Siegel or no Siegel, the CBO did a fine job

I normally do not comment on Mike Siegel's blog because he has long since lost all credibility with me as a scientist.  He praises any study -- no matter how poorly done -- if it supports his ideological position that he is the one ethical voice in tobacco and trashes any study -- no matter how well done -- if it does not.

His latest commentary, on my praise for the Congressional Budget Office analysis of the effect of a 50 cent tobacco tax increase (indexed to inflation) on federal spending so seriously misrepresents my position that it warrants response.

Here is what Siegel said:

"Had the analysis shown the opposite, I would be willing to bet that Dr. Glantz would have criticized the report rather than embracing it. And I'm sure his main line of attack would be the following: How could the authors possibly have considered as a benefit of this policy the fact that thousands of people would live longer and thus be eligible for Social Security benefits? Furthermore, I'm confident that he would have been quick to point out that this strategy was used by Philip Morris in arguing against a cigarette tax increase in the Czech Republic, and that the company disavowed its own report and apologized after being attacked in the international media.

"Is Dr. Glantz actually suggesting that it is perfectly appropriate to consider as a factor in public health policy analysis the fact that if we save lives, we are going to end up increasing the deficit because people will live long enough to enjoy Social Security and Medicare benefits? Unfortunately, that's precisely the argument that is implied by his embracing of the CBO report and his condoning of this perverse approach to public health policy analysis."

This bit of innuendo conflates two issues:

The first issue is a technical one about how to construct a credible model to project the economic impacts of a tobacco control or other health policy.

The second is a policy question about how long a time horizon is appropriate for decision making and whether something (mammography) is worth doing even if it costs money.

My blog post on the CBO model addressed the first issue.  As I said there, the CBO model is the best long-term model done to date.

In saying that, I did not make any comment on what the appropriate time horizon for policy making should be or say that only health policies that save money should be implemented.  (Neither did the CBO.)  Having a reasonable projection of effects over a long time does not require that one use that time for policy making. 

Just to be clear, here is what I think about the second issue:

• The appropriate time horizon for health policy decision making is 10 years, 20 if you really stretch it.  (That does not mean that one ought not try an develop longer-term projections, but decisions should be based on time horizons over which the projections are most reliable.)  As the CBO notes, the uncertainty in projections grows with time.
• Policies to improve health are worth doing even if they cost money.  If society did not accept this position, we would eliminate almost all medical care, certainly for the elderly or people with serious disabilities.

The reality is that actual policy decisions are usually made with very short time horizons, often just months.  (The current Congressional debate over the fiscal cliff is but one example of such behavior.  Here in California, I used to joke that the long term financial thinkers work with a 2 year time horizon.  Until Jerry Brown became governor, In recent years the long term visionaries seemed to be working more on a 9 month time horizon.)  That is why I have put a lot of energy into estimating the short-term effects of tobacco control policies.  (See, for example, our analyses of the impacts of the California and Arizona tobacco control programs and the immediate impact of smokefree laws on hospital admissions for heart attacks and other disorders).  Urging that 10 to 20 years is the appropriate horizon for health decisions does not make it invalid to do your best to develop longer term models.

It is also important to remember that, in addition to estimating fiscal effects, the CBO estimated how many people's lives would be extended by a 50 cent tobacco tax indexed to inflation:

• By 2021, almost 1.4 million adults would be nonsmokers, including about 10,000 adults who would not have otherwise have survived to that year.
• By 2035, 63,000 additional adults would be alive because of the higher cigarette tax.
• By 2085 more than 3 million adults would be nonsmokers because of the policy, including 200,000 who would otherwise have died earlier.

Siegel forgot to mention that, too.