Meet a self-admitted nerd who uses an all-the-data, all-the-time approach to take the risk out of insuring people in risky jobs.
Some day you might look way, way up at a tiny figure working hundreds of feet above the ground on a high-voltage transmission line and wonder: Could that guy’s cholesterol levels possibly have much influence on his life expectancy?
Melissa Kuecker ponders unlikely questions like that all day long. As the Senior Life Insurance Underwriter for The Union Labor Life Insurance Company, a subsidiary of Ullico Inc.—a company that insures union labor members whose work can make mainstream carriers nervous—it’s her job to size up risk from a distinctly different vantage point.
Some on-the-job risks may seem downright death defying—at least to us deskbound types—but they’re all in a day’s work to that guy on the transmission tower, and Melissa really wants him to have life insurance. Union Labor Life looks past the perils of the job so that the workers who build and maintain our infrastructure and keep our communities safe can keep their own families financially secure, too. She cares very much about things like cholesterol levels and the conditions they may signal, as well as their severity and nuance, because that makes all the difference in issuing more policies to this cohort. That’s why Union Labor Life begins its supplemental life underwriting procedure with Irix Risk Score and treats Prescription Data and Medical Data as foundational—an approach that allows Melissa to quickly triage applicants while also satisfying her inner data nerd.
Some underwriters are still a bit ill at ease with predictive models for risk assessment, but Melissa is completely at home with them and uses Risk Score like a champ. We were intrigued by her fearless use of this insurtech tool and Union Labor Life’s perspective on risk, so we pressed her with some questions that may give confidence to others, too.
Given Union Labor Life’s union heritage, is there a sense of solidarity—that the company is looking out for its union brothers and sisters?
Absolutely. Our life insurance products are specialized for union members. We do not discriminate against them because of their tough jobs. We know we are taking on that risk. Union Labor Life was founded because we are passionate about getting coverage for union members who put themselves at risk and do hard work that matters. That has been our number one priority for almost a century. These families deserve financial protection, and we help them get it.
Underwriting has traditionally been focused on understanding causality. For example, smoking isn’t merely associated with higher mortality; we understand the causal mechanism behind many of smoking’s harms. That causality is blurred when using predictive models. Did you have to overcome any of your own skepticism when it came to trusting a predictive model?
It was easier for me because I evaluated social security disability claims before I became an underwriter, so I have reviewed a lot of medical records and have been exposed to multitudes of medical conditions. With that experience, I was already comfortable with the billing, procedure, and diagnosis codes that are important inputs into Risk Score.
Since that was a language that I already spoke, I loved being able to sit down with the Irix team to go through all of the medical codes and time periods and assign the risk higher or lower based on the knowledge that the IRIX team had and my own. That eased any fears I might’ve had. Now, by the time I start underwriting a case, I’ve already got a score from the model, but I can also go into the Rules Engine and look at all the yellow rules to make sure the system’s working the way I want it to work—that it maps to our underwriting and reinsurance guidelines, and that there’s nothing in the data that would make me change the decision.
Traditionally, underwriters started from an assumption of best class; no matter what new information came in, that applicant’s insurability could stay the same but if it changed, it only got worse. A model like Risk Score starts from an assumption of average mortality; the data that comes in can move the score up or down. Have you found that using the model allows you to insure some people who would previously have been declined?
Yes. We had a blanket medical question on our simplified issue business. It was only one question, but it listed several conditions like diabetes, uncontrolled high blood pressure, kidney disease; it went through almost every body system and if an applicant answered yes to any of that, they were declined.
Now, even if we get a yes to that question, we still run them through the model. If that comes up red, we decline them; but if it comes up yellow or green, I review the medical data. I know there’s something to find in there—they answered yes for a reason—but often I find they have diabetes or hypertension that’s well controlled. The last time I checked, I’d had 52 applications where they answered yes to the medical question, but the Risk Score was okay; I wrote 41 of them.
The sheer volume of data available on many applicants can be intimidating, which is one reason many underwriters prefer the clarity of a score, or the repeatability of rules engine calibrated to their guidelines. You use those tools but seem happy to get into those weeds, too.
The more data, the better. When I’m underwriting, I love how easy it is to sort through the medical and prescription information by date, frequency, and by claim code count. It is a game changer and makes the more persistent medical conditions stand out easily. Each medical provider is listed and that makes the process of obtaining an APS very efficient.
Is knowing what you’re looking for useful, even in times when you need additional information such as EHR or an APS?
Yes, before the implementation of IRIX I would ask a member for their primary care doctor information, order an APS, and find that they’d accidentally given you the name of their eye doctor. Because we order Medical Data on everyone, I can go through and see, they’ve seen a gastroenterologist, they’ve seen a cardiologist, they’ve been to the hospital; if the cardiac condition is the one I’m concerned about, I know exactly what to order.
I will say when an applicant has a yellow Irix and I order an APS, I check back to see whether I was leaning one way or the other and then ask myself, did the APS change my decision? It’s nice to have a little bit of reflection on that. And of course, sometimes you get the APS and think, “We dodged a bullet there,” but even when I realize that all the relevant information was available from the start, it’s still reassuring.
Your approach seems very methodical and analytical, but it almost feels as if you’re going after some deeper truth about the applicant—almost the way an investigative reporter or detective might.
Sometimes it feels like that. I have had applicants who report that they have never been arrested but when I see their motor vehicle records, I see that they have had DUIs, their license has been revoked, or they have even spent a stint in jail. That is why I feel it is important to get all their information together to create the entire picture and then evaluate the risk.
We’re not scared off by the simple fact that an applicant’s job calls for a hard hat and steel-toed boots, safety harness, or even body armor. We get it; smart underwriting is our protective gear. We want to look at all the factors with the goal of getting them protection, while also making sure that we are protecting ourselves as a company and making the appropriate determinations.
