Policy

Valuing Health: How the UCSF School of Nursing Helped One Economist to Thrive

May 2011Wendy Max
PhD

Wendy Max, PhD, is professor of health economics in the Department of Social and Behavioral Sciences and co-director of the Institute for Health & Aging in the School of Nursing at UCSF. Max joined the faculty in 1987. She holds a PhD degree in economics from the University of Colorado at Boulder and a BA degree in history and economics from Stanford University.

Max conducts research on health economics and the impact of illness, and loves teaching and working with the accomplished and insightful students and postdoctoral scholars at UCSF. She has consulted with state and federal agencies, research projects at UCSF, and other universities, private companies and foundations.

Max had the wonderful good fortune to meet mentor extraordinaire Dorothy Rice, ScD, early in her career, and for many years, the two of them collaborated on studies related to the cost of illness and injury. Rice imparted her wisdom on everything from how to present research findings in such a way as to get policymakers’ attention to how to have a career while raising three sons.

Max and Rice’s studies included an investigation of the cost of caring for Alzheimer’s patients living at home and in nursing homes, studies of the lifetime burden of injury, and an evaluation of California’s motorcycle helmet law. These studies helped to refine and improve the methodologies used to conduct cost-related research, while at the same time serving to document the economic impact of illness and injury in a way that also had an impact on policymaking.

Max’s recent research has focused on modeling the economic impact of tobacco on health and health care expenditures. She and her colleagues have made UCSF one of the key tobacco economics research centers in the world. They have estimated national costs of smoking, costs to Medicare and Medicaid, tobacco-related costs in California and in India, the effect of tobacco control efforts on health and health expenditures over a lifetime, and the impact of smoking on communities of color.

Max is proud of the role she played in the 1990s in helping to bring the tobacco industry to its knees. She worked with a number of economists, attorneys and state attorneys general as an expert witness when they sued the tobacco industry to recoup Medicaid costs that the states had incurred as a result of tobacco-related diseases.

She is currently developing models of the impact of secondhand smoke exposure on health care expenditures in California and the United States. She is also modeling the impact of California tobacco control expenditures on health care costs, and is estimating the cost of smoking among those with mental illness.

Max has delighted in working with colleagues who have enticed her to collaborate on interesting projects that vary widely in topic and geography. Accordingly, her other current projects include developing and evaluating an integrated primary health care program in rural Malawi, looking at the economic impact of Hurricane Katrina on the primary health care safety net in New Orleans, determining the cost of hoarding and cluttering behaviors in San Francisco, and analyzing the incidence and cost of pedestrian injuries in San Francisco.

Max loves to travel and has taken advantage of every opportunity to travel for work or play, having visited six continents so far. She loves photography, hiking and listening to books. In addition to her academic achievements, Max’s proudest accomplishments include her three sons, Tayson, Kyler and Zarek, as well as the recent completion of her 1,000th hike around her favorite circuit in the Stanford University foothills. She and her husband, Robert Siegel, live in Palo Alto, where they occasionally run into each other between world travel stints and late nights in the office.

The Thirtieth Helen Nahm Research Lecture – April 23, 2010

The Helen Nahm Research Lecture Award recognizes a UCSF School of Nursing faculty member or graduate who has made an outstanding contribution to nursing science and research. It honors the late Helen Nahm, professor emerita and dean of the UCSF School of Nursing from 1958 to 1969. Nahm passed away in May 1992. In her early retirement, Nahm organized the directors of schools of nursing and health care agencies in the San Francisco area to facilitate the professional development of the new graduate nurse and to improve the quality of nursing care. Her many honors and awards, including honorary fellowship in the American Academy of Nursing in 1978, attest to her leadership in the profession. Through her research findings and numerous articles and addresses, Nahm shared her ideas and vision with a new generation of nurses. It is to exemplify this excellence of scholarship and dedication of purpose that the Helen Nahm Research Lecture Award was instituted. The lectureship was inaugurated in 1981.

 

LECTURE:

Valuing Health: How the UCSF School of Nursing Helped One Economist to Thrive

Wendy Max Thank you very much. Let me say that I’m absolutely thrilled to be here today as the Helen Nahm Lecturer and to see all of my friends and colleagues and family and everyone out here. It’s very, very exciting for me.

I’d like to start out with a question today. What is an economist like me doing at the UCSF School of Nursing? This is no ordinary school of nursing. When I was young, I had a very naïve view of what nursing was, probably from TV and the movies. I thought that nurses were women (of course!) in white dresses who wore funny hats and took care of patients in hospital beds.

But now I know that nursing science is so much more than just that. And actually, some of this is due to the vision of Helen Nahm. It was under her deanship that our School of Nursing added a Department of Social and Behavioral Sciences, and she broadened the vision of nursing education to include a very strong research dimension. It is partly due to this that it’s possible, 50 years later, for a health economist like me to be standing here before you.

What Economics Contributes to Nursing Science

What is it that economics can contribute to nursing science and our understanding of health care? Economics helps us to understand the underlying constraints to what we do or what we would like to be able to do. It also helps us determine what we can do with resources that we have, and how to do things most efficiently and cost-effectively.

A good friend of mine told me that she thinks what I do is to lay bare the economic underpinning of some of our most critical social problems, so that effective policy solutions can be devised. I hope I can live up to that, at least in part. I have focused my research career on understanding the economic implications of illness and injury – really the broader question of how to value health. So let me just say a little bit about my overall career path before I focus in on some of the areas I’ve worked in.

My work has really fallen into three key areas. The first one, including the first project I ever was involved in here, is related to injury. It started with a report to Congress. But after that report, where we estimated the lifetime cost of all injuries, the work continued to grow. We studied head injuries, firearm injuries, and did a later report to Congress on the cost of intimate partner violence. I conducted an economic evaluation of California’s motorcycle helmet law, and I have a current project now, working with colleagues at the San Francisco Injury Center, looking at the cost of pedestrian injuries in San Francisco.

Another area where I’ve done a lot of work relates to aging and chronic illness. The first project I was involved in was a study looking at the cost of both formal and unpaid caregiving costs for Alzheimer’s patients. A number of other studies followed that, looking at costs of osteoporosis and a number of different types of cancer. I have two current projects related to primary health care: one developing an integrated primary health care system in Malawi and another project looking at the primary health care safety net that has been developed, redeveloped really, after Hurricane Katrina in New Orleans.

Another area of my research is tobacco – the economics of tobacco and tobacco control – and I’m going to tell you a lot more about the work I’ve done in that area. Don’t worry. I’m not going to talk about all the different studies. I was going to put arrows up here, but it got to be rather a mess. Suffice it to say that I’ve had the opportunity to work on a number of different projects as an economist here and to be involved in a number of very important health issues that have been tremendously satisfying to work on.

During the rest of my time today, there are two things I’d like to do. First, I will talk about my research on the economics of tobacco over the last 20 years. Then I want to pass along 10 lessons that I’ve learