George Furlong – Senior Vice President, Managed Care Program Outcomes Analysis – Sedgwick
Stephany A. Rockwell – Risk Manager, Risk Management Finance – JBS
As workers’ compensation program success becomes increasingly dependent on how data can be used to take action, this session at the 2018 National Workers’ Compensation and Disability Conference takes a pragmatic look at available decision making tools.
Big data is about combining multiple, disparate data sets into integrated and usable information. This allows identification of patterns, breaking down silos, and mining through large amounts of data allowing for better decision making and validating impacts of change. While there is much to learn and predict from the breadth of data collection, humans must factor in business decisions.
Sources of data that can provide better predictive information and processes improvements for risk management decisions and claims trending include:
- Risk management and claims management systems
- Human Resources information systems
- Telephonic case management data
- Field case management data
- Utilization review data
- Medical Payment network information
- Pharmacy network information
- Evidence based disability and medical information
Predictive Modeling utilizes these compilations of big data to identify patterns and can provide historical modeling based on predetermined variables affecting a projected outcome or problem. It is critical to maintain the human review of these models to effectively adapt business accordingly.
AI, or Artificial Intelligence then develops pattern matching through machine learning and deep learning. In practical application, this can help identify points in claims management where early intervention practices can improve the overall outcome of the entire experience. Specialized AI provides an opportunity for the business to take a specific problem to the data sets to help identify where improvements and solutions can be effective.
Robotics Process Automation (bots) replicate human actions on the computer. They work particularly well on repetitive processes like adjudication, medical review, etc. The are good for processes difficult to custom code, use with vendor systems that cannot be modified, and in processes crossing multiple systems that cannot be easily integrated.
When leveraging AI in practice, identify data sources, requirements and organization needs, understand the problems you want AI to solve, and employ the human element: data scientist, business, prescriptive modeling. These are particularly great tools to ensure consistency in adjustor onboarding and continued role support.
Leveraging technology in practice optimizes decision-making through large loss modeling, litigation modeling, and clinical modeling. What are the driving factors of risk being identified and how can you use this information to support changes in the claims experience? Another practical application of AI are engaging chat bots. For example, website traffic could largely be injured workers looking for more information. Implementing chat bots relieves strain on other resources, directs person to relevant info, and also collects valuable data directly from employees in the claims process that can improve modeling and processes for future claimants.
Taking time to better understand the multiple AI tools available, how they can influence process and decision-making, and navigating through any challenges to develop a sound AI strategy can be invaluable in managing the costs and experience of the claims process.