Our chief lesson learned in this arena is that, without validated, well-documented software programs to associate profiles from tens of thousands of remains with scores of direct and indirect reference samples, the matching process is untenable.The midst of a victim identification project is a difficult time to be beta-testing new versions of software.
The archiving and management of the vast amount of data involved in a DNA-based identification of mass disaster victims is an enormous challenge. Because data must be retrieved, compared, and integrated reliably and efficiently, it is crucial to have sophisticated software.
In June 2005, the National Institute of Justice published Mass Fatality Incidents: A Guide for Human Forensic Identification (NCJ 199758). The study that resulted in the guide found that:
Ideally, an experienced IT laboratory staff member should be involved in the management of a mass fatality DNA identification effort. IT should be the cornerstone of quality control, and the IT department should continually be searching for ways to improve work processes and turnaround time.
For example, one way to increase productivity in a mass fatality identification effort is to have the Laboratory Information Management System (LIMS) produce a daily progress report for the media and elected officials. It is preferable to develop this capability before an incident occurs, because it is extremely difficult to achieve this level of IT sophistication in the midst of a mass fatality response.
Exhibit 14 shows different functions that an IT system can support in a DNA laboratory. The arrows follow the basic flow of samples and data. In highly automated laboratories, these procedures will be monitored or controlled through the LIMS.
Regardless of their level of automation, all laboratories employ these systems, in some form or another, during routine casework and offender processing. To effectively support a mass fatality incident response, however, some of these systems require special features.
Prior to 9/11, mass disasters in the United States were relatively small in scale, allowing simple spreadsheet approaches to be sufficient for data management. However, even small-scale disasters require scrupulous data management. Although software programs exist for data management, access, and statistical analyses, the magnitude of the World Trade Center (WTC) disaster demanded enhanced capabilities.
Data-handling systems are needed to integrate any customized software as well as to provide a middleware system for connection and integration of different software components. Computer software must be able to assist with many functions. It must:
In the WTC identification effort, the Office of the Chief Medical Examiner (OCME) contracted with a private vendor that developed software with the above-listed requirements in mind. However, because the software was developed in the midst of the identification effort, and was not previously documented and validated "shelfware," the majority of the kinship analyses and (initially) the remains-matching had to be conducted using several commercially available programs. These programs were supplemented with customized patches developed by Kinship and Data Analysis Panel (KADAP) members who were deployed to the OCME. Without such software, the success of the WTC identification effort—nearly 1,600 identifications made and nearly 20,000 remains profiled—would not have been possible.
Another software tool used in the WTC identification project was from the FBI's Combined DNA Index System (CODIS). Two of CODIS' four files—the Missing Persons and Unidentified Human Remains Index (CODISmp) and the Reference Samples from Personal Items and Family Index—allow the search of DNA profiles. The use of mtDNA profiles as a screening system is facilitated by the introduction of the CODISmp system. Although designed for missing persons, the system may be used to search for DNA profiles of mass disaster victims.
Go to Next Section: Sample Accessioning/Laboratory Information Management System Requirements
Download printer-friendly versions: