Selecting the right Verbal Autopsy (VA) instrument is a critical decision that balances international standards with local realities. Since you're already familiar with Civil Registration and Vital Statistics (CRVS) monitoring, we can jump directly into the technical specifics. The primary choice for most programs is between the World Health Organization's 2016 and 2022 VA instruments.

The 2022 version is a significant refinement of the 2016 model. It was redesigned based on extensive field data to be more efficient without sacrificing accuracy. The main goal was to shorten the interview time and reduce the burden on both interviewers and bereaved families.

Regardless of which instrument you choose, its core function is to collect symptom and sign indicators that can be mapped to a cause of death. This mapping is not arbitrary; it must align with the International Classification of Diseases (ICD). Automated diagnostic algorithms, like SmartVA, analyze the pattern of 'yes'/'no' answers to assign a probable cause of death with a corresponding ICD code.

The instrument is designed so that specific combinations of indicators point toward diagnoses in either ICD-10 or the newer ICD-11. For example, a series of questions about fever, headache, stiff neck, and confusion in an adult helps differentiate between meningitis, malaria, and other febrile illnesses, each with a distinct ICD code.

A common challenge is adapting a standardized WHO instrument to a local context. You might need to translate it into a local language or rephrase questions to be culturally appropriate. However, this carries a significant technical risk.

Automated diagnostic algorithms are calibrated to the exact wording and structure of the standard instrument. Changing a question, even slightly, can break the logic of these algorithms. Altering the skip patterns is particularly dangerous, as it can prevent key diagnostic questions from being asked, rendering the collected data unusable for automated analysis.

The VA instrument is not a single questionnaire but a set of modules. The interview starts with general questions, then directs the interviewer to the appropriate module based on the decedent's age and sex.

Each module contains carefully designed skip patterns. For example, in the maternal module, a 'yes' answer to whether the woman was pregnant at the time of death (or died within 42 days of delivery) triggers a detailed set of questions about hemorrhage, eclampsia, and infection. If the answer is 'no,' the entire section is skipped, and the interview proceeds with the standard adult questions.

Similarly, in the neonate module, questions about the baby's condition at birth (e.g., 'Did the baby cry immediately?') are fundamental. The answers guide the interviewer through questions about birth asphyxia, preterm complications, or neonatal infections, while skipping irrelevant sections.

Let's test your understanding of selecting and using VA instruments.

Proper instrument selection and implementation are foundational for generating reliable mortality data. By understanding the technical differences between versions and the critical need for standardization, you can ensure your VA program produces data that is both locally relevant and internationally comparable.