India has long been thought to have more snakebites than any other country. However, inadequate hospital-based reporting has resulted in estimates of total annual snakebite mortality ranging widely from about 1,300 to 50,000. We calculated direct estimates of snakebite mortality from a national mortality survey.
Methods and Findings
We conducted a nationally representative study of 123,000 deaths from 6,671 randomly selected areas in 2001–03. Full-time, non-medical field workers interviewed living respondents about all deaths. The underlying causes were independently coded by two of 130 trained physicians. Discrepancies were resolved by anonymous reconciliation or, failing that, by adjudication.
A total of 562 deaths (0.47% of total deaths) were assigned to snakebites. Snakebite deaths occurred mostly in rural areas (97%), were more common in males (59%) than females (41%), and peaked at ages 15–29 years (25%) and during the monsoon months of June to September. This proportion represents about 45,900 annual snakebite deaths nationally (99% CI 40,900 to 50,900) or an annual age-standardised rate of 4.1/100,000 (99% CI 3.6–4.5), with higher rates in rural areas (5.4/100,000; 99% CI 4.8–6.0), and with the highest state rate in Andhra Pradesh (6.2). Annual snakebite deaths were greatest in the states of Uttar Pradesh (8,700), Andhra Pradesh (5,200), and Bihar (4,500).
Snakebite remains an underestimated cause of accidental death in modern India. Because a large proportion of global totals of snakebites arise from India, global snakebite totals might also be underestimated. Community education, appropriate training of medical staff and better distribution of antivenom, especially to the 13 states with the highest prevalence, could reduce snakebite deaths in India.
Earlier hospital based reports estimate about 1,300 to 50,000 annual deaths from snakebites per year in India. Here, we present the first ever direct estimates from a national mortality survey of 1.1 million homes in 2001–03. Full-time, non-medical field workers interviewed living respondents about all deaths. The underlying causes were independently coded by two of 130 trained physicians. The study found 562 deaths (0.47% of total deaths) were assigned to snakebites, mostly in rural areas, and more commonly among males than females and peaking at ages 15–29. Snakebites also occurred more often during the rainy monsoon season. This proportion represents about 45,900 annual snakebite deaths nationally (99% CI 40,900 to 50,900) or an annual age-standardised rate of 4.1/100,000 (99% CI 3.6–4.5), with higher rates in rural areas (5.4) and with the highest rate in the state of Andhra Pradesh (6.2). Annual snakebite deaths were greatest in the states of Uttar Pradesh (8,700), Andhra Pradesh (5,200), and Bihar (4,500). Thus, snakebite remains an underestimated cause of accidental death in modern India, causing about one death for every two HIV-related deaths. Because a large proportion of global totals of snakebites arise from India, global snakebite totals might also be underestimated. Effective interventions involving education and antivenom provision would reduce snakebite deaths in India.
Citation: Mohapatra B, Warrell DA, Suraweera W, Bhatia P, Dhingra N, Jotkar RM, et al. (2011) Snakebite Mortality in India: A Nationally Representative Mortality Survey. PLoS Negl Trop Dis 5(4): e1018. doi:10.1371/journal.pntd.0001018
Editor: John Owusu Gyapong, Ghana Health Service, Ghana
Received: November 25, 2010; Accepted: February 15, 2011; Published: April 12, 2011
Copyright: © 2011 Mohapatra et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The Registrar General of India and the Sample Registration System are funded by the Government of India. The Million Death Study is funded by the Fogarty International Centre of the US National Institutes of Health (http://www.nih.gov/; grant R01 TW05991-01) and the Li Ka Shing Knowledge Institute and the Keenan Research Centre at St. Michael's Hospital, University of Toronto (http://www.stmichaelshospital.com/knowledgeinstitute/). PJ is supported by the Canada Research Chair program. The senior author had full access to all the data and had final responsibility for the decision to submit for publication on behalf of the authors. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Alexander the Great invaded India in 326 BC, and was greatly impressed by the skill of Indian physicians; especially in the treatment of snakebites . Since then, India has remained notorious for its venomous snakes and the effects of their bites. With its surrounding seas, India is inhabited by more than 60 species of venomous snakes – some of which are abundant and can cause severe envenoming . Spectacled cobra (Naja naja), common krait (Bungarus caeruleus), saw-scaled viper (Echis carinatus) and Russell's viper (Daboia russelii) have long been recognised as the most important, but other species may cause fatal snakebites in particular areas, such as the central Asian cobra (Naja oxiana) in the far north-west, monocellate cobra (N. kaouthia) in the north-east, greater black krait (B. niger) in the far north-east, Wall's and Sind kraits (B. walli and B. sindanus) in the east and west and hump-nosed pit-viper (Hypnale hypnale) in the south-west coast and Western Ghats .
Joseph Fayrer of the Indian Medical Service first quantified human snakebite deaths in 1869 for about half of “British India” (including modern Pakistan, Bangladesh and Burma), finding that 11,416 people had died of snakebites . Subsequent estimates of human deaths from snakebite prior to Indian Independence ranged from 7,400 to 20,000 per year –. Government of India hospitals from all but six states reported only 1,364 snakebite deaths in 2008  but this is widely believed to be an under-report as many victims of snakebite choose village-based traditional therapists and most die outside government hospitals. Community-based surveys in some localities have shown much higher annual mortality rates, ranging widely from 16.4 deaths/100,000 in West Bengal  to 161/100,000 in the neighbouring Nepal Terai . However, such focal data cannot be extrapolated to provide national or even state totals because of the heterogeneity of snakebite incidence. These uncertainties have resulted in indirect estimates of annual snakebite mortality in India that varied from approximately 1,300 to 50,000 , , –.
To fill this gap in knowledge, we estimated snakebite deaths directly from a large continuing study of mortality in India.
Ethics approval for the Million Deaths Study (MDS) was obtained from the Post Graduate Institute of Medical Research, St. John's Research Institute and St. Michael's Hospital, Toronto, Ontario, Canada –.
Most deaths in rural India take place at home without prior attention by any qualified healthcare worker, so most causes are not medically certified –. Other approaches are therefore needed to help determine the probable causes of such deaths. The Registrar General of India (RGI) organises the Sample Registration System (SRS), which monitors all births and deaths in a nationally representative selection of 1.1 million homes throughout all 28 states and seven union territories of India. India was divided into approximately one million areas for the 1991 census, each with about 1,000 inhabitants. In 1993, the RGI randomly selected 6,671 of these areas to be represented in the SRS. Household characteristics were recorded and then enumerated twice yearly thereafter, documenting new births and deaths, but not the causes of death .
Since 2002, one of 800 non-medical field staff (trained by the RGI in appropriate fieldwork methods) visited each SRS area every six months to record a written narrative (in the local language) for each death from families or other reliable informants. In addition to the narratives, answers to standard questions about the deaths were also recorded in the field report. Fieldwork quality control methods were used routinely, including random re-sampling by teams reporting directly to the study investigators , . This survey is part of the MDS, which seeks to assign causes to all deaths in SRS areas for the period between 2001–14 –, –.
These field reports, or ‘verbal autopsies’, were emailed randomly (based on the language of the narrative) to at least two of 130 collaborating physicians trained in disease coding. Physicians worked independently to assess the probable underlying cause of death, assigning each case a three-character International Classification of Diseases (ICD; 10th revision) code . Any differences between the two coders were resolved by anonymous reconciliation between them (asking each to reconsider) or, for persisting differences, adjudication by a third physician (3% or 15/562 of snakebite deaths, and 18% or 22,845/122,848 of all deaths). The physician coders' training and their written guidelines (available online ) instructed them to use their best medical judgement to determine the most probable cause of death. Field reports could not be collected on 12% of the identified deaths due to migration or change of residence. As these missing deaths were mostly random, a systematic misclassification in cause of death was unlikely. We used logistic regression to quantify the odds of snakebite versus other deaths for gender, state, religion, education, occupation, place of death and season. Risk is measured compared to the reference group of lowest risk for each variable. Climate data on rainfall and temperature were obtained for each state from the India Meteorological Department –.
