Ante-mortem Data Collection and Victim Identification

Ante-mortem data collection begins with whoever last knew the missing person. In conflict contexts, this is typically a close family member, and the interview is conducted by a trained investigator working from a structured protocol. The INTERPOL DVI AM form provides a standardised structure: physical description (height, weight, build, eye and hair colour), medical history (surgeries, implants, healed fractures, tattoos, dental treatment), last known clothing and footwear, and personal effects carried. The interview also records the family's DNA reference eligibility: biological parents, children, and siblings each provide different reference value depending on which relative is being sought.

The interview itself requires care. Families are frequently traumatised, may be uncertain about details from years or decades earlier, and may have partial or conflicting memories. A good interviewer creates the conditions for accurate recall rather than pressuring for certainty. Leading questions ('Was he wearing a blue jacket?') are avoided in favour of open prompts ('Can you describe what he was wearing?'). The session is documented verbatim where possible and signed by the family member as the formal AM record.

• Medical records: hospital surgical notes, X-rays of healed fractures, dental treatment records, records of prosthetics or implants. These are the highest-quality AM data but often missing in conflict-affected populations.
• Photographs: facial, full-body, and any that show distinguishing features (tattoos, birthmarks, healed wounds). Used for facial approximation comparison and for clothing identification.
• Dental records: charts or radiographs if available; the gold standard when present. Family descriptions of dental appearance (missing teeth, crowns, gaps) are recorded as secondary data.
• DNA reference sample: blood from close biological relatives. The foundation of the DNA identification pathway and the most consistently obtainable AM data type in the absence of records.
• Clothing description: as detailed as possible, including fabric, colour, brand, and any modifications (repairs, monograms, personalisation). Cross-referenced with post-mortem clothing inventory.

The INTERPOL DVI framework uses standardised yellow (AM) and pink (PM) forms so that data recorded by different teams in different countries can be compared in a common structure. The forms mirror each other: the dental section of the AM form has the same layout as the dental section of the PM form, so a comparison is made field-by-field. Differences that are explainable by post-mortem change (tooth loss, soft tissue decomposition) are distinguished from differences that are genuine discrepancies indicating a mismatch.

The reconciliation process involves a team of specialists, typically including an odontologist, a fingerprint examiner, a DNA specialist, and a case manager. Each specialist reviews their discipline's section of the AM/PM comparison and records a conclusion: consistent, inconsistent, or inconclusive. A formal identification requires at least one primary identifier (DNA match, dental match, or fingerprint match) or a combination of secondary identifiers strong enough to meet the applicable legal standard. The case manager compiles the reconciliation report, which becomes the legal identification document.

STR profiling for mass identification follows the same chemistry as routine forensic DNA typing, but the operational context differs in two important ways: the reference samples are biological relatives rather than the missing person themselves, and the skeletal material is often highly degraded, requiring modified extraction and amplification protocols.

Bone samples for DNA extraction are taken from dense cortical bone, typically the femoral shaft, which resists degradation better than cancellous tissue. The sample is powdered under cryogenic conditions to maximise surface area for extraction, then treated with inhibitor-removal reagents that remove humic acids and other soil compounds that block PCR amplification. Mini-STR primers amplify shorter DNA fragments than standard forensic primers, which is important because degraded bone DNA is fragmented, and short fragments are more likely to survive for amplification.

The resulting profile is compared against the family reference database using statistical software that calculates the likelihood ratio for each candidate family reference pair. The LR represents the probability of the observed profile data under the hypothesis that the bone and the reference are biologically related, divided by the probability under the hypothesis that they are unrelated. For identification purposes in mass casualty contexts, likelihood ratios of 10,000 or higher are typically required as a minimum, though organisations often apply higher thresholds in practice.

Clothing and personal effects do not reach the evidential standard of DNA, but they serve two important functions in mass identification. First, they drive prioritisation: when a family has described a specific, unusual garment, and that garment appears in the post-mortem inventory, the case is flagged for immediate DNA comparison rather than waiting in a queue. Second, they provide corroboration: when a DNA identification is confirmed, the presence of described clothing strengthens the overall case and can matter in subsequent legal proceedings about circumstances of death.

Post-mortem clothing processing requires care. Garments are removed in the mortuary under controlled conditions, photographed systematically (front, back, label, detail), described using a standardised vocabulary, and inventoried. The inventory is cross-referenced against the AM clothing descriptions on file. Partially preserved or degraded garments can be compared to family photographs, sometimes allowing recognition of a specific pattern or a tailoring detail that was not standard.

• Manufacturing labels: brand, size, and country of manufacture can narrow the population of potential owners and sometimes date the garment.
• Custom features: repairs, alterations, monograms, and personalised additions recorded by families can be unique enough to assist recognition.
• Associated documents: identity papers, photographs, and letters found with remains can directly indicate identity, though document-based identification is not accepted as a primary identifier under DVI standards because documents can be separated from their owners.
• Jewellery: engraved, personalised, or culturally specific items that families can specifically describe.

Secondary mass graves frequently contain commingled remains: multiple individuals whose skeletal elements are mixed and cannot be assigned to specific individuals by visual inspection alone. Before individual identification can be attempted, the assemblage must be characterised as a whole. The minimum number of individuals (MNI) is calculated from the most frequently occurring non-duplicate element, which is usually the left femoral head in adult assemblages. If there are 47 left femoral heads, MNI is at least 47.

DNA provides a complementary count by generating a set of unique post-mortem profiles from the assemblage. If 120 unique profiles are generated from an assemblage where MNI was 47, the skeletal MNI underestimated the count because the most-common element was not present for all individuals (some were only represented by elements other than the left femoral head). The DNA count gives a better-estimated total, constrained by the assumption that the DNA extraction was sensitive enough to profile all individuals represented.

The scale of the Srebrenica programme is the best-documented demonstration of DNA identification capability in a conflict context. The estimated total killed in the July 1995 Srebrenica events is approximately 8,100 individuals. ICMP's reference database, as of 2023, held over 22,000 blood reference samples from relatives. This exceeds the victim count because many victims had multiple relatives who donated, which actually increased matching confidence.

Over 7,000 unique identifications have been confirmed, using both DNA and non-DNA methods. This represents an identification rate of approximately 88 percent of the estimated victim population. It was achieved across remains found in over 100 separate grave sites, including primary and secondary graves spread across a corridor of approximately 80 kilometres. Many individuals were identified from fragments scattered across multiple sites; the record involves one individual whose remains have been collected from four separate grave sites, up to 50 kilometres apart.

The programme is not complete. Hundreds of cases remain open because remains have not yet been located or because the bone material recovered is too degraded for current profiling methods. Ongoing exhumation and improved laboratory techniques continue to generate new identifications each year.