Using already-solved cases of a mass disaster event for prioritizing the search among remaining victims: a Bayesian approach

Abstract

This work presents a new method for assisting in the identification process of missing persons in several contexts, such as enforced disappearances. We apply a Bayesian technique to incorporate non-genetic variables in the construction of prior information. in that way, we can learn from the already-solved cases of a particular mass event of death, and use that information to guide the search among remaining victims. this paper describes a particular application to the proposed method to the identification of human remains of the so-called disappeared during the last dictatorship in Argentina, which lasted from 1976 until 1983. Potential applications of the techniques presented hereby, however, are much wider. the central idea of our work is to take advantage of the already-solved cases within a certain event to use the gathered knowledge to assist in the investigation process, enabling the construction of prioritized rankings of victims that could correspond to each certain unidentified human remains.The process of identification that guides searches in contexts such as disaster victim identification (DVI), missing person identification (MPI), migration and other situations of violence (OSV) requires the collection of background information from different sources (e.g. legal courts documents, testimonies from survivors, witnesses and families of the missing) 1 . The identification process is essential not only for the sake of Justice and for humanitarian reasons 2 but also to offer answers to victims' families and friends 3-6 . The process of identification usually includes both, (i) the construction of hypotheses of identity from the analysis of such background information that needs to be evaluated at a later stage through genetic evidence, and (ii) the validation of the information gathered from a genetic DNA-led process through the comparison of the ante-mortem and post-mortem information. It is our aim in this paper to describe a general method which could contribute to the investigation process by taking advantage of the already-solved cases of a particular mass death event, to use that elicited knowledge for guiding new searches of related unidentified human remains (UHR). Whenever a pattern does exist within the already-solved cases, the method presented here allows us to make predictions in the identification process of the cases still unsolved, and it also makes it possible to minimize any bias from the researcher. Predictions are understood as the act of prioritizing some individuals over others to be more likely related to certain UHR within the same event. The available information is: (i) information regarding the context of the mass death event, such as date and place in which the event has occurred and the total number of victims, (ii) a database with information of reported victims who are potential candidates to correspond with a set of UHR. That database also includes non-genetic variables amenable to be modeled mathematically in the search for patterns, and (iii) information of the set of alr...

Making decisions in missing person identification cases with low statistical power

“…Finally, in the case of BNDG, baby theft or child trafficking, information such as age, geography and birth date, could complement the DNA-based identification. Methods for formalizing and incorporating this type of data in database searches have been suggested [24,25].…”

Prioritising family members for genotyping in missing person cases: A general approach combining the statistical power of exclusion and inclusion

“…Large-scale cases of the same kind from other parts of the world include those reported in [5,10]. Also related to our work is the recent paper by Caridi et al [11], using Bayesian methods to prioritise identifications among victims of the Argentinian military dictatorship. A recent phenomenon in the search for the Argentinian grandchildren is the emergence of missing great grandchildren.…”

The search process: Integrating the investigation and identification of missing and unidentified persons

“…), hot spots maps (key points in migration routes, places of executions, human remains deposits, etc.). Mathematical models enable predictions or prioritization of some persons more likely to be connected to specific events for identification purposes [ 5 , 18 , 19 ]. The goal is not only to reach an identification but rather a timely one that reduces the missing person's exposure to risk of harm, while also ensuring their prompt return or communication with family to reduce their suffering and uncertainty.…”