The proportion of cause specific deaths in each age category was weighted by the inverse probability of household selection within rural and urban sub divisions of each state, to account for the sampling design . Using methods described earlier –, –, the resulting proportion of deaths from each cause was applied to the United Nations (UN) population division estimates of deaths in India in 2005  (9.8 million, upper and lower limits 9.4–10.3 million) to generate cause- specific death totals and rates. The UN totals (which undergo independent demographic review ) were used because the SRS underestimates adult mortality rates by about 10% –. The UN totals are not affected by the 12% of the SRS-enumerated deaths that were unavailable for interview in our survey. Totals for 2005 were used because they: (i) were most complete; (ii) could be compared to the available Indian Census projections for 2006; and (iii) captured information prior to the implementation of a new national health program in rural areas . However, applying the 2001–03 proportions to the 2005 total deaths did not introduce major biases since there was little change in the yearly distribution in snakebite deaths in our survey, or in the annual number of deaths reported from snakebites in routine national hospital surveillance data between 2003 and 2008 .
Snakebite deaths in study and nationally
Of the 643 deaths coded by physicians as ICD-10 codes X20–X29 (contact with venomous animals and plants), 523 (81%) were coded as X20 (venomous snakes) and review of these yielded no misclassified causes. Central re-examination of the symptoms and key words found 39 of 45 deaths coded as X27 (animals) and X29 (uncertain) to be snakebite deaths. We excluded 75 deaths coded as X21–X25 (various arthropods), X26 (marine organisms) and X28 (plants).
Among all 122,848 deaths, 2,179 of the deaths that were randomly chosen to be re-interviewed by independent teams were eventually matched to the identical houses and individuals of the MDS. Of the 2,179 re-sampled deaths, 9 were coded as snakebites, and 7 of these were found in the MDS. Thus, the sensitivity and specificity of the SRS field survey, assuming the re-sample deaths are the standard comparison, was 78% (7/9) and 100% (2,170/2,170), respectively.
A total of 562 of the 122,848 deaths (0.47% weighted by sampling probability or 0.46% unweighted) were from snakebites (Table 1). Almost all snakebite deaths (544 or 97%) were in rural areas. More men (330, 59%) than women (232, 41%) died from snakebites (overall ratio of 1.4 to 1). The proportion of all deaths from snakebites was highest at ages 5–14 years. Only 23% (127/562) of the deaths occurred in a hospital or other healthcare facility.
Expressed as national totals, snakebites caused 45,900 deaths in India in 2005 (99% CI 40,900 to 50,900). The age-standardised death rate per 100,000 population per year was 4.1 (99% CI 3.6–4.5) nationally and was 5.4 (99% CI 4.8–6.0) in rural areas.
Risk factors and seasonality
Figure 1 shows the odds ratios for snakebite deaths versus other deaths, adjusted for age, gender, and for high prevalence states (13 states with age-standardised snakebite death rates greater than 3 per 100,000) versus other states. The risks of snakebite deaths were significantly increased among Hindus and farmers/labourers, deaths occurring outside home, and during the monsoon months of June to September (Figures 1 and 2). In contrast, gender and education were not significantly associated with risk of snakebite death. About 5,000–7,000 snakebite deaths per month occurred during the monsoon period, compared to less than 2,000 deaths in the winter months. Monthly numbers of snakebite deaths correlated with rainfall (R = 0.93, p<.0001) and mean minimum temperature (R = 0.80, p = 0.0017), but not with mean maximum temperature (R = 0.35, p = 0.2585; Figure 2).
Odds ratio after adjusting for age, gender and states with a high prevalence of snakebite deaths (see definition in Table 2). Occupation ‘Other’ includes students and house wives.
Rainfall amount (mm) is cumulative daily rainfall for the past 24 hours measured by the India Meteorological Department , . Maximum and minimum temperatures are also measured daily and presented as monthly averages across the 13 snakebite high prevalence states. Pearson correlation coefficients between snakebite mortality and weather were: (i) rainfall; 0.93 (p<0.0001); (ii) minimum temperature: 0.80 (p = 0.0017); (iii) maximum temperature: 0.35 (p = 0.2585).
State mortality patterns
Annual age-standardised mortality rates per 100,000 from snakebite varied between states, from 3.0 (Maharashtra) to 6.2 (Andhra Pradesh) in the 13 states with highest prevalence (average 4.5) compared to 1.8 in the rest of the country (Table 2; Figure 3). Total deaths were highest in Uttar Pradesh (8,700), Andhra Pradesh (5,200), and Bihar (4,500). The age and gender of snakebite deaths also varied by region, although these differences were not significant due to the small numbers of snakebite deaths in each state. Deaths at ages 5–14 years were prominent in the states of Jharkhand and Orissa, whereas deaths at older ages were prominent in Andhra Pradesh, Bihar, Madhya Pradesh, and Uttar Pradesh (data not shown). In Bihar, Madhya Pradesh, Maharashtra and Uttar Pradesh, female deaths exceeded male deaths (Table 2).
