Ian R asked a really good question yesterday about the economics of prevention.
Isn’t it significantly less of a money loser to pay out for PrEP than to pay the various claims of an HIV-positive patient?
This is a common thought and it is one that is highly dependent on the answers to a couple of questions. Prevention is not always net cost savings. Let’s work through some of the questions to see some of the considerations that go into the economics of prevention.
Are we analyzing cost savings from the point of view of society in general?
Are we analyzing costs from the point of view of the current payer?
These are two very different questions. Social costs allow us to think long term and value benefits ten or fifteen years down the road even if the costs are happening in this budget cycle. Social costs and benefits allow us to consider costs that can not be captured by the payer such as improved quality of life and less stress or higher work productivity due to better health.
If we are only looking at the current payer and we are looking at an intervention that occurs on non-Medicare covered people, we need to model churn. Odds are fairly high that an avoided medical event that would have happened next week will be captured by the current payer. Odds are very low that an avoided medical event that would have happened in ten years will be captured by the current payer.
What costs and benefits are included?
What are we valuing? A full societal cost accounting of the benefits and harms of smoking reduction programs can show immediate health improvements but significant pension and elderly healthcare cost increases if the national old age programs have inherent founders’ debt and are on a pay as you go system. Longer life spans lead to longer pension collection spans and longer (and perhaps lower annual) medical spending trajectories. That might be relevant. It might not be.
How many people are needed to be treated to get a benefit?
How many people needed to be treated to get a incremental benefit? What is the Number Needed to Treat (NNT)?
There are very few things where one preventative intervention will lead to one avoided event. The most applicable interventions like steroid inhalers to prevent asthma hospitalizations need a half dozen or more interventions to produce one positive event. Cheap interventions like aspirin to minimize cardiac events needs over a thousand people to take baby aspirin to avoid one cardiac event.
The cost-benefit calculation for a cost saving intervention is:
NNT*Cost per Intervention <= Cost of a single marginal avoided event
Targeted screenings and interventions aim to reduce the effective NNT.
When do the benefits occur and at what discount rate?
Flu shots provide almost immediate benefits. The value of the future is almost as high as the value of the present.
HPV vaccines provide benefits that don’t kick in for many years. The value of that future is less than the value of the present but there are lots of ways to place differing valuations on the future. We could apply the US Treasury long term rate or we could apply an administrative rate. Small changes in the chosen discount rate can lead to widely different present value projections.
Do costs always matter?
Some prevention efforts may lead to higher net costs (however we have determined them). And that can be okay if we think that the improved health is worthwhile. Prevention is mainly about preventing bad outcomes and not about saving money. Saving money sometimes is an intended goal but the goal of improving mortality or quality of life even at a net increase in expense is also a legitimate goal.
Prevention should be judged on what it is seeking to be doing instead of expecting it to prevent disease and other negative outcomes while also saving money.