicide rate at a sensitive age.
The introduction of pesticide bans in Sri Lanka appears to have contributed to the dramatic reduction in suicide rates in young men and older women, albeit with evidence of limited method substitution. Other countries where pesticide related suicide are a major problem should follow Sri Lanka’s example by restricting access to the most toxic pesticides. Unlike in many other countries, young women in Sri Lanka have the highest rates of suicide amongst women suggesting that method restriction alone may not be adequate in reducing the higher rates in this group. Further work investigating specific risk factors in this group is needed to help inform future preventative strategies.
Low and middle income countries
Akaike information criterion.
We would like to thank the Registrar General for Sri Lanka, the Department of Police, Division of Statistics, Sri Lanka, and Ministry of Health for Sri Lanka for providing us with the data for this paper. In particular we would like to thank Mr S.A.S.Bandulasena from the Department of Census and Statistics and Ms. Pushpa Gunasekara from the Medical Statistics Unit in Sri Lanka.
- De Silva V, Hanwella R, Senanayake M: Age and sex specific suicide rates in Sri Lanka from 1995–2011. Sri Lanka J Psychiatry. 2013, 3 (2): 7-11.Google Scholar
- Patel V, Ramasundarahettige C, Vijayakumar L, Thakur JS, Gajalakshmi V, Gururaj G, Suraweera W, Jha P, Million Death Study C: Suicide mortality in India: a nationally representative survey. Lancet. 2012, 379 (9834): 2343-2351.View ArticlePubMedPubMed CentralGoogle Scholar
- Thalagala N: Suicide trends in Sri Lanka 1880–2006; social, demographic and geographical variations. J College of Community Physicians Sri Lanka. 2011, 14 (1): 24-32.View ArticleGoogle Scholar
- Gunnell D, Eddleston M: Suicide by intentional ingestion of pesticides: a continuing tragedy in developing countries. Int J Epidemiol. 2003, 32 (6): 902-909.View ArticlePubMedPubMed CentralGoogle Scholar
- Gunnell D, Hawton K, Murray V, Garnier R, Bismuth C, Fagg J, Simkin S: Use of paracetamol for suicide and non-fatal poisoning in the UK and France: are restrictions on availability justified?. J Epidemiol Community Health. 1997, 51 (2): 175-179.View ArticlePubMedPubMed CentralGoogle Scholar
- Eddleston M, Eyer P, Worek F, Mohamed F, Senarathna L, von Meyer L, Juszczak E, Hittarage A, Azhar S, Dissanayake W, Sheriff MH, Szinicz L, Dawson AH, Buckley NA: Differences between organophosphorus insecticides in human self-poisoning: a prospective cohort study. Lancet. 2005, 366 (9495): 1452-1459.View ArticlePubMedGoogle Scholar
- Ratnayeke L: Suicide and crisis intervention in rural communities in Sri Lanka. Crisis. 1996, 17 (4): 149-151. 154View ArticlePubMedGoogle Scholar
- Gunnell D, Fernando R, Hewagama M, Priyangika WD, Konradsen F, Eddleston M: The impact of pesticide regulations on suicide in Sri Lanka. Int J Epidemiol. 2007, 36 (6): 1235-1242.View ArticlePubMedPubMed CentralGoogle Scholar
- Pearson M, Zwi AB, Buckley NA, Manuweera G, Fernando R, Dawson AH, McDuie-Ra D: Policymaking ‘under the radar’: a case study of pesticide regulation to prevent intentional poisoning in Sri Lanka. Health Policy Plan. 2013Google Scholar
- Dawson AH, Buckley NA: Toxicologists in public health–Following the path of Louis Roche (based on the Louis Roche lecture “An accidental toxicologist in public health”, Bordeaux, 2010). Clin Toxicol (Phila). 2011, 49 (2): 94-101.View ArticleGoogle Scholar
- Department of Census and Statistics: Sri Lanka Labour Force Survey. Edited by: Planning MoFa. 2012, Department of Census and Statistics, http://www.statistics.gov.lk/samplesurvey/Labour%20Force%20survey_Annual%20Report_2012-final.pdf,Google Scholar
- Akaike H: A new look at the statistical model identification. IEEE Trans Autom Control. 1974, 19 (6): 716-723.View ArticleGoogle Scholar
- Bell A, Jones K: The impossibility of separating age, period and cohort effects. Soc Sci Med. 2013, 93: 163-165.View ArticlePubMedGoogle Scholar
- Marecek J, Senadheera C: ‘I drank it to put an end to me’: Narrating girls’ suicide and self-harm in Sri Lanka. Contr Indian Sociol. 2012, 46 (1–2): 53-82.View ArticleGoogle Scholar
- Eddleston M, Sudarshan K, Senthilkumaran M, Reginald K, Karalliedde L, Senarathna L, de Silva D, Rezvi Sheriff MH, Buckley NA, Gunnell D: Patterns of hospital transfer for self-poisoned patients in rural Sri Lanka: implications for estimating the incidence of self-poisoning in the developing world. Bull World Health Organ. 2006, 84 (4): 276-282.View ArticlePubMedPubMed CentralGoogle Scholar
- Roberts DM, Karunarathna A, Buckley NA, Manuweera G, Sheriff MH, Eddleston M: Influence of pesticide regulation on acute poisoning deaths in Sri Lanka. Bull World Health Organ. 2003, 81 (11): 789-798.PubMedGoogle Scholar
- Suicide rates (per 100,000), by gender, India, 1980–2009. [http://www.who.int/mental_health/media/indi.pdf]
- Wang CW, Chan CL, Yip PS: Suicide rates in China from 2002 to 2011: an update. Soc Psychiatry Psychiatr Epidemiol. 2013, 49 (6): 929-41.View ArticlePubMedGoogle Scholar
- Suicide rates (per 100,000), by gender, China (mainland, selected rural areas), 1987–1999. [http://www.who.int/mental_health/media/chinrural.pdf]
- Kim SY, Kim MH, Kawachi I, Cho Y: Comparative epidemiology of suicide in South Korea and Japan: effects of age, gender and suicide methods. Crisis. 2011, 32 (1): 5-14.View ArticlePubMedGoogle Scholar
- Chang SS, Lu TH, Eddleston M, Konradsen F, Sterne JA, Lin JJ, Gunnell D: Factors associated with the decline in suicide by pesticide poisoning in Taiwan: a time trend analysis, 1987–2010. Clin Toxicol (Phila). 2012, 50 (6): 471-480.View ArticleGoogle Scholar
- Yip PS, Liu KY, Hu J, Song XM: Suicide rates in China during a decade of rapid social changes. Soc Psychiatry Psychiatr Epidemiol. 2005, 40 (10): 792-798.View ArticlePubMedGoogle Scholar
- Phillips MR, Li X, Zhang Y: Suicide rates in China, 1995–99. Lancet. 2002, 359 (9309): 835-840.View ArticlePubMedGoogle Scholar
- Ji J, Kleinman A, Becker AE: Suicide in contemporary China: a review of China’s distinctive suicide demographics in their sociocultural context. Harv Rev Psychiatry. 2001, 9 (1): 1-12.View ArticlePubMedGoogle Scholar
- Phillips MR, Yang G, Zhang Y, Wang L, Ji H, Zhou M: Risk factors for suicide in China: a national case–control psychological autopsy study. Lancet. 2002, 360 (9347): 1728-1736.View ArticlePubMedGoogle Scholar
- Daigle MS: Suicide prevention through means restriction: assessing the risk of substitution: a critical review and synthesis. Accid Anal Prev. 2005, 37 (4): 625-632.View ArticlePubMedGoogle Scholar
- Gunnell D, Miller M: Strategies to prevent suicide. BMJ. 2010, 341: c3054-View ArticlePubMedGoogle Scholar
- Yip PSF, Caine E, Yousuf S, Chang S-S, Wu KC-C, Chen Y-Y: Means restriction for suicide prevention. Lancet. 2012, 379 (9834): 2393-2399.View ArticlePubMedGoogle Scholar
