The flowers are blooming. The mad rush of lost ID cards and questions about new benefits has faded into distant memory. That means that it's time to start planning for next year! Your insurance carriers are calling you to schedule a lengthy meeting to review your annual data from the prior year and attempting sell you additional products and offerings.
What the data actually tells you
Like many of you, my world this time of the year revolves around data, data and more data. I have stacks of presentations, spreadsheets of summaries and statistical models galore. But what does the data actually tell you? What actionable items can be found?
Maybe it's just me, but it seems everywhere I go lately people are talking about data. I recently heard Dr. David Newman speak at the 2011 Employee Health Care Conference. He's one of the driving forces behind thennt.com, a website that looks at scientific studies and tries to scientifically determine what helps or hurts patients, or is yet to be determined.
Some of the items he highlighted that were determined to have no benefit were the use of antibiotics for ear infections and breast cancer screenings to prevent death by breast cancer.
There were also items that were unclear about whether the benefit outweighs the harm, such as the use of statins for the prevention of heart disease. And items that shockingly cause more harm than good, such as the PSA test to screen for prostate cancer. Many of these topics are heavily promoted as part of employer wellness programs, but the research is clear about the lack of efficacy of these treatments.
Slicing and dicing your data
In another data-centric talk presented by Mercer, an analyst spoke about looking at carrier data in a different format and focusing on four areas: high-cost and high-use, high-cost but low-use, low-cost but high-use, and low-cost and low-use. In analyzing your data in this format, you can see the areas that you need to focus on.
For example, high-cost but low-use might be your catastrophic claimants. There are probably individuals in that population that have rare conditions. Intervention for this group is likely to be centered on case management and determining the appropriate stage of intervention. In contrast, the high-cost and high-use quadrant is likely to include those chronic conditions that would best be served by a condition management program, incentives or alternate plan design, based on the conditions that are part of that grouping.
As I started delving into my data, my management team also wanted to examine certain trends: demographics, cost variations by geographic region, work locations, specific conditions and several other items that border on the impossible.
The underlying questions in all of this were: What drives our costs? Where can we impact the overall health of employees and their dependents? Where can we impact cost?
Correlation does not equal causation
After looking at endless cuts of data, mapping them out by trying to group them via a MECE (Mutually Exclusive Collectively Exhaustive) analysis and trying to find causes for each of the items, I'm coming to the conclusion that not enough data is captured for us to be able to make sound conclusions.
For example, how can we easily run a report on the week a woman delivers a baby at various hospitals and then take that data and see if it really has an impact on the health of the baby and the mother?
Even in the minuscule number of cases where we can find a correlation between two sets of data points, I am reminded that correlation does not equal causation, particularly in health care. There are multiple factors that impact health.
If employers plan to continue offering health care, then we need to take an active role in truly understanding data and partnering more effectively with the research community.
We need to stop having our data held hostage and start making our own demands about what we want to see reported. How will we ever be able to impact cost, quality, and efficacy if we don't have a clear picture of those elements in our own employee populations?
Contributing Editor Shana Sweeney - a self-proclaimed geek and political junkie - is a benefits professional at Google. She is an SPHR with degrees in politics and human resources. She has more than a decade of experience working in various industries, including high-tech, utilities, manufacturing and health insurance. She can be reached at
