Verbal autopsy to assess causes of mortality among the economically productive age group in the tribal region of Melghat, central India

Understanding the cause of death (CoD) is crucial for policy framing and planning programmes to improve health indicators. A reliable assessment of disease-specific mortality rates is nearly impossible in developing countries because the underlying cause (UC) is unknown or relevant information is not recorded due to poor reach and coverage of the death registration system1. Verbal autopsy (VA) is a retrospective inquiry systematically conducted involving the family members of the deceased to collect information pertaining to the symptoms before death through predetermined questions to describe the CoD for home deaths2,3. VA will enable evidence-based decision-making and can guide public health priorities in communities without physician certified deaths4,5.

In most resource constrained, hilly, forest, rural-tribal areas of world with high mortality, documentation of death information and medical certification of CoD is nearly impossible because of linited healthcare access, low health literacy in tribal populations and a high proportion (67 to 75%) of deaths which occur outside of the healthcare setting6,7. The physician-certified verbal autopsy (PCVA) approach is reportedly more effective in determining CoD in community deaths8.

MAHAN trust, a non-governmental organization (NGO) is continually providing holistic medical services to poor tribals of Melghat for the last 25 years, conducted a 17 yr study which revealed that the age-specific mortality rate (ASMR) in Melghat was >400/100000 population2, which is 2.25 times greater than the rest of India (178 in 2016). Deaths in the economically productive age group can have detrimental effects on the community, nation and family members, particularly children as it also reportedly increases malnutrition9 and deaths10.

This study was conducted to ascertain the CoD in the 16-60 yr age group in Melghat, which has previously been reported to have a disproportionately high age-specific mortality rate as compared to the Indian average. In order for intervention to prevent these premature deaths through the village based healthcare for the economically productive age group people (EPAGP), as a first step, VA in EPAGP was needed to know the CoDs to plan the programmes to reduce mortality. The objective of this study was to determine the underlying cause (UC) for cause-specific mortality and to determine contributory causes (CC) for all causes of mortality by VA in EPAGP in Melghat.

Material & Methods

A community based longitudinal prospective study was conducted from January 2004 to December 2020 in EPAGP in 34 villages of Melghat. The study was ethically approved by the Institutional Review Board of MAHAN trust. A written informed consent was obtained from all participants and the Gram Sabha of the villages.

Source population: Melghat is a hilly, forest, difficult to access, highly impoverished tribal area in Amravati district of Maharashtra. It is spread over 4000 km², across 320 villages. The population is 300,000 with 84 per cent poor tribal, 50 per cent semiliterate/illiterate, mostly small farmers/labourers. Melghat has poor transportation, grossly inadequate health services and low health seeking behaviour, leading to high mortality11.

Sample size: In all 34 villages (10% of total 320 villages) from two blocks of Melghat were considered for this study (Fig. 1). A minimum of 50 per cent of total mortality was decided as the criteria for VA, considering the feasibility of such exercise in the tribal population. The World Health Organization (WHO) has recommended cluster sampling and minimum population size of 2000-20000. A two stage cluster sampling technique was employed. The first stage is a block and the second is the villages within each block. Eight clusters/villages from five zones were randomly picked up by the lottery method. Finally, 34 clusters were selected based on the willingness of the villagers and village health workers (VHWs) to participate and were block stratified. In the year 2016, two villages were dropped from the study due to administrative reasons and hence data was analysed from 32 villages. The sample unit comprised all deaths in the EPAGP in the selected villages. The total number of deaths during the period was 1011. Verbal autopsies were conducted in 763 (75.46%) deaths.

Close

Fig. 1

Selection of study participants.

Export to PPT

Inclusion and exclusion criteria: De-facto method (home and hospital deaths in the study area) was used12. Only close relatives of all deceased individuals who gave consent were included in this study. Those who migrated (≥6 months ago) were excluded from the study.

Results

The completeness of death reporting was ~95 per cent. The population of the 32 villages was 29,356, with 15,002 (51.1%) males and 14,354 (48.9%) females. The minimum sample size needed was 500 VA. Seven hundred and sixty three VA were randomly undertaken to improve the confidence interval and accuracy. There were no significant differentials in the response rates over time or by age/sex of the deceased. Only response rates varied across surveyors.

Table I shows age sex distribution of the study population and occurrence of deaths during three study periods. In the 21-30 yr category, during 2010-2015 and 2016-2020, the proportion of male deaths was significantly higher than that of females (P<0.001 and P=0.032, respectively). In the 41-50 yr category, the proportion of male deaths was significantly higher than females (P=0.034) for the period 2010-2015. In the 51-60 yr category, for the period 2016-2020, the male deaths were significantly higher than that of females (P=0.003). Overall, for the last two study periods, the male deaths were significantly higher than females. The male deaths (all age groups (16-60 yr), 21-30 and 51-60 yr groups) were significantly higher than that of females for total 17 yr (P<0.001, P<0.001 and P<0.001, respectively). There was a significant increase in deaths per 10,0000 population per year in the age group of 41-50 yr and 51-60 yr as compared to previous age groups. Table II shows the age wise distribution of cause-specific mortality during three study periods.

CNS infections were more common among the older age groups as compared to the younger one. TB was the most common CoDs in all age groups except in the 51-60 yr age group.

Table III provides the gender-wise distribution of CoDs and CSMF during the three study periods. The subjects comprised of 41.68 per cent females and 58.32 per cent males.

TB was the leading UC as well as CC of deaths acros sall age groups, equally distributed across both genders. Jaundice was the second leading CoD followed by heart diseases and diarrhoeal diseases. Heart diseases, jaundice, homicide, alcohol intoxication, tetanus, drowning, snakebite and alcohol liver diseases were more common in males. Suicide and road traffic accidents (RTAs) were significantly more common in males than females (P=0.046 and 0.018, respectively). Deaths due to tetanus in Melghat were significantly higher (7.75 per 100000 population per year) as compared to the rest of India (1.12 per 100,000 UI) and equal to the highly impoverished countries such as south Sudan (7.62 per 100,000 UI)17. Out of total deaths, 50 per cent of deaths are due to preventable infections. Deaths due to TB were also higher in Melghat than rest of the India.

Table IV and Figure 2 provides the seasonal pattern of major UCs of death during the three periods. During the period 2011-2015, the proportion of deaths due to TB and jaundice was significantly higher in the winter and summer seasons as compared to the rainy season. Further, the deaths due to TB were significantly higher in winter as compared to other seasons (P<0.001) and the deaths due to diarrhoeal diseases were significantly higher in the rainy season as compared to other seasons (P=0.013). The proportion of CNS infection was also significantly higher in the rainy season as compared to other seasons (P=0.002) and deaths due to heart diseases were significantly higher during the summer season (P=0.044). Chronic obstructive airway disease (COAD) was higher during the monsoon.

Close

Fig. 2

Seasonal variation of causes of deaths during the study period (2004-2020). CNS, central nervous system; COAD, chronic obstructive airway disease

Export to PPT

Discussion

In this study, the deaths in EPAGP were found to be higher than the national average. Health outcomes in this population have been reported to be poor because of poverty, low heath literacy, inaccessible and grossly inadequate healthcare services, low health seeking behaviour18, poor hygiene19 and high prevalence of addiction20.

A high rate of premature adult mortality is reported as a major issue at the population level as it creates a negative impact on families and communities21. CoD across EPAGP in India and worldwide shows respiratory infections, diarrhoeal diseases and non-communicable diseases as the major CoD1,3,22, a similar trend was also observed in Melghat.

In the present study, TB was found to be the most common underlying CoD across all age groups indicating shortfalls in outreach of the existing TB control programme and the need for extensive revision of such programmes suitable for the tribal areas. The prevalence of TB is high in the tribal area of Melghat (>0.4%), twice as compared to the national average due to low health literacy, crowded living conditions, delayed diagnosis, high prevalence of smoking and malnutrition, inaccessible health facilities and poverty23,24. Jaundice was reportedly the second most common UC across all age groups. This could be due to abuse of single use unsterile injections to multiple patients25 by unlicensed healthcare providers, multiple sexual partners26 drinking impure water22 and blood transfusion without testing for hepatitis virus27. Heart and diarrhoeal diseases were the third and fourth leading CoD. High diarrhoeal mortality is because of impure drinking water (90%) and impure food, inadequate hand hygiene and a high prevalence of malnutrition (body mass index<18.5)28. Intensive vaccination can prevent many deaths due to infections.

Heart diseases were the second leading CoD in the age group of 41-60 yr. The shift from infectious to non-communicable diseases in the adult age group shows the epidemiological transition in India29. The risk factors for heart diseases in Melghat are high tobacco and alcohol use, high prevalence of hypertension (10% in EPAGP), low birth weight (43%)30,31 and severe malnutrition in the age group of 0-5 yr (20%)30. Suicide is the sixth most common UC CoD mainly seen in 16-40 yr’ group indicating significant mental health issues in younger population warranting attention32. Many of the suicides were under the influence of alcohol or due to family disputes. Depression and anxiety may go unnoticed in the tribal areas in the context of limited family support. The greatest burden of suicide is seen in young people in developing countries33.

The probable causes of significantly high death rate in males in the present study were high risk behaviour of males e.g., smoking, alcoholism, more outdoor activities, occupational hazards and mental health issues, etc. These findings are in coraboration with other studies7. In the present study, jaundice deaths were more frequent in males due to more high risk behaviour e.g., multiple sex partners, alcoholism (50%), drinking impure water because of more frequent outdoor activities similar to previous reports4,7. Diarrhoeal deaths were more common among males, because of more outdoor work particularly due to consumption of impure outdoor water and impure food, and lesser tendency for handwashing with soap. Similar findings have been found in other low-income settings, such as in Gaza, where the prevalence of diarrhoea is reportedly higher in males (5.4/100) as compared to females (1.3/100)34. On the other hand in Melghat death due to heart diseases (coronary artery disease) are reportedly more common in males due to the higher prevalence of smoking, alcoholism and hypertension20.

In the present study, suicide deaths were more common in males than females similar to previous reports, indicating significantly more personal/social reasons and mental disorderin males35. Many suicides in Melghat were recorded as under the influence of alcohol. In present study, COAD (chronic obstructive airway disease) or COPD (chronic obstructive pulmonary disease) deaths were more common in males due to more smoking and outdoor pollution. RTA as CoD was significantly more in males than females. Motorbikes are the important mode of transport in Melghat due to inadequate public transport. Driving motorbikes, especially under the influence of alcohol is significantly higher amongst the tribal males. Similar findings were found in Iran36.

In the present study, significantly higher deaths occurred during the rainy season. The difference in the proportion of deaths across season was significant (P=0.002). It is because of more deaths due to diarrhoeal diseases, CNS infections and COAD during the rainy season. This highlights that the onset of rainy season opens the door to various infections and diseases due to unhygienic conditions. Poor transport connectivity during this season further inhibits access to hospitals. In the present study, TB deaths were found to be maximum in the winter season in Melghat when the temperature reaches to 2°C and people live in enclosed huts without proper ventilation leading to a greater risk of TB exposure. It is in contrast to the majority of studies (n=49), where TB decreased during winter24. In present study, diarrhoeal deaths were significantly more common during the rainy season, similar to previous studies37. It is because of more impure drinking water and more flies infecting food during the rainy season. There is no system for water purification and control of flies in Melghat. Open defecation (>90%) exacerbates more contamination of water and contamination by flies (unpublished data). Deaths due to CNS infections were more common in Melghat during the rainy season because of unhygienic surroundings in villages, poor drainage systems and water logging leading to mosquito breeding and more cerebral malaria/viral encephalitis, similar to the findings from a rural hospital in Uganda38.

In our study, COAD was more common during the rainy season with increased morbidity and mortality, due to the increased prevalence of respiratory viral infections6.

We used these CoD by VA to plan cluster randomized control trial (CRCT) to reduce ASMR. Our CRCT resulted in significant reduction in ASMR in intervention villages as compared to control villages (P<0.001), demonstrating the reliability of VA. Accurate information of CoD is essential for health policy, planning, monitoring, evaluation, comparisons, public attention and reducing premature deaths39.

This study was not without limitations. The study team chose to use the physician-reviewer method for diagnosis. Diagnosis of TB was not made by bacteriological confirmation. The accuracy of VA to know the CoDs was 75 per cent. This was a long-term follow up study of 16 yr. During the study, the interviewers were changed and the quality of data obtained could have been affected due to the poor educational status of responders. Furthermore, some villages were dropped out of the study and new villages were added in between the tenure. This may have limited the interpretation of trends in CoD over time.

Overall, our data showed that communicable diseases are the major CoDs (59.18%) in EPAGP. Providing oral rehydration solution (ORS) to all diarrhoea cases can save many lives immediately. Improving access to TB diagnosis and treatment services and ensuring treatment adherence can break the chain of transmission and prevent unnecessary deaths due to TB. It is essential to strengthen the routine death registration systems to cover all the locations and population groups across India.

As per our study finding from 2004 to 2021, the mortality rate in EPAGP among tribals in Melghat is more than 400 per 100,000 population of that age group per year. If we extrapolate our findings to all tribal areas of India, then, more than 208,846 people of EPAGP are dying every year40.