Validity of verbal autopsy procedures for determining cause of death in Tanzania

Abstract

Summaryobjectives To validate verbal autopsy (VA) procedures for use in sample vital registration. Verbal autopsy is an important method for deriving cause-specific mortality estimates where disease burdens are greatest and routine cause-specific mortality data do not exist.methods Verbal autopsies and medical records (MR) were collected for 3123 deaths in the perinatal/ neonatal period, post-neonatal <5 age group, and for ages of 5 years and over in Tanzania. Causes of death were assigned by physician panels using the International Classification of Disease, revision 10. Validity was measured by: cause-specific mortality fractions (CSMF); sensitivity; specificity and positive predictive value. Medical record diagnoses were scored for degree of uncertainty, and sensitivity and specificity adjusted. Criteria for evaluating VA performance in generating true proportional mortality were applied.results Verbal autopsy produced accurate CSMFs for nine causes in different age groups: birth asphyxia; intrauterine complications; pneumonia; HIV/AIDS; malaria (adults); tuberculosis; cerebrovascular diseases; injuries and direct maternal causes. Results for 20 other causes approached the threshold for good performance.conclusions Verbal autopsy reliably estimated CSMFs for diseases of public health importance in all age groups. Further validation is needed to assess reasons for lack of positive results for some conditions. keywords cause of death, autopsy/methods, epidemiological measurements, epidemiological methods, vital statistics, child mortality Background

Ascertaining causes of neonatal deaths using verbal autopsy: current methods and challenges

“…Physician review with 2-3 physicians was the most common method used to assign cause of death from the VA. 10,13,14,18,[20][21][22][23][24][25][26][27][28][29][30][32][33][34] In the event of conflicting diagnoses among physician reviewers, discussions to reach consensus 7,16,26,32 or a third party 'majority rules' were applied. 22,25 In cases where all three disagreed, the cause was often labeled as 'undetermined'.…”

“…22,25 In cases where all three disagreed, the cause was often labeled as 'undetermined'. 18,21 Predefined diagnostic categories and clinical guidelines, 7,21,24 prescribed lists of cause of death criteria, 26 and International Classification of Disease, 10th Edition (ICD-10) codes, 23,29,34 or reference articles, 13 helped guide cause of death assignment; however, in some studies, criteria to reach the diagnosis was left to the physician's discretion. 28 Computer-programmed combinations of signs and symptoms (computer algorithms) of the fatal illness were also used to assign neonatal causes of death.…”

“…This may be due, in part, to varying definitions for the reference standards. For example, Kalter et al 15 developed reference standard criteria based on physicians' diagnoses combined with objective illness signs and laboratory and radiologic findings, whereas Setal et al 29 provided physicians with ICD-10 codes to identify reference standard cases. Table 2 illustrates the variations in the reference standard diagnoses used for the three primary causes of neonatal deathFsepsis, prematurity, and birth asphyxia.…”

“…The open history was reviewed using a 'sign and symptom duration matrix' to guide the analysis, and diagnoses made from the closed-ended modules were based on predetermined field case definitions or algorithms. 7 Setal et al 29 employed Current methods and challenges of neonatal VA N Thatte et al physician review of the closed-ended modules and assignment of diagnoses based on current ICD-10 codes.…”

“…38 Lack of validated algorithms Stillbirths account for an additional, estimated 3.2 million deaths globally each year, 39 however, there have been no well-validated algorithms to identify stillbirths from a VA interview, let alone specific causes of stillbirths. Setal et al 29 included stillbirths in Current methods and challenges of neonatal VA N Thatte et al their validation study; however, VA underestimated the stillbirth rate in this setting with relatively low sensitivity but good specificity (61 and 84% respectively). Iriya et al also included stillbirths in their validation study and reported 80% sensitivity and 76% specificity, 14 illustrating the need for additional research on stillbirth modules for use in VA.…”