Development of the Verbal Autopsy Instrument for COVID-19 (VAIC)

Abstract

Background Improving accuracy of identification of COVID-19-related deaths is essential to public health surveillance and research. The verbal autopsy, an established strategy involving an interview with a decedent’s caregiver or witness using a semi-structured questionnaire, may improve accurate counting of COVID-19-related deaths. Objective To develop and pilot-test the Verbal Autopsy Instrument for COVID-19 (VAIC) and a death adjudication protocol using it. Methods/Key Results We used a multi-step process to design the VAIC and a protocol for its use. We developed a preliminary version of a verbal autopsy instrument specifically for COVID. We then pilot-tested this instrument by interviewing respondents about the deaths of 15 adults aged ≥65 during the initial COVID-19 surge in New York City. We modified it after the first 5 interviews. We then reviewed the VAIC and clinical information for the 15 deaths and developed a death adjudication process/algorithm to determine whether the underlying cause of death was definitely (40% of these pilot cases), probably (33%), possibly (13%), or unlikely/definitely not (13%) COVID-19-related. We noted differences between the adjudicated cause of death and a death certificate. Conclusions The VAIC and a death adjudication protocol using it may improve accuracy in identifying COVID-19-related deaths. Supplementary Information The online version contains supplementary material available at 10.1007/s11606-021-06842-1.

The Mortality Burden from COVID in Low-Income Settings: Evidence from Verbal Autopsies in India

“…However, the procedure for mapping answers to diagnostic codes had to be modified for a new diagnosis, SARS-CoV-2. The specific modification we employed has not yet been validated – although other modifications to the WHO VA instrument to address COVID have been 24,25 . Nevertheless, assessments of VA modifications to capture SARS-CoV-2 show there is error and diagnosis should be complemented with minimally invasive techniques such as ultrasound when the decedent’s cadaver is available 26 .…”

Accuracy of the Verbal Autopsy questionnaire in the diagnosis of COVID-19 deaths in a Brazilian capital

“…Most studies, however, have reported favorable outcomes for the VA, albeit based on poor quality and low robustness comparisons of data on cause of death, which affects the questionnaire validation 27 . Modifications have therefore been made to improve the accuracy of the VA, such as the creation of a diagnostic stratification system to categorize cases as unlikely, possible, probable, and definite 28 , and a version of the VA that uses a community-based mechanism to identify suspected COVID-19-related deaths at the population level 29 .…”

The impact of the COVID-19 pandemic on healthcare access and utilisation in South Sudan: a cross-sectional mixed methods study

“…Second, to supplement official COVID-19 surveillance, point-in-time questionnaires, patient interviews or other verbal reporting mechanisms may help improve surveillance of COVID-19 transmission in contexts where testing or surveillance is limited, particularly in countries where the national government is coping with significant resource and capacity constraints, such as in South Sudan [56,57]. Assessing relatively rare cases of mortality which may carry stigma through survey instruments in a fragile or conflictaffected context, however carries limitations as well, often leading to underreporting, as found in relation to measurement of maternal mortality in Afghanistan [58].…”