Linked In Improved decision making: Three keys to getting value from data analytics - Sedgwick

Improved decision making: Three keys to getting value from data analytics

In today’s information age, data is everywhere. We’re inundated with it. But gathering data is just one piece of a much larger puzzle. Data isn’t very useful if nothing is done with it. It’s the analysis that leads to using data effectively to make better decisions that is the true key to success.

Workers' compensation is no exception. Increasingly, the success of your workers' compensation program is dependent on how data is being used to take action. If you’re not seeing the value of data analytics, here are three things you need to start seeing a return on your investment.

1. A large enough data set to provide value

Working with an administrator with a vast data warehouse and the ability to analyze hundreds of data fields is critical. Are they able to integrate multiple sources of data together to provide deeper insight? Are you analyzing just your own data or are you able to compare your data with others in your industry? Analyzing your data alone could cause you to miss valuable insights that can be gained by looking outside your organization. With access to the largest integrated data set in the industry, Sedgwick is taking the lead in this area in a way that no one else can.

2. The ability to translate data analysis into actionable strategies

Having access to analyze data is just the first step, but even more important is the ability to use that data to develop a predictive model that results in actionable strategies to drive positive outcomes. This is where many organizations get stuck. They have lots of data but haven’t figured out how to take it to the next level and use it in a way that brings about results.

Working with an administrator that has developed a predictive model will help uncover and identify meaningful patterns that may not be evident and can have a significant effect on the outcome of a claim. It can prevent the tendency to ignore risk variables on claims that appear innocent to the naked eye, missing out on opportunities to change the course of these claims. Analyzing historical claim data driven by machine learning, can lead to a much more accurate way of identifying the variables with a propensity to impact costs. Then effectively using these models to prescribe alternate actions can change the course of a claim.

3. Bringing data analytics and business resources together for added insight

Relying solely on the data to build the model can be misleading, however. Bringing in the business/ claims and clinical expertise together to provide input is integral to the design of the predictive model and can bring a deeper level of understanding. For example, claims that indicate high risk may not actually need a different approach. Managing all claims that hit a risk profile without the business/clinical expertise to provide context could add cost to the system with diluted benefits.

Nurses and examiners, who are educated on the drivers behind the scoring, can provide valuable input on the model that can lead to developing strategies for mitigating future risk and ensure that these claims are handled with the right resources early on to reduce escalation.

To illustrate the dangers of relying on data alone, we can look at a simplified scenario – the outcome for patients with pneumonia with a comorbidity of asthma. The data may show that asthmatic patients with pneumonia historically have good outcomes. Relying solely on the outcomes data could lead to the conclusion that it is not necessary to admit these patients to the hospital since it will drive up costs. However, further discussions with clinical resources brings added insight that these positive outcomes are the direct result of always admitting asthmatics to the hospital. While this may be an oversimplified example, this scenario shows how data alone could lead to inaccurate conclusions in an attempt to mitigate costs, but that ultimately lead to a higher risk of complications and increased costs down the road.

Leveraging the powerful combination of multiple data sources, machine learning / predictive model analysis, and the right experts to provide input on the model make all the difference. It’s what allows better decision making – what we at Sedgwick call decision optimization. A model that has been developed with these three elements can then successfully be used to identify meaningful patterns, prescribe alternate actions that change the course of the claim, and ultimately drive improved outcomes.

Join me this Thursday from 3:00 – 4:00 pm at the National Workers' Compensation and Disability conference for a co-presented session on Improved decision making: The evolution of data, analytics, and technology. Session information:

Stay tuned to the Sedgwick blog for future posts on this important topic.

Back to Blog
Back to top