Bridging Data and Software in Health Care

How I Combined Self-Learned Technical Skills and Traditional Actuarial Skills

By Shea Parkes, FSA, MAAA

After 10 years and the equivalent of at least three careers at Milliman I now spend my days producing data-focused solutions. I’m a statistically focused “intrapreneur” – an entrepreneur who works within a large organization.

In all of my career phases, my biggest joy has been consistently learning new skills. SOA exams grounded me in the ideas of actuarial credibility and the intricacies of the U.S. health care system. Consulting helped me focus on solving relevant business problems.

As part of my traditional consulting work, I developed solutions that had the added benefit of working for multiple clients. My colleagues and I started with a “copy and paste” approach as we moved solutions from client to client. As we got better at problem-solving with applied statistics, our solutions started living longer and “copy and paste” was no longer adequate. To maintain long-term successful solutions, we realized we needed to demonstrate a deeper understanding of software development and further explore the world of data analytics.

The software development profession encourages self-learning. My personal learning style is to consume a torrent of text. I read a mix of the latest blogs and authoritative textbooks. The blogs provide a picture of modern best practices and pain points, while the textbooks impart a deeper understanding of complex concepts.  I would jump from topic-to-topic as it related to my task or duty at hand – that way I reinforced what I was reading with applied practice.

Around the same time, I began participating in predictive modelling competitions with some of my coworkers.  They proved an invaluable source of learning for all forms of advanced analytics; especially when the top competitors would share nuggets of knowledge after a competition.  Often, it was not just that they used some fancy prediction algorithm, but the power of their disciplined approach to architecting and growing a complete solution that smoothly went from raw data to near-optimal predictions.  We often struggled to iterate responsibly when developing our own solutions.  This caused us to double down on our studies of software development.  I reached a tipping point in my career when we transitioned away from spreadsheets and embraced modern revision control systems for everything we did.

The ability to produce robust, reusable, extensible, testable, maintainable, and automated solutions is invaluable. While I don’t consider myself a software development expert, I can confidently say I know more than the average actuary. I can often view problems and solutions from software developers’ perspectives, work with the tools they use, and offer meaningful contributions to the analytics components of our products.

In particular, the solid data intuition gained during my earlier years as an actuary has been a great asset in the field of software development. Not only can I quickly dismiss some results as incorrect, I can provide suggestions about which stage in an analytics pipeline most likely contains the responsible errors.

We should always be in the practice of learning new techniques and trying new products. If actuaries, in particular, want to compete with data scientists and statisticians, it will require getting our hands dirty. It will also give us real world experience that will steer us into the next paradigm of model building and help us deliver and implement the next valuable solution.

Sharing is caring.
  • Subscribe to our feed
  • Tweet about this post
  • Share this post on Facebook
  • Share this post on Google
  • Share this post on LinkedIn


No responses to "Bridging Data and Software in Health Care"

There are no comments yet, add one below.

Leave a Comment