When measuring benefits’ impacts, outcomes are essential 

As a former benefits leader, I’ve spent plenty of hours looking at population data to better understand trends and areas of opportunity. Looking at our top drivers of spend was typically a good place to start. After I identified an area of focus, I would brainstorm and hypothesize how to solve the problem, implement a program or service (sometimes home grown and sometimes through a third party), and ultimately evaluate its efficacy. In other words, did it have the impact I thought it would?

Understanding the outcomes — and ultimately the ROI or VOI of a program or service — isn’t always straightforward. You can’t just “test” a health plan benefit by offering it to some plan members and not others. In order to measure the cost avoidance (or perceived cost avoidance) of a program or service, you often have to compare outcomes of those who participated in the benefit to those who are their “match” (demographics, conditions, etc.) within the population but who did not participate in the program. For physical health programs, you may have the ability to look at changes in clinical risk factors, respective claims spend, and changes in prescription use in order to assess the ROI of your program. For example, when measuring the effectiveness of a diabetes treatment or program’s impact on A1C, you can look at A1C measurements before and after a benefits implementation to see if the clinical values are moving into a controlled state.

Another example comes from the fertility space, where there are several metrics that can be tracked to prove the efficacy of a program: pregnancy rate, miscarriage rate, singleton and multiple birth rates, and live birth rate. Ideally, you want this data pulled directly from your experiences, not from extrapolated averages based on nationally reported data.  

As we all know, outcomes are important to evaluating ROI; but while most solutions can provide some sort of data, not every data point carries the same weight. This often requires benefits teams and their advisor to dig a bit deeper to get the real story. 

Why specific employee outcomes are difficult to obtain    

Outcomes are key to evaluating a solution, but obtaining valid outcomes can be harder than one might think. When it comes to gathering data, obstacles abound. In the fertility space, clinics report their outcomes to the CDC and SART, but due to the nature of the treatments, these reports are often delayed (e.g., CDC and SART’s data are two years behind). Plus, you need member results and feedback to calculate outcomes, and not all benefit solutions have an ecosystem set up where they receive actual information.  

Another obstacle to data gathering stems from what’s going on behind the scenes. With regard to maternity reporting, there’s a desire to reduce C-section rates, but the way a C-section is coded by the obstetrician can vary to ensure that insurance pays for the operation. So even when you can look at claims data, you aren’t getting the whole story. Many vendors are resigned to receive overall clinic outcomes or make guesses based on claims or self-reported data. 

Adding another layer of complexity is the fact that benefits providers can choose different benchmarks to report, allowing them to show their specific outcomes in the best light. This makes it hard to compare benefits solutions and get an accurate picture. One provider might benchmark against national averages while another uses state averages, or fertility outcomes may note that a “multiples” rate includes twins, while another notes it as triplets or more.   

As you can see, the trouble lies in the fact that it’s often  not clear how the outcomes were calculated, where the data came from, and what it all means. It’s extremely hard to get actual outcomes, and most of the time, outcomes are inferred from claims, estimates, etc., which is essentially a guessing game.  

Finding the right data to get a complete picture 

So how do you know what data to trust? The good news is there are ways to look at data retrospectively to get the whole story. Benefits can be set up to receive actual outcomes, and then researchers can look at outcomes from all those who utilized the solution and then compare that with data from an equivalent population (of the same demographic, age, conditions, etc.) who didn’t go through the program.  

Companies that publicly publish their outcomes, along with their methodology and data sources, are enabling employers to see their impact and evaluate if the benefit can be meaningful for their population. And many take it one step further by having their data and outcomes legitimized by an independent third party. 

As a benefits leader, I always strived to use actual outcomes data when possible and armed myself with specific questions to determine if I had the right data and how that solution could be applied to my population. Here are some preliminary questions to ask a benefits provider: 

• What is the vendor’s network model?  
• How was the data collected? (e.g., claims, member survey, direct from providers, from clinic-level data, extrapolated from nationally reported data)
• Does this data represent the vendor’s entire book of business or is it a smaller sample?  
• How does the vendor calculate the outcomes? What is the calculation behind each metric?   
• Has the data been validated by an independent third party?  
• Does the vendor report outcomes every year? For what years does the vendor report outcomes? What are they benchmarking against?  
• When can the next update of outcomes be expected? 

Ask the right questions, dig deeper, then verify 

Beyond the hard numbers, there are human stories. For the employee, better outcomes mean more success stories. It means that employees are getting the support they need and deserve from their employers. That could mean getting their A1C in range and coming off insulin, or in the case of fertility, bringing a baby home from the hospital. 

  However, to help employees achieve these goals, advisors and employers need to be sure solutions are delivering what is promised. It’s impossible to understand the overall cost and value of a benefit if you can’t first understand the outcomes. Once you obtain the outcomes data required for an evaluation, it’s critical to see if the methodology used to calculate the outcomes has been validated by a reputable independent organization. Vendors have an obligation to provide outcomes data and a clear explanation of their methodology, so you can effectively and confidently evaluate and choose the best solution for the employee population. 

Arielle Bogorad is SVP of Employer Market Strategy at Progyny.