What role can AI play in workers' compensation?
Suppose workers’ compensation carriers had more widely adopted artificial intelligence during the early stages of the opioid crisis. How might it have been used to mitigate the extent of it and, in effect, save lives?
Artificial Intelligence
Artificial intelligence (AI) is the simulation of human intelligence in machines programmed to think like humans and mimic their actions. This technology encompasses learning, reasoning, problem-solving, perception, and language understanding capabilities. Its capabilities include:
Learning: AI systems can learn and adapt to new information or stimuli from their environment.Problem-Solving: Through algorithms and decision-making capabilities, AI can solve problems, often optimizing for specified objectives.Automation: AI enables automation of various tasks and processes, reducing human intervention.Data Analysis: AI can analyze and interpret vast datasets, deriving insights that can inform decision-making.Natural Language Processing: AI enables interaction and communication by understanding and generating human language.Pattern Recognition: AI can identify patterns and anomalies in various forms of data.
AI in workers’ compensation
With its multifaceted capabilities, AI is significantly reshaping the workers’ compensation industry today, addressing its inherent challenges, and enhancing its operational, clinical and customer service elements. By intertwining intelligent technology with human expertise, the industry is optimizing its processes, improving its service delivery, and ensuring better outcomes for injured workers navigating their recovery journey. Examples include:
Pricing and underwriting are affected, employing machine learning models to develop more accurate, dynamic pricing for policies, considering numerous variables that might influence risk. Implementing AI algorithms to automate the underwriting process, analyzing data points to evaluate risk, and determining policy terms efficiently are also utilized.
In risk management, AI can be implemented to analyze historical data and identify patterns or predictors of incidents to enhance preventative measures and leverage data to provide insights and recommendations for improving workplace safety and minimizing the likelihood of injuries.
Operating efficiency is impacted by leveraging AI to enhance operational workflows, ensuring tasks are completed optimally and resources are allocated effectively. Other examples include implementing AI to automate data entry tasks, minimizing manual errors and improving data accuracy, as well as utilizing AI for intelligent document management, such as auto-categorizing, summarizing and retrieving documents.
Increasingly positive claims outcomes can be facilitated and influenced by an ecosystem of relevant technologies which converge to improve recovery predictions, provide better program administration and communication, and support the continuous monitoring and reporting to benefit the injured worker, the employer, and the insuring organization.
Assisting carriers with the opioid crisis
While the use of AI is becoming prevalent today, focusing on AI-specific use cases like the opioid epidemic and preventing opioid misuse can lead to significant changes in workers’ compensation claims. The opioid crisis has impacted many facets of our lives in the last 20 years. In the 1990s and 2000s, tools like artificial intelligence were in their infancy. As AI improved, along with technology to automate workflow and better integration of data across systems, adjusters have been able to become more strategic in their approaches to proactively manage workers’ compensation claims and improve patient outcomes Here are just some of the ways AI can assist carriers in managing the complexities of the opioid crisis.
Data analytics for identifying patterns
Risk Identification: AI can analyze data to identify workers at risk of developing opioid dependency based on prescription patterns, duration of medication, and other relevant variables. Whereas opioids play a crucial part in pain management and many workers’ injuries are painful, comparisons of prescription patterns for similar injuries could be monitored. Anomalies such as the size of prescriptions and the number of renewals prescribed could be identified as important.
Pattern Recognition: Identifying patterns in prescribing opioids, such as specific doctors prescribing significantly more than peers or specific cases receiving high dosages, can highlight potential issues. One of the most potent uses of artificial intelligence in claims analysis is comparing the quality and type of care workers receive for similar injuries from different doctors and clinics. Based solely on outcomes measured by time out of work, number of treatments provided, total treatment cost, and other factors, algorithms can assign grades to doctors and clinics to treat similar injuries. Identifying claims that significantly exceed averages leads to giving D and E grades. Carriers can then evaluate these providers’ commonalities compared to those with A and B grades. Opioid utilization would likely have been one differentiating factor.
Enhanced claims management
Predictive Analytics: AI can help predict which workers might require long-term opioid use or might be at risk of developing a dependency based on various factors, including the nature of the injury, prescribed opioids, and past medical history. A form of artificial intelligence, predictive models can be built to forecast a claim’s likely duration and expected cost at the onset of a claim and from the information gathered early in it. Algorithm-driven scores for the likelihood of opioid utilization can also be developed. Claims that moved down a path to longer duration and higher cost can be flagged for investigation, as can claims that involve opioids where the models did not expect their use.
