ट्रांसलेशनल मेडिसिन
खुला एक्सेस

अमूर्त

Cause of Death by Validation from Verbal Autopsy in Selected Sites of Rural South India

Basavarajaiah Doddagangavadi Mariyappa, Bhamidipathi Narsimhamurthy

Background: Mortality data bycause of death for the entire cross-section of the population in the country is essential for informed decision making in the health sector. Information on cause of death in India is poor because, lack of data base and existing mortality record is incomplete at national level. The identification of cause of death is not feasible due to paucity of medical personnel’s, complexity of datasets and health care management etc. Apart from our Practical experience, the validated verbal autopsy tool is very much important for health administrative convenience and management of medical and non-medical staffs for eliciting information on cause of death at population level. Cause-specific data on mortalityform the core of any health planning strategy. The size and geographical distribution of occurrence of deaths bytheircausesfor different age-group,sex,residence and otherimmense characteristics added greatest value to the public health planners, medical scientist, epidemiologists and researchers. Methods: The Pilot based study was conducted in selected sites of Rural India, the field investigators were collected primary data sets from varied geographical location, an interviewers has visited the sites for extrapolation of accurate cause of death at population level (blinded to the hospital diagnosis it was conducted at greatest accuracy). In a nutshell, the enumerators were collected survey data sets from pretested questionnaires, consequently investigators were adopted various driven mechanisms forcollecting elicited symptoms, signs of the disease conditions etc. All data sets were collected from the direct and focus group interview respectively from selected sites. The study components has derived main and pilot based, In case of main study, field Investigators were collected mortality information in the selected sites through direct interview, focus group interview and also visited door-to-door steps for gathering information pertain to cause of death by survey method. Simultaneously, the medical officers were documented cause of death on various diseases. Finally, observed and unobserved (hospital data) data sets were simulated from logit probability model to test the hypothetical statement about the population. Results: The present study derived two components via pilot and main study, in case of pilot-based study major cause of death was pregnancy and child birth complications with predicted good sensitivity (85.7%), specificity (100%), and positive predictive value (100%) (Odd ratio 4.65; p=0.0012; CI 95%0.63-0.74). Among neonatal deaths, the major cause of death was seen in perinatal (65.0%) (Specificity 95.5% sensitivity 62.5% and PPV 87.5%) odd ratio 3.47; p=0.002; CI95% 0.55-0.71) among child mortality the major cause of death was “infectious parasitic (25.0%) with good specificity ( 71.0%); sensitivity (85.7%); PPV (68.0%) and odd ratio was 4.88;p=0.0001; CI 95% 0.19- 0.28) As we noticed, in maternal deaths, the major cause of death was “pregnancy and child birth (36.0%) besides with specificity (86.7%), sensitivity (50.0%) and PPV (92.9%) odd ratio was 3.11; p=0.002; CI 95% 0.28-0.39). Mortality due to accidents and injuries was (9.0%) with sensitivity was (100%), specificity (95.7%), and positive predictive value was (80.0%) odd ratio was 1.98 p=0.08; CI 95% 6.33-10.19). In case of main study, among adults 204 (29.30%) p=0.018 death due to cardio vascular diseases, 96 (13.81%) p=0.00, respiratory diseases, 80 (11.5%) p=0.011, accidents, 65 (9.3%) p=0.00, senility, 64 (9.2%) p=0.116 gastro-intestinal tract conditions, 60 (8.6%) p=0.001, cancer, 34 (4.9%) p=0.000 and renal failure. Cardio vascular diseases rank first for both years of Potential Life Lost (PLL) and cause-specific mortality rates irrespective of gender.

Conclusion: The verbal autopsy is the instrument, as can be used to estimate the distribution on causes of deaths in rural areas, since the vital registration is ill-defined. The death due to non-communicable diseases accounts (65.30%) on all deaths aged between 20 years and above in the selected sites: the study clearly depicted that; a strong emergency call for public health action would be needed.