Discussion
As demonstrated in this study, SCD is a serious disease in the aboriginal population of the Gudalur Valley, as there is a substantially higher risk of premature death. . Despite the high prevalence of SCD among the aboriginal population, there have been only a few studies documenting the causes of death and mortality rates.16 The mortality rate identified in this study is comparable to that observed in Brazilian patients (16.8 %) with SCD.17
Based on our newborn screening data we found an SCT rate of 17.1% and a SCD rate of 1.1% respectively. The SCT rate of 8.4% by Hb electrophoresis could be falsely low because of duplications as some individuals would have been tested more than once as their Unique ID numbers were recorded inaccurately. Of the 16 newborns detected to have SCD screened over 7 years (Dec 2011 – 2018), there were no deaths in the <5 age group. This low mortality is intriguing as it contrasts with other studies in low resource settings18,19 and especially since none of the patients received prophylactic penicillin or pneumococcal vaccine. However, three of the 16 <5 children were admitted at GAH with respiratory infections and one with acute splenic sequestration requiring blood transfusion. Survival of all the <5 children in this aboriginal SCD population may thus, potentially, be related to timely medical care.
Seventy percent (70%) of deaths were in the 6-29 age group, with the 19-29 y age group having the highest mortality (45.4%) with 24.6% of deaths in the 6-18 age group. In the study from Brazil, out of 281 SCD deaths, most occurred in the 6-11 year age group (35.9%)16.
The median age of death in the non-SCD aboriginals population of Gudalur during the study period was 55 years, which is 20 years more than the median age of death of 25 years in those with SCD. These data suggest significant premature mortality in this SCD aboriginal population. The high rate of early mortality reported in this study stands in contrast to the report by Feroze et al from the adjacent district of Wayanad which suggested that SCD disease is of a milder phenotype in the aboriginal population.10 As the authors state, their study of 23 patients with SCD was too small to draw any conclusions about disease phenotype. Further, they provided no details of the setting in which the individuals received medical care or the completeness of the collection of clinical complications. It is of interest that the authors report the differences in literacy and outcomes between different tribes in the area. They too observed that the Paniyas had the lowest literacy and the highest mortality. Thus, the study of disease phenotype in aboriginal populations who are faced with extreme isolation and severe disadvantage must be carried out in the context of comprehensive care, close follow up and surveillance, and accounting for social determinants of health.
The predominance of deaths in males is inconsistent with the overall cohort characteristics (Table 1) but is consistent with the greater survival in women as reported in other studies17,20
Just under half the deaths occurred at home despite the availability of free government ambulance services and free medical care at GAH. The persistence of barriers to care in a community owned, community centered system of care in this extremely disadvantaged community bears further investigation.
The causes of death in our SCD population were the sequelae of the disease itself. None of these patients died of non-communicable diseases, suicides, or trauma, which were the common causes of death in the non-SCD aboriginal community. In our cohort, the overwhelming majority of the deaths occurred due to acute chest syndrome, anemia, and sepsis which have been described as leading causes of death in other studies.21,22 The absence of deaths due to trauma can be explained by the fact that, unlike the non-SCD aboriginal population, they do not undertake hazardous work.
The Paniya tribe formed the greater part of the diagnosed SCD cases (64%), which is consistent with their being the largest group, constituting 58% of the aboriginal population in the Gudalur Valley. It is not surprising that all the deaths occurred in the Paniya tribe. They constitute the most disadvantaged of the four aboriginal groups and were, until recently, bonded laborers, a modern form of slavery. Despite their liberation from bonded labor, the Paniyas continue to experience several ’vulnerability traps’ of isolation, education deficits,10 indebtedness, alcohol use, loss of community bonds, and traditional leadership. These factors may explain why health outcomes in the Paniya population continue to be poorer than the other aboriginal populations in the area. These findings may suggest that improving health outcomes in this vulnerable population, would require a broad strategy that addresses the social determinants of health and their intersections with healthcare delivery and outcomes. It is noteworthy that although the Mullu Kurumbas form only 7% of the population, 17.8% of the SCD patients were from this group. This is probably because of the practice of endogamy within this small tribal community
Observed mortality has decreased between 2008 to 2018. Early screening for SCD and the availability of a comprehensive health system with community-based programs may account for some of this improvement. It may also be secondary to higher economic and social status resulting from increased accessibility and availability of education and employment.
A major limitation of the study is that many of the deaths occurred outside the hospital setting. However, since the study was conducted in the context of a community directed and community based comprehensive care, we are confident that we have been able to track all births and deaths in this community.  All aboriginal deaths in this area are reported to the Area Nurses and AMS (Aboriginal Advancement Society) members, who support and are also often present for the last rites. Deaths of individuals working outside the area are uncommon but are also reported. The Area Nurses are all from the aboriginal community and have extensive personal connections. Of note, the accurate age of the older adults in the aboriginal population is difficult to determine because of the absence of birth certificates and varied self-reporting of age. The age of death in this group was determined using the village census list. However, the age of the younger people was accurate as it was taken from their birth records which are more organized since 1990 when ASHWINI started and a Unique ID number was given to each individual.
We have presented the results over 11 years and hence could not account for multiple population-level changes that may have occurred during the study period. As there were improvements in the educational and financial conditions within the aboriginal population during this period, the decrease in mortality rate observed cannot be attributed to the availability of comprehensive and integrated sickle cell care alone. Furthermore, the events preceding home deaths were analyzed and documented using WHO VA tool12. The causes of mortality recorded in this survey may not be entirely reliable considering the time that may have elapsed since death, introducing recall bias. Also, there could have been more than one reason for death, but we only recorded the primary cause of death.