Automation: A critical part of deploying artificial intelligence systems in a claims operation is their automation. Scores and alerts must be calculated every time new data on a claim becomes available. Adjusters typically have over a hundred open claims they are working on and cannot possibly screen each one for new data daily. Having algorithms identify which claims need attention and systems that funnel this information to the adjuster are essential to streamlining the claims process, enabling quicker and more efficient delivery of alternative therapies and interventions that could minimize opioid usage.
Reporting: Cloud-based core systems offer visual dashboards reflecting the progress of a claim. This allows managers to react in real time to the inflow of new information which AI is capable of combining in insightful ways. Further, intelligent workflows feed off this data to route workloads according to recommended actions or based on a variety of factors to ensure capacity alignment and overall efficiency. Insurers benefit from deeper level claims functions ranging from predictive risk scores to litigation propensity.
Optimized treatment plans
Personalized Treatment: AI algorithms could tailor pain management and rehabilitation plans for injured workers, considering their health history, injury, and recovery progress. This is not meant to supplant the adjuster’s expertise but to give her a second set of eyes on each claim that creates a reference point.
Alternative Therapy Recommendations: AI could suggest alternative therapies, such as physical therapy or non-opioid pain relief, which could be more effective and have a lower risk of dependency. These recommendations could be based on analyzing their effectiveness in past claims where opioids were not prescribed or were minimally used.
Fraud detection and prevention
Unmasking unethical practices: Using AI’s pattern recognition can help to identify prescribers or pharmacists involved in unscrupulous practices, such as over-prescribing opioids or collaborating in “pill mills.” Prosecuting fraud requires the involvement of law enforcement. Still, at a minimum, prescribers thought to be acting unscrupulously could be flagged and avoided in states where prescribing care is legal. If a worker continues to see a potentially unscrupulous physician, he could be counseled and advised against the danger of becoming opioid dependent.
Ensuring compliance: AI can be used to ensure prescribing practices follow guidelines and identify outliers for further investigation. Doctors and clinics thought to be unethical could be confronted by insurance companies with evidence of their behavior. Reporting their behavior to their state’s medical and legal authorities could be used as a “stick” to get them to uphold accepted standards of care.
Support systems for injured workers
Monitoring Progress: AI can keep a close watch on the worker’s recovery process, flagging any deviation that might indicate a developing dependency. Again, the automation and continuous monitoring of a claim’s progress is essential here. Suppose events occur that are out of pattern with a claim’s current and expected path. In that case, claims staff can be alerted to take appropriate corrective action.
Virtual Assistants: The evolution and acceptance of virtual assistance, empowered by generative AI, will provide support, answer queries, and guide workers through their recovery journey.
Legal and regulatory compliance
Regulatory Adherence: Employing AI may be done to monitor and ensure that prescription practices, claims management, and other processes adhere to regulatory standards and guidelines.
Legal Support: AI can be used to assist in navigating through legal cases related to opioid misuse or malpractice, providing data-driven insights and ensuring documentation integrity. Through natural language processing of structured and unstructured data, AI can organize and analyze the mass of documents that comprise a complex, lengthy claim.
Developing and implementing education programs
Targeted Education: Development of targeted education programs for workers and employers on opioid misuse, alternative treatments, and safe practices can be facilitated by utilizing AI. With the wealth of claim case outcomes that carriers possess, they could deploy AI to profile different claim practices against claim outcomes as measured by the time out of work and the amount of lost wages paid. Many states do not compensate workers for their total wages when they are out of work. Minimizing this loss of income and downtime can become the focus of targeted education that seeks to get the injured worker on a better path to full recovery, minimizing the use of opioids.
Information Dissemination: Employing AI can ensure that information on safe opioid use, risks, and alternative treatments is readily available and easily understandable.
In conclusion, AI is a powerful tool. Along with better technology, enhanced training, and automation, it has the potential to identify early possible opioid abuse and mitigate the long-term impact to injured workers. Through intelligent data analysis, predictive analytics, personalized interventions, and more robust support systems, AI offers a multipronged approach to managing opioid prescriptions more effectively, thereby reducing the risk of dependency and its associated challenges. Combining technology with a compassionate, human-centric approach is pivotal in navigating the nuanced challenges of the opioid crisis, ensuring that workers are supported, safe and healthy through their recovery journey.
