Use a cross-functional approach: Obtain representation and contribution from across the organization, with inputs from Finance, Data, and Clinical/ Healthcare Specialist resources, at a minimum. Clinical/Healthcare Specialists ensure your model is answering key questions, contains accurate assumptions, and reflects current thinking. Your data team, working in collaboration will create efficiencies in balancing what you want to know and what data is available to support this investigation. They can help build for a refreshable model and possibly introduce new ideas to the team based on data possibilities. The finance team will guide calculation accuracy and are usually great partners to help with organizational buy-in and create confidence in the model.

Build for ranges: Create a model framework that invites ranges both as inputs and outputs. The template provided contains both range estimates and is constructed for three case scenarios to help accommodate different organizational approaches. This increases the usability of the model, as ranges communicate a realistic spectrum. By understanding how large the range can be, you innately create context for anyone using the model outputs.

Create in parallel: Again, collaboration is key. Let data analysis help inform increasing specificity for your cost model. As you develop your model, assess the data available to support increased granularity and actionability. Bring in member demographics like age, gender, geography and use historical health insights along with HR data to consider health risks, role type and job type considerations. Health plans might also wish to analyze by industry, geography, plan type, or other business attributes for additional aggregated insights.

Build assumptions as variables: What we know is rapidly changing and evolving. By building assumptions in as variables, rather than use within formulas, you can update your assumptions more easily and completely. Showing the values as variables on the model itself also creates transparency so it is clear what is used in calculations.

Make the model self-documenting: Document within the model itself. Cite and link to sources, show assumptions as variables as shared above, and create a high level of transparency to help answer as many questions as possible. Make it easy to check sources to see if recommendations or assumptions merit updating based on evolving information.

Download the HDMS COVID-19 Cost Model to see and play with a working model. If you are an HDMS client, reach out to your customer experience team for a more detailed and customizable model that works cohesively with your HDMS implementation.