Death rates are standardised to 2005 UN population estimates for India . The vertical bars represent the state wise estimated deaths (in thousands). Total snakebite deaths for the 13 states with high-prevalence of snakebite death are 42,800 or 93% of the national total (these states have about 85% of the total estimated population of India). States where the snakebite death rate was below 3/100,000 or where populations are less than 10 million are not shown. The states with high-prevalence of snakebite deaths are: AP-Andhra Pradesh, BR-Bihar, CG-Chhattisgarh, GJ-Gujarat, JH-Jharkhand, KA-Karnataka, MP Madhya Pradesh, MH-Maharashtra, OR-Orissa, RJ- Rajasthan, TN-Tamil Nadu, UP-Uttar Pradesh, WB-West Bengal.
Snakebite remains an important cause of accidental death in modern India, and its public health importance has been systematically underestimated. The estimated total of 45,900 (95% CI 40,900–50,900) national snakebite deaths in 2005 constitutes about 5% of all injury deaths and nearly 0.5% of all deaths in India. It is more than 30-fold higher than the number declared from official hospital returns . The underreporting of snake bite deaths has a number of possible causes. Most importantly, it is well known that many patients are treated and die outside health facilities – especially in rural areas. Thus rural diseases, be they acute fever deaths from malaria and other infections  or bites from snakes or mammals (rabies; ), are underestimated by routine hospital data. Moreover, even hospital deaths may be missed or not reported as official government returns vary in their reliability, as shown from a study of snakebites in Sri Lanka . The true burden of mortality from snakebite revealed by our study is similar in magnitude to that of some higher profile infectious diseases; for example, there is one snakebite death for every two AIDS deaths in India . Consequently, snakebite control programmes should be prioritised to a level commensurate with this burden.
Our data suggest underestimation in recent global estimates of mortality from snakebite deaths : the upper bounds of recent annual estimates were 94,000 deaths globally and 15,000 deaths in India. This total for India is only about one-third of the snake bite deaths detected in our study. The incidence of snakebite deaths per 100,000 population per year in a recent community-based study in Bangladesh was similar to ours , suggesting that much of South Asia might have thousands more snakebite deaths than is currently assumed. Considering the widely accepted gross underestimation of snakebite deaths in Africa , it seems highly probable that well over 100,000 people die of snakebite in the world each year.
A minimal number of non-fatal snakebites in India may be estimated with far less certainty. Indian data from routine public sector hospitals  are clearly under-reports of deaths (recording only 1 in 5 of the deaths we estimated to have occurred in hospital). Nonetheless, the ratio of non-fatal bites (about 140,000) to fatal bites (about 2,200) in these hospital data from 2003–08 (about 64∶1) is informative of the relative burden of bites to deaths. Very crudely, even if we halve the fatal/nonfatal bite ratio to 32, this would suggest at least 1.4 million non-fatal bites corresponding to the 45,000 fatal bites. The actual number of non-fatal bites in India may well be far higher, as the community-based study in Bangladesh found about 100 non-fatal bites for each death .
Our study has limitations; notably the misclassification of snakebite deaths. However, snakebites are dramatic, distinctive and memorable events for the victim's family and neighbours, making them more easily recognizable by verbal autopsy. We observed a reasonably high sensitivity and specificity when compared to re-sampled deaths. Confusion with arthropod bites and stings is unlikely because of the different circumstances, size and behaviour of the causative animal and the course of envenoming. For example, most deaths from hymenoptera stings result from rapidly evolving anaphylaxis. Kraits (important agents of snakebite death in South Asia) may unobtrusively envenom sleeping victims, who may die after developing severe abdominal pain, descending paralysis, respiratory failure and convulsions . Such deaths might not be associated with snakebite at all. These examples suggest possible underestimation of deaths in our data.
Since the numbers of deaths observed in each state were small, the estimated totals for each state are uncertain. However, the state distribution is broadly consistent with that reported by the RGI survey of deaths in selected rural areas in the 1990s . The marked geographic variation across states in our study is similar to that in a country-wide survey conducted during the period 1941–45, which identified Bengal, Bihar, Tamil Nadu, Uttar Pradesh, Madhya Pradesh, Maharashtra and Orissa as having the highest death rates from snakebite . Moreover, despite the obvious underestimates in hospitalised data , their geographical distribution of bites and deaths were similar to what we observed from household reports of deaths. The marked differences in snakebite mortality between states of India may be attributable to variations in human, snake and prey populations, and in local attitudes  and health services. The 13 states with the highest snakebite mortality are inhabited by the four most common deadly venomous snakes: Naja naja, Bungarus caeruleus, Echis carinatus and Daboia russelii. With the exception of E. carinatus, which favours open wasteland, these are widely distributed species of the plains and low hills where most Indians live. While some species can inhabit altitudes of up to 2,700 metres , this is exceptional and higher mountainous regions have considerably lower death rates.
As found in an earlier study , the peak age group of snakebite deaths is 15–29 years (25% or 142/562). However, the relative risk of dying from snakebite versus another cause was greater at ages 5–14 years. The peak age range and gender associated with snakebite mortality varied between states, perhaps reflecting differences in the relative numbers of children and women involved in agricultural work –. The slight excess among Hindus may reflect more tolerance of snakes and greater use of traditional treatments . Snakebites and snakebite fatalities peak during the monsoon season in India ,  and worldwide , probably reflecting agricultural activity, flooding, increased snake activity, and abundance of their natural prey.
Only 23% of the snakebite deaths identified in our survey occurred in hospital, consistent with an earlier study from five states . This emphasises three points: (i) hospital-based data reflect poorly the national burden of fatal snakebites; (ii) inadequacy of current treatment of snakebite in India; and (iii) vulnerability of snakebite victims outside hospital. Practicable solutions include strengthening surveillance to allow a more accurate perception of the magnitude of the problem, improving community education to reduce the incidence of snakebites and speed up the transfer of bitten patients to medical care, improving the training of medical staff at all levels of the health service (including implementation of the new WHO guidelines ), and deployment of appropriate antivenoms and other interventional tools where they are needed in rural health facilities to decrease case fatality –. In addition, phylogenetic and venom studies are needed to ensure appropriate design of antivenoms to cover the species responsible for serious envenoming.
The Registrar General of India has managed the SRS since they established the survey in 1971, and is collaborating with several of the authors in the ongoing Million Death Study. All study materials are available at www.cghr.org. The opinions expressed here are those of the authors and do not necessarily represent those of the Government of India. We thank Ansely Wong for comments and Brendon Pezzack for graphics.
The Million Death Study Collaborators: Indian Academic Partners (in alphabetical order)
1. Department of Community Medicine Gujarat Medical College, Ahmedabad: DV Bala, P Seth KN Trivedi
2. Department of Community Medicine Kolkatta Medical College, Kolkatta: SK Roy
3. Department of Community Medicine Regional Institute of Medical Sciences, Imphal: L Usharani
4. Department of Community Medicine S.C.B. Medical College, Cuttack, Orissa: Dr. B Mohapatra
5. Department of Community Medicine SMS Medical College, Jaipur: AK Bharadwaj, R Gupta
6. Epidemiological Research Center, Chennai: V Gajalakshmi, CV Kanimozhi
7. Gandhi Medical College, Bhopal: RP Dikshit, S Sorangi
8. Healis-Seskarhia Institute of Public Health, Mumbai: PC Gupta, MS Pednekar, S Sreevidya
9. Indian Institute of Health & Family Welfare, Hyderabad: P Bhatia
10. Institute of Health Systems Research, Hyderabad: P Mahapatra (until 2004)
11. St. John's Research Institute, St. John's Academy of Health Sciences, Bangalore: R Jotkar, S Rao-Seshadri
12. King George Medical College, Lucknow: S Awasthi
13. Regional Medical Research Center ICMR Institute, Bhubaneshwar: AS Karketta, SK Dar
14. School of Preventative Oncology, Patna: DN Sinha
15. School of Public Health Post Graduate Institute of Medical Education and Research, Chandigarh: N Kaur, R Kumar, JS Thakur
16. Tata Memorial Hospital, Mumbai: RA Badwe, RP Dikshit, K Mohandas
The Million Death Study Collaborators: Lead Partners
1. Office of the Registrar-General India, New Delhi: C Chandramouli (Registrar General of India [RGI]), RC Sethi, B Mishra, AK Saxena, MS Thapa, N Kumar, JK Banthia and DK Sikri (former RGIs)
2. Million Death Study Coordinating Centre for Global Health Research (CGHR), Li Ka Shing Knowledge Institute, St. Michael's Hospita,l Dalla Lana School of Public Health, University of Toronto Canada: DG Bassani, P Jha (Principal Investigator), R Jotkar, R Kamadod, B Pezzack, S Rao-Seshadri, P Rodriguez, J Sudhir, C Ramasundarahettige, W Suraweera
Conceived and designed the experiments: PJ DAW BM. Performed the experiments: PJ DAW. Analyzed the data: PJ DAW WS PSR. Wrote the paper: PJ DAW BM. Analysis planning: PB BM WS DAW PJ. Statistical analysis: WS PJ. Data interpretation, critical revision of the paper, and approval of the final paper: BM DAW WS PB ND RMJ PSR KM RW PJ.
