By Jeremy F. Heinnickel, Esq., Saul Ewing
The insurance industry is no stranger to collecting and analyzing data. Long before the internet (and even computers), underwriters routinely used historical data for risk selection and pricing. What has changed in the past several years is the dramatic increase in the amount and types of data available to insurers. Each day, massive stores of real-time data are being created from sources like social media, internet browsing histories, mobile devices, and cloud computing platforms. This information boom is often referred to as “big data.”
This large influx of data is of little use if it cannot be effectively organized and analyzed. Analytics is often used to achieve these goals. Data analytics is the science of drawing insights from sources of raw information. One subset of data analytics, predictive analytics, is particularly useful for insurers. Predictive analytics uses data, algorithms, and machine learning techniques to identify the likelihood of future outcomes based on historical data.
While underwriting is often cited as the main beneficiary of the big data revolution, it also has tremendous potential in claims handling and management. Predictive analytics can be particularly helpful to claims handlers because it can assist them in identifying patterns in historical data that can be applied to current or future claims. Identifying patterns in claims and losses can be useful in many different areas. Below are just a few examples:
- Fraud Detection - Fraud detection is currently the most prevalent use of big data and analytics. Using data analytics and artificial intelligence, an insurance carrier can more rapidly and consistently identify suspicious claims that may not have been identified by looking at a single claim file. Potentially suspicious claims are often then referred to an insurer’s special investigations unit (SIU) for further analysis.
- Claims Triage / Claims Management – Resource allocation is a challenge for every claims department. Many insurers utilize data analytics to assign scores and/or rank claims based on risk, allowing them to focus additional resources on high-risk claims. This saves time and resources internally while increasing claim-processing efficiency to improve customer satisfaction.
- Setting Reserves / Settlement – Using predictive analytics, an insurer can compare factors associated with new and pending claims against those of past losses. Analysis of the values of past claims fitting similar fact patterns can help an insurer assess the appropriate reserve and settlement values for current losses.
- Rapid Payment Situations - In emergency situations, insurance companies are faced with a much higher volume of claims, and often decisions on these claims need to be made quickly. In these situations, claims are vulnerable to overpayment. Using big data and analytics, claims handlers can more quickly evaluate claims, assess the appropriate value based on historical data, and identify potentially fraudulent claims in time-pressured situations.
While the use of big data and data analytics in claims has great potential, there are several hurdles to overcome in implementing these tools. One of the most significant barriers to adoption in the insurance industry is cost. Establishing systems to collect, store, and protect data can come at a significant expense. Larger carriers can spread the fixed costs of developing/purchasing these systems over thousands of claims. Smaller carriers with a lower volume of claims may have more difficulty in justifying the costs.
The other factor that is most often cited as a barrier to innovation in the insurance industry is regulation and legal concerns. Privacy issues are paramount any time data is being collected and used (especially from consumers). Numerous federal laws that potentially apply to the collection, storage and use of data, including the Health Insurance Portability and Accountability Act (HIPAA), the Gramm-Leach-Bliley Act, the Children’s Online Privacy Protection Act, and the Fair Credit Reporting Act. Many states have also adopted laws to protect privacy, with California being a leader in this area.
Cybersecurity is also a significant consideration for any company storing and/or transferring data. Many states have laws that establish requirements in the event of a security breach. For example, Pennsylvania’s Breach of Personal Information Notification Act applies to any “entity that maintains, stores or manages computerized data that includes personal information” on a Pennsylvania resident and requires the collecting entity to provide notification to those residents who are affected by a security breach of the computerized data. Furthermore, while not yet adopted in many states, the National Association of Insurance Commissioners recently adopted the Insurance Data and Security Model Law, which establishes data security standards and standards for investigation of and notification to the insurance commissioner of cybersecurity events. A similar cybersecurity law is already in effect in New York.
Finally, over-reliance on data analytics and other forecasting tools can be problematic. For example, while an automated fraud screen process may be able to identify the indication of fraud in a claim, many companies refer the claim for review by the SIU before it is denied. As with other forecasting tools, predictive analytics indicate probabilities, not certainties. The safest approach is to use these tools to assist but not supplant a claim handler’s thought process.
This article was previously published in the PAMIC Pulse.
Jeremy Heinnickel is an attorney from the Harrisburg office of Saul Ewing Arnstein & Lehr LLP. He devotes his entire practice to advising insurance companies and producers. He has extensive experience counseling insurance clients on regulatory, corporate, and transactional matters. He also litigates insurance-related cases in state and federal courts, and before regulatory agencies. Over the past several years, he has focused on the impact of technology on the insurance industry, and particularly InsurTech.