- Gunnell D, Middleton N, Frankel S: Method availability and the prevention of suicide–a re-analysis of secular trends in England and Wales 1950–1975. Soc Psychiatry Psychiatr Epidemiol. 2000, 35 (10): 437-443.View ArticlePubMedGoogle Scholar
- Page A, Milner A, Morrell S, Taylor R: The role of under-employment and unemployment in recent birth cohort effects in Australian suicide. Soc Sci Med. 2013, 93: 155-162.View ArticlePubMedGoogle Scholar
- Etzersdorfer E, Piribauer F, Sonneck G: Sex differential for suicide among Austrian age cohorts. Acta Psychiatr Scand. 1996, 93 (4): 240-245.View ArticlePubMedGoogle Scholar
- Ajdacic-Gross V, Bopp M, Gostynski M, Lauber C, Gutzwiller F, Rossler W: Age-period-cohort analysis of Swiss suicide data, 1881–2000. Eur Arch Psychiatry Clin Neurosci. 2006, 256 (4): 207-214.View ArticlePubMedGoogle Scholar
- Snowdon J, Hunt GE: Age, period and cohort effects on suicide rates in Australia, 1919–1999. Acta Psychiatr Scand. 2002, 105 (4): 265-270.View ArticlePubMedGoogle Scholar
- Lynskey M, Degenhardt L, Hall W: Cohort trends in youth suicide in Australia 1964–1997. Aust N Z J Psychiatry. 2000, 34 (3): 408-412.PubMedGoogle Scholar
- Kalediene R: Time trends in suicide mortality in Lithuania. Acta Psychiatr Scand. 1999, 99 (6): 419-422.View ArticlePubMedGoogle Scholar
- Granizo JJ, Guallar E, Rodriguez-Artalejo F: Age-period-cohort analysis of suicide mortality rates in Spain, 1959–1991. Int J Epidemiol. 1996, 25 (4): 814-820.View ArticlePubMedGoogle Scholar
- Rodrigues NC, Werneck GL: Age-period-cohort analysis of suicide rates in Rio de Janeiro, Brazil, 1979–1998. Soc Psychiatry Psychiatr Epidemiol. 2005, 40 (3): 192-196.View ArticlePubMedGoogle Scholar
- Phillips JA, Robin AV, Nugent CN, Idler EL: Understanding recent changes in suicide rates among the middle-aged: period or cohort effects?. Public Health Rep. 2010, 125 (5): 680-688.PubMedPubMed CentralGoogle Scholar
- Skegg K, Cox B: Suicide in New Zealand 1957–1986: the influence of age, period and birth-cohort. Aust N Z J Psychiatry. 1991, 25 (2): 181-190.PubMedGoogle Scholar
- Gunnell D, Middleton N, Whitley E, Dorling D, Frankel S: Influence of cohort effects on patterns of suicide in England and Wales, 1950–1999. Br J Psychiatry. 2003, 182: 164-170.View ArticlePubMedGoogle Scholar
- Odagiri Y, Uchida H, Nakano M: Gender differences in age, period, and birth-cohort effects on the suicide mortality rate in Japan, 1985–2006. Asia Pac J Public Health. 2011, 23 (4): 581-587.View ArticlePubMedGoogle Scholar
- Kwon JW, Chun H, Cho SI: A closer look at the increase in suicide rates in South Korea from 1986–2005. BMC Public Health. 2009, 9: 72-View ArticlePubMedPubMed CentralGoogle Scholar
- Eddleston M: Patterns and problems of deliberate self-poisoning in the developing world. QJM. 2000, 93 (11): 715-731.View ArticlePubMedGoogle Scholar
- De Alwis M: The changing role of women in Sri Lankan Society. Soc Research. 2002, 69 (3): 675-691.Google Scholar
- De Silva WI: Family transition in South Asia: provision of social services and social protection. Asia Pac Popul J. 2005, 20 (2): 13-Google Scholar
- Widger T: Suffering, frustration, and anger: class, gender and history in Sri Lankan suicide stories. Cult Med Psychiatry. 2012, 36 (2): 225-244.View ArticlePubMedGoogle Scholar
- Marecek J: Young women’s suicide in Sri Lanka: cultural, ecological, and psychological factors. Asia J Counsell. 2006, 13 (1): 63-92.Google Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2458/14/839/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.