Introduction

Diarrheal diseases, pneumonia, and other infectious diseases are leading causes of death among children younger than five years in low and middle income countries and also in India [1], [2]. However, little is known about the causes of death in children after age five years. The Global Burden of Disease and Risk Factors (GBD) estimates that in 2004 there were approximately 69 000 deaths from pneumonia and 1000 deaths from diarrheal diseases among children aged 5–14 years in India, accounting for approximately 20% of all deaths at these ages [3]. However, medical certification of deaths in India and other south Asian countries is uncommon. Thus, estimates of mortality in this age group are largely derived from models based on estimates from countries where vital registration data is available and older, non-representative studies of causes of death in India.

The objective of this study is to provide the first ever direct national and sub-national estimates of deaths from the major infectious causes among Indian children ages 5 to 14 years, with a focus on diarrheal diseases and pneumonia.

Methods

Results

There were a total of 3 855 deaths in children aged 5 to 14 years from 2001 to 2003 in the study sample. Based on the total number of deaths by gender, the overall all cause mortality rate for boys and girls aged 5 to 14 years was 130 and 148 per 100,000 respectively. Table 1 shows the top causes of death at these ages. In boys aged 5 to 14 years, diarrhea was the cause of 12·0% (99% CI; 8·7 to 16·3) and pneumonia was the cause of 4·3% (99% CI; 2·5 to 7·3) of all deaths. In girls aged 5 to 9 years, diarrhea was the cause of 17·3% (99%CI; 13·3 to 22·3) and pneumonia the cause of 6·9% (99% CI; 4·5 to 10·3) of all deaths. In the 10 to 14 year old age group, diarrhea was the cause of death of 17·0% (99% CI; 14·2 to 20·2) of boys and 20·6% (99% CI; 17·7 to 23·9) of girls whereas pneumonia was the cause of 10·9% (99% CI; 8·6 to 13·7) and 13·6% (99% CI 11·1 to 16·5) of boy and girls deaths respectively. The distribution of ICD-10 codes for the diarrheal and pneumonia deaths are shown in Table S2. Malaria (8·3%, 99% CI; 6·7 to 9·0), central nervous system infections (6·1%, 99% CI; 5·1 to 7·4), and measles (4·5%, 99% CI; 3·6 to 5·5) constituted the remaining top five infectious diseases causes of death in all children aged 5 to 14 years. The remainder of the deaths in children aged 5 to 14 years were due to injury (21·6%, 99% CI; 19·0 to 24·3), non-communicable diseases (20·6%, 99% CI; 18·8 to 22·5), and other infectious diseases (11·7%, 99% CI; 10·3 to 13·2). Overall, 86% of deaths occurred in rural areas (range from 82% for non communicable diseases to 94% for malaria) and only 16% occurred in a health facility. The most common symptoms preceding death from the major infectious diseases are shown in Table 2. The remainder of this paper will focus on diarrheal and pneumonia mortality since a complete analysis of malaria mortality in India (including children) is published elsewhere [11].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Causes of Death in Children 5 to 14 Years old and Comparison to Global Burden of Disease Estimates.

https://doi.org/10.1371/journal.pone.0020119.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Presence of Symptoms by Cause of Death (%).

https://doi.org/10.1371/journal.pone.0020119.t002

The analysis of physician agreement demonstrates that physicians were highly likely to agree on both diarrhea and pneumonia as causes of death during the first review of the verbal autopsy (Table S3). The overall agreement was 94% for both causes. For diarrheal diseases inter-rater agreement (Cohen's kappa coefficient) was 0·79 (99% CI; 0·75–0·82) and for pneumonia it was 0·67 (99% CI; 0·61–0·72). Initial physician agreement exceeded 90% for boy and girl deaths for both diarrheal diseases and pneumonia.

We estimate there were approximately 59 300 (99% CI; 51 900 to 68 700) deaths due to diarrhea and 34 200 (99% CI; 28 700 to 42 300) deaths due to pneumonia in children aged 5 to 14 years in 2005. These correspond to mortality rates of 24·1 (99% CI; 21·1 to 27·9) and 13·9 (99% CI; 11·7 to 17·2) per 100 000 persons, respectively (Figure 1). Proportional mortality from both diarrhea and pneumonia were highest in the north of the country (Figure 2). Diarrheal deaths peaked in the period between June and August (Figure 3). There were significant differences in diarrhea mortality rates by region and gender. Girls had an approximately 50% higher mortality rate due to diarrhea than boys (43% greater in ages 5 to 9 years and 61% greater in ages 10 to 14 years). These differences result in approximately 35% more annual diarrheal deaths among girls (n = 34 200) compared to boys (n = 25 100). There was important regional variation; the diarrhea mortality rate was higher than 25 deaths per 100 000 children aged 5 to 14 years in the northeast, east, and central regions and lower than 12 in the north, west, and south regions. At ages 5 to 14 years, the mortality rate from pneumonia was approximately 50% higher in girls compared to boys, and in the 5 to 9 year old age group it was more than 60% higher in girls compared to boys. Similar to diarrhea mortality, the highest pneumonia mortality was seen in the east, central, and north east regions and the pneumonia mortality rate in the Central region (22·0 per 100 000 (99% CI; 17·0 to 29·3) was nearly 25 times higher than in the South (0·9 per 100 000 (99% CI; 0·2 to 6·2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Estimated total deaths and mortality rates from diarrhea and pneumonia in children aged 5 to 14 years.

https://doi.org/10.1371/journal.pone.0020119.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Diarrhea (left) and Pneumonia (right) Proportional Mortality.

https://doi.org/10.1371/journal.pone.0020119.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Seasonality of Diarrhea and Pneumonia Deaths.

https://doi.org/10.1371/journal.pone.0020119.g003

Discussion

This is the first study to directly estimate causes of death in children aged 5 to 14 years in India. Infectious diseases cause over 196 000 annual deaths (approximately 60% of all deaths) in this age group. Diarrhea, pneumonia, and malaria are the three main infectious causes of death. We estimate that there were approximately 59 000 deaths due to diarrhea in 2005 (19% and 15% of all deaths in the 5 to 9 year and 10 to 14 year categories respectively) and 34 000 deaths due to pneumonia (13% of 5–9 years old deaths and 7% of 10 to 14 year old deaths).

One of the most important findings of our study is the gender difference in mortality; more girls than boys that died from not only infectious diseases in general but also each of the specific infectious diseases studied. For both diarrhea and pneumonia, mortality rates for girls were nearly 50% higher than boys. Increased mortality rates among Indian girls has previously been shown for children at ages 1 to 59 months [2] and these results show that excess mortality due to infectious diseases among girls extends into older childhood as well. This apparent gender bias may be enacted via parents being less likely to immunize [12], [13], seek medical attention [14], and/or being less likely to use appropriate antibiotic therapy for their sick female children [14]. The gender difference in infectious disease related mortality suggests that there are significant numbers of potential lives saved if mortality rates in girls could be reduced to those already seen in boys. As one example, if the national diarrhea mortality rate observed among children 5 to 14 years were as low as that observed among boys in the south (7·6 per 100,000 people, 99% CI; 4·1 to 13·8), there would be approximately 40 000 less diarrheal deaths nationally in this age group, a reduction of approximately 70%.

A review of data from the Registrar General of India's Survey of Causes of Death (SCD) [15] suggests that diarrhea mortality in older children has been high in India for some years. The SCD was conducted in a sample of villages from selected rural Primary Health Centres throughout India's major states. Table S4 shows the trends in mortality from diarrhea (diarrhea, cholera, food poisoning or dysentery) and pneumonia (pneumonia and whooping cough) from the SCD between 1991 and 1998. In the SCD, the annual proportion of deaths among children 5 to 14 years of age ranged from a low of 8·1% to a high of 13·1% for diarrhea and from a low of 6·4% to a high of 10·8% for pneumonia compared to 18% and 10% respectively in our study. While using a different methodology from our study, the SCD estimates support our findings that diarrhea is a major cause of death even among children older than 5 years of age. It is also worthwhile to examine other sources of information on causes of death in urban children aged 5 to 14 years. The Medically Certified Causes of Death (MCCD) [16] collects data in certain hospitals, generally in urban areas which are selected by the Chief Registrar of Births and Deaths. In 2004 the MCCD estimated that 2·5% of urban deaths in this age group were caused by diarrhea and 8·1% by pneumonia.

A recent systematic review examined published studies to estimate diarrhea morbidity among children aged 5 to 14 years and found that the median annual incidence rate was 67·5 episodes (IQR 15·4–91·7) per 100 person-years in the WHO South East Asia Region [17]. The same review included six studies of diarrhea mortality in older children, adolescents, and adults; however, only two of these studies included children aged 5 to 14 years [17]. One of these studies, conducted in Indonesia, found that diarrheal diseases were the cause of 22·2% of all deaths in children aged 5 to 14 years [18].

A larger study that used physician coded verbal autopsies in Senegal found that 12·5% of all deaths in the 5 to 14 year old age group were caused by diarrhea [19]. Though limited in number and in scope, the published literature is consistent in its findings. Our study now provides the most reliable evidence to date for India, based on the largest number of deaths, and the first from South Asia, to corroborate that diarrhea is a major cause of death in children older than five years.

For diarrheal diseases, we used the commonly used definition of Intestinal Infectious Diseases from the ICD-10 (codes A00 to A09). Of all diarrheal deaths studied, approximately 80% were physician coded as A09 (infectious gastroenteritis and colitis) and 16% were coded as A01 (typhoid and paratyphoid fevers) (Table S2). In young children, typhoid is less likely to be associated with diarrhea symptoms and more likely to include fever, constitutional symptoms, gastrointestinal disturbances and hepatosplenomegaly [20], [21]. The different clinical presentations of deaths coded as gastroenteritis and typhoid fevers is evident in Table 2. Overall, 16·3% (110/675) of the diarrheal deaths were coded as typhoid/paratyphoid, corresponding to approximately 9000 annual diarrheal deaths due to typhoid and paratyphoid fever in children aged 5 to 14 years in India.

Diarrhea, pneumonia, central nervous system infections, and measles represent nearly 130 000 annual deaths in this age group and are all, to a degree, vaccine preventable diseases. While the microbiologic etiology of severe diarrheal disease in older children remains incompletely understood, both rotavirus and Salmonella typhi are likely players and both are preventable by safe and effective vaccines. Streptococcus pneumoniae and Haemophilus influenza type B are well recognized blood, lung, and central nervous system pathogens and these two organisms have been estimated to cause over 200 000 deaths in all ages in India annually [22], [23]. Infections from both of these organisms, in addition to Neisseria meningitides, another major bacterial cause of central nervous system infections, are largely preventable through the use of existing vaccines. Measles, has been effectively eliminated in much of the world through vaccination. The availability of effective preventive tools and the fact that control of infectious diseases (including diarrhea, pneumonia, central nervous system infections, and measles) have not been made a priority in this age group suggest that there is significant potential to reduce mortality from these causes among older children in India.

There are several limitations to this study, the major one being that verbal autopsy based studies may be biased by misclassification of causes of death. An analysis of over 27 000 childhood deaths in the MDS did find that physicians were more likely to agree on the categorical cause of death in children aged 5 to 14 years than in any other age category [24]. Additionally, the symptoms based tabulation (table 2) is very much consistent with what would be expected given the clinical presentation of the various conditions. However, this does not necessarily validate that the cause of death agreed upon by the physicians is in fact correct. Validation studies comparing verbal autopsy classification of major causes of death to known hospital diagnoses or death certificates as the gold standard have been conducted for children under 5, adults, and maternal deaths [25]–[30], however, it is likely that there are differences between hospital deaths and deaths that occur at home. To our knowledge, no such validation study has been performed specifically on deaths between the ages of 5 to 14 years. Multiple verbal autopsy validation studies in children show that the sensitivity and specificity of verbal autopsy diagnosed diarrhea and pneumonia versus medically certified cause of death is in general reasonable. For diarrhea, studies from the Philippines [31], Namibia [32], Bangladesh [33], South Africa [27], Uganda [33], and Nicaragua [33], [34] show sensitivities ranging from 51% to 88% and specificities ranging from 54% to 100%. For pneumonia, the same studies [27], [31]–[34] show sensitivities ranging from 58% to 91% and specificities from 37% to 91%. While there is an inherent possibility of misclassification in any individual death, for major causes of death verbal autopsy does provide estimations of overall causes of death within the population sufficient to make actionable policy recommendations.

Another limitation is that we were unable to determine the contribution of underlying medical co-morbidities, such as caloric or protein malnutrition, micronutrient deficiencies, immune deficiencies, and others, that may have contributed to the deaths. We are also unable to determine role that poor access to health services, most specifically oral or intravenous rehydration for diarrheal disease and antibiotic therapy and respiratory supportive care for pneumonia, may have played in the deaths of these children. However, the fact that over 85% of both the diarrheal and pneumonia deaths in this study occurred in rural areas and only 12% of diarrhea deaths and 10% of pneumonia deaths occurred in health centres suggests that the majority of care for these children, if any, was provided locally and likely by those lacking comprehensive health training.

Supporting Information

Acknowledgments

The Registrar General of India established the SRS in 1971, has continued it ever since, and is collaborating with several of the authors on the ongoing Million Death Study. The opinions expressed in this paper are those of the authors and do not necessarily represent those of the Government of India. We thank Brendon Pezzack and Peter Rodriguez for data support.