November 15, 2012
Be a Mathematical Wizard Who Can Explain Your Work to Policymakers
The recent Conversation with the Director was rich in mathematical equations done on the fly with pen and paper, lively discussions about the relevance of healthcare modeling, as well as ideas on HIV prevention.
CDC Director Tom Frieden, MD, MPH, and Stephanie Sansom, PhD, MPP, MPH, lead of the Prevention Modeling and Economics team, Division of HIV/AIDS Prevention (DHAP), share a strong interest and knowledge about programs and interventions to prevent HIV, and get high risk people tested and into care.
Sansom gave credit to her team for analyzing and turning data into economic evaluation models. The team develops healthcare models aimed at shedding light on the cost effectiveness of HIV prevention interventions and making the most efficient allocation of HIV prevention funds.
“The question that we started out answering most of the time – which is, ‘How cost-effective is a particular intervention’? – I think has really started to expand in an interesting way to ‘we have a limited pot of money, how do we best allocate it?’” said Sansom.
“One of the challenges of cost-effectiveness analysis is that people tend not to believe us when we say it is saving money,” said Frieden. “There is lots of confusion between cost saving and cost effectiveness.”
Sansom paused before reflecting, “It is an interesting problem; one we face all the time. And when we hire people onto the team, we sort of let them know it’s not sufficient that you are a great mathematical wizard. You really have to be able to explain what you are doing to policymakers.”
Frieden agreed, adding, “I think your point of avoiding complexity is very important. That the only good models are the models where you can see under the hood and understand what they are saying and how they got to that conclusion.”
Precise Questions + Good Data = Better Models
Sansom recalled that when she started working for CDC in 1998, the use of cost effectiveness to evaluate and prioritize prevention programs was not always popular. But now cost effectiveness and the use of modeling is more widely embraced, and in fact, Sansom believes that sometimes people can be overly optimistic about what models can do. “Sometimes people think, you know, the more vague and elusive the question is, the more we should use a model to address it; when honestly, I think the better your data, the better you are going to be able to use a model. The very big, open-ended questions are very hard to address.”
Sansom shared with Frieden many of the models and analyses the team has and is working on. She brought with her copies of presentations and slides that the director was eager to peruse. “Here, give them to me, and I will give them right back. I will give them right back, I promise!,” he said, bringing Sansom to laughter.
A Peek at How Modelers Work
The director wanted to learn how modelers work, so using CDC data, both experts got pen and paper and started analyzing trends. One of the three equations they developed was on HIV prevention.
“Because the lifetime treatment cost of HIV is so high, most interventions are cost-effective. As you can see from what we worked out here, you don’t have to prevent very much HIV to be cost-effective, and that is where the resource allocation modeling has gotten much more informative.”
The equation below demonstrates the cost effectiveness and cost saving of HIV screening. It shows that with 1,000 people tested, only 8 new cases were identified and only half a case was prevented, but even so, the program was cost saving.
Total Tested: | 1,000 people |
---|
HIV seroprevalence | 0.80 percent |
Cost negative test | $20.00 |
Cost positive test | $85.00 |
Number of positives identified | 8 |
Total testing cost | $20,520 |
---|
When diagnosis otherwise would have occurred | 1 year later |
Increased seroawareness time | 8 person-years |
Annual percent reduction in transmission rate among people aware of their HIV infection | 6.24 percent |
Total reduction in transmissions among people newly aware of HIV diagnosis | 0.50 |
---|
Lifetime cost of an HIV infection | $391,000 |
Savings from reduced transmission | $195,187 |
Net savings (from reduced transmission minus cost of testing) | $174,667 |
---|
CDC experts have developed other models, and Sansom spoke about how results from these models can also show which interventions are more cost-efficient and allow for better resource allocation. The data-driven duo agreed that in an era of limited funding for prevention, the most efficient use of available resources is critical.
CDC Connects Story Manager: María-Belén Moran.