This paper presents a summary description of NIJ-funded research that determined whether data from cases of suspected elder abuse can be supplied to a computer program that can determine whether financial abuse of the elderly person has occurred ("pure" financial abuse), as well as whether financial abuse may have been accompanied by physical abuse ("hybrid" financial abuse).
The research team used a statewide dataset of 8,800 confirmed cases of elder abuse in Texas that had been investigated by adult protective services agencies. The data were randomly split 80/20. The larger dataset was used to program the computer to detect patterns for financial exploitation and also differentiate between "pure" financial exploitation and "hybrid" financial exploitation. The smaller dataset was used to test the computer models on accuracy in classifying the cases of financial exploitation. The researchers determined that the computer algorithms they developed were reliable in predicting clients who had experienced financial exploitation compared with those who had experienced other forms of elder abuse. The main factors that distinguished the financial exploitation from other abuse types were misuses of financial assets. In distinguishing between "pure" financial exploitation and "hybrid" financial exploitation, the computer algorithms made modest improvements in prediction accuracy compared to chance. The most significant factor that distinguished between "pure" and "hybrid" financial abuse was the presence of an apparent injury, such as bruises and skin tears. The researchers anticipate the data algorithms being transformed into web-based applications.
Downloads
Similar Publications
- PEARR Tool Training and Implementation: Building Awareness of Violence and Human Trafficking in a Hospital System
- FY 2024 Solicitation Overview: Public Safety Officers’ Benefits (PSOB) National Law Enforcement Survivor Support and National Firefighter and First Responder Survivor Support Programs Competitive Grant Solicitation
- Assessment of Sexual Assault Kit (SAK) Evidence Selection Leading to Development of SAK Evidence Machine-Learning Model (SAK-ML Model)