- 1. Jaggi OP (2000) Medicine in India: Modern Period (History of Science, Philosophy and Culture in Indian Civilization, Vol. IX: Part 1). New Delhi: Oxford University Press. 384 p.
- 2. Whitaker R, Captain A (2004) Snakes of India: The Field Guide. Chennai: Draco Books. 495 p.
- 3. Fayrer J (1882) Destruction of life in India by poisonous snakes. Nature 27: 205–208.
- 4. British Medical Journal (1892) Snake bites. BMJ 2: 620.
- 5. British Medical Journal (1927) Indian sanitary reports. BMJ 1: 538–539.
- 6. Swaroop S, Grab B (1954) Snakebite mortality in the world. Bull World Health Organ 10(1): 35–76.
- 7. Government of India, Central Bureau of Health IntelligenceHealth Status Indicators, National Health Profile 2007 and 2008 (Provisional): 18.104.22.168 State/UT wise Cases and Deaths Due to Snake Bite in India. 107–108. Available: http://cbhidghs.nic.in/writereaddata/mainlinkFile/Health Status Indicators.pdf (accessed 2011).
- 8. Hati AK, Mandal M, De MK, Mukherjee H, Hati RN (1992) Epidemiology of snake bite in the district of Burdwan, West Bengal. J Indian Med Assoc 90(6): 145–147.
- 9. Sharma SK, Chappuis F, Jha N, Bovier PA, Loutan L, et al. (2004) Impact of snake bites and determinants of fatal outcomes in southeastern Nepal. Am J Trop Med Hyg 71(2): 234–238.
- 10. Kasturiratne A, Wickremasinghe AR, de Silva N, Gunawardena NK, Pathmeswaran A, et al. (2008) The Global Burden of Snakebite: A Literature Analysis and Modelling Based on Regional Estimates of Envenoming and Deaths. PLoS Med 5(11): e218. doi:10.1371/journal.pmed.0050218.
- 11. Chippaux JP (1998) Snake-bites: appraisal of the global situation. Bull World Health Organ 76(5): 515–524.
- 12. WHO SEARO (2010) Guidelines on Management of Snake-bites. New Delhi: WHO Regional Office for South-East Asia. Available: http://www.searo.who.int/LinkFiles/BCT_snake_bite_guidelines.pdf (accessed 2011).
- 13. World Health Organisation (2007) Rabies and envenomings: a neglected public health issue. Geneva: World Health Organisation.
- 14. Jha P, Gajalakshmi V, Gupta PC, Kumar R, Mony P, et al. (2006) Prospective Study of One Million Deaths in India: Rationale, Design, and Validation Results. PLoS Med 3(2): e18. doi:10.1371/journal.pmed.0030018.
- 15. Registrar General of India (2009) Report on causes of death in India, 2001–2003: Sample Registration System. New Delhi: Government of India. Available: http://cghr.org/publications/Causes_of_death_2001-03_20RGI_CGHR.pdf (accessed 2011).
- 16. Registrar General of India (2004) Sample Registration System. New Delhi: Government of India.
- 17. Jha P, Jacob B, Gajalakshmi V, Gupta PC, Dhingra N, et al. (2008) A nationally representative case-control study of smoking and death in India. N Engl J Med 358: 1137–1147.
- 18. Jha P, Kumar R, Khera A, Bhattacharya M, Arora P, et al. (2010) HIV mortality and infection in India: estimates from a nationally-representative mortality survey of 1.1 million homes. BMJ 340: c621.
- 19. Dhingra N, Jha P, Sharma VP, Cohen AA, Jotkar RM, et al. for the Million Death Study CollaboratorsAdult and child malaria mortality in India: a nationally representative mortality survey. Lancet 2010; 376: 1768–74.
- 20. World Health Organization (1994) International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD – 10). Geneva: World Health Organization.
- 21. Sinha D, Dikshit R, Kumar V, Gajalakshmi V, Dhingra N, et al. (2006) Technical document VII: Health care professional's manual for assigning causes of death based on RHIME household reports. Toronto: Centre for Global Health Research, University of Toronto. Available: www.cghr.org/project.htm (accessed 2011).
- 22. Rajeevan M, Bhate J, Kale JD, Lal B (2006) High resolution daily gridded rainfall data for the Indian region: Analyses of break and active monsoon spells. Current Science 91: 296–305.
- 23. Srivastava AK, Rajeevan M, Kshirsagar SR (2008) Development of a high resolution daily gridded temperature data set (1969–2005) for the Indian: National Climate Centre Research Report. Pune: India Meteorological Department, Government of India.
- 24. United Nations Department of Economic and Social Affairs, Population Division (2009) World Population Prospects, the 2008 Revision. New York: World Health Organization. Available: http://esa.un.org/peps/peps_interpolated-data.htm (accessed 2011).
- 25. Bhat MPN (2002) Completeness of India's Sample Registration System: An assessment using the general growth balance method. Population Studies 56: 119–134.
- 26. Saikia N, Jasilionis D, Ram F, Shkolnikov VM (2009) Trends in geographical mortality differentials in India. Rostock: Max Planck Institute for Demographic Research.
- 27. Jha P, Laxminarayan R (2009) Choosing health: an entitlement for all Indians. Toronto: Centre for Global Health Research, University of Toronto. 132 p.
- 28. Knobel DL, Cleaveland S, Coleman PG, Fèvre EM, Meltzer MI, Miranda ME, Shaw A, Zinsstag J, Meslin FXRe-evaluating the burden of rabies in Africa and Asia. Bull World Health Organ 2005 May;83(5): 360–8.
- 29. Fox S, Rathuwithana AC, Kasturiratne A, Lalloo DG, de Silva HJ (2006) Underestimation of snakebite mortality by hospital statistics in the Monaragala district of Sri Lanka. Trans R Soc Trop Med Hyg 100: 693–695.
- 30. Rahman R, Faiz MA, Selim S, Rahman B, Basher A, et al. Annual incidence of snake bite in rural Bangladesh. PLoS Negl Trop Dis 2010 Oct 26;4(10): e860.
- 31. Ariaratnam CA, Sheriff MH, Theakston RD, Warrell DA (2008) Distinctive epidemiologic and clinical features of common krait (Bungarus caeruleus) bites in Sri Lanka. Am J Trop Med Hyg 79(3): 458–462.
- 32. Registrar General of India (2001) Survey of Causes of Death (Rural), Statistical Report: 1998. New Delhi: Government of India.
- 33. Sawai Y, Honma M (1976) Snake bites in India. In: Ohsaka A, Hayashi K, Sawai Y, editors. Animal, Plant and Microbial Toxins Vol 1: Biochemistry. New York: Plenum. pp. 451–460.
- 34. Banerjee SR (1993) Agricultural child labor in West Bengal. Indian Pediatr 30(12): 1425–1429.
- 35. Mukerji AB, Mehta S (1975) Female participation in agricultural labour in India: Patterns and Associations. J of Econ & Social Geography 66(2): 103–107.
- 36. Ahuja ML, Singh G (1954) Snake bite in India. Ind J Med Res 42(4): 661–686.
- 37. Warrell DA (2010) Snake bite. Lancet 375: 77–88.
- 38. Gutiérrez JM, Williams D, Fan HW, Warrell DA (2010) Snakebite envenoming from a global perspective: Towards an integrated approach. Toxicon 56: 1223–35.