In conclusion, the data presented in this manuscript indicate that SCD is a disease with high premature mortality among the aboriginal population of the Gudalur Valley.  The recognition of the severity of SCD in this population provides the rationale for the design and implementation of newborn screening and comprehensive care for SCD. These programs must also address the socioeconomic and cultural barriers to the delivery of care to these extremely disadvantaged aboriginal populations.
Conflicts of Interest: None declared
Acknowledgments: We acknowledge Ms.Malathi Manikandan and Ms. Pradeepa for the data collection. We thank Drs. Anna Oommen,  Roshina Sunny, and Meredith Brooks for their valuable advice. We thank the Government of Tamil Nadu for their support.
References :
  1. Piel FB, Hay SI, Gupta S, Weatherall DJ, Williams TN. Global burden of sickle cell anemia in children <5, 2010-2050: modelling based on demographics, excess mortality, and interventions. PLoS Med. 2013;10: e1001484.
  2. McGann PT. Sickle Cell Anemia: An Underappreciated and Unaddressed Contributor to Global Childhood Mortality. The Journal of Pediatrics 2014 July;165(1), 18-22
  3. Ware RE, de Montalembert M, Tshilolo L, Abboud MR. Sickle cell disease. Lancet 2017;390(10091):311-323.
  4. Lehman H, Cutbush M. Sickle cell trait in southern India. Brit Med J 1952; 1: 404-5.
  5. Kaur M, Das GP, Verma IC. Sickle cell trait & disease among aboriginal communities in Orissa, Madhya Pradesh & Kerala: Indian J Med Res 1997 Mar;105:111-6
  6. Census of India 2011. Office of the Registrar General and Census Commissioner.
Ministry of Home Affairs, Govt of India.http://www.censusindia.gov.in
  1. Therrell BL, Lloyd- Puryear MA., Eckman JR., Mann MY Newborn Screening for sickle cell disease in the United States: A review of data spanning 2 decades Semin Perinatol 2015 Apr;39(3):238-51
  2. Lubeck D, Agodoa I, Bhakta N et al. Estimated life expectancy and income of patients with sickle cell disease compared with those without sickle cell disease. JAMA Network Open. 2019;2(11):e1915374.
  3. Kar BC, Satapathy RK, Kulozik AE, Kulozik M, Sirr S, Serjeant BE, et al. Sickle cell disease in Orissa State, India. Lancet. 1986;ii:1198–201.
  4. Feroze M, Aravindan K. Sickle cell disease in Wayanad, Kerala: Gene frequencies and disease characteristics. National Medical Journal of India 2001; 14:267-270
  5. Nimgaonkar V, Krishnamurti L, Prabhakar H, Menon N. Comprehensive integrated care for patients with sickle cell disease in a remote aboriginal population in southern India. Pediatr Blood Cancer 2014; 61: 702-705.
  6. World Health Organization. Verbal autopsy standards: 2016 WHO verbal autopsy instrument. Geneva, Switzerland: World Health Organization 2016. Available from: http://www.who.int/healthinfo/statistics/verbalautopsystandards/en/
  7. Anderson and Newman. Revisiting the behavioral model and access to medical care:   does it matter? Journal of Health and Social Behavior. 1995; 36:1-10. Framework of Health Services Utilization.
  8. Kumar S, Crouse Quinn S, Kim KH, Musa D, Hilyard, KM, & Freimuth VS. (n.d.). The Social Ecological Model as a Framework for Determinants of 2009 H1N1 Influenza Vaccine Uptake in the US.https://doi.org/10.1177/1090198111415105
  9. Registration of  Births and Deaths Act 1969 Accessed at www.censusindia.gov.in.
  10. Shah P, Bhagat VM, Patel K, Patel C. Pattern of mortality in sickle cell disease: an autopsy study. Int J Res Med Sci. 2017;5:2115-2119
  11. Lobo CLC, Nascimento EMD, Jesus LJC, Freitas TG, Lugon JR, Ballas SK. Mortality in children, adolescents, and adults with sickle cell anemia in Rio de Janeiro, Brazil. Hematol Transfus Cell Ther. 2018;40(1):37–42
  12. Thomas AN, Pattison C, Sergeant GR. Causes of death in sickle-cell disease in Jamaica. BMJ 1982:285(633-635)
  13. Grosse SD, Odame I, Atrash HK, Amendah DD, Piel FB, Williams TN. Sickle cell disease in Africa: a neglected cause of early childhood mortality Am J Prev Med 2011 Dec;41(6 Suppl 4): S398-405
  14. Wierenga KJ, Hambleton IR, Lewis NA. Survival estimates for patients with homozygous sickle-cell disease in Jamaica: a clinic-based population study. Lancet 2001;357(9257):80–3.
  15. Al-Suliman A, Elsarraf N.A, Baqishi M, Homrany H, Bousbiah J, Farouk E. Patterns of mortality in adult sickle cell disease in Al-Hasa region of Saudi Arabia. Ann Saudi Med 2006; 26:487-488.
  16. Serjeant GR, Chin N, Asnani MR, Serjeant BE, Mason P, Hambleton I.R, Knight-Madden JM. Causes of death and early life determinants of survival in homozygous sickle cell disease: The Jamaican cohort study from birth. PLoS One.2018;13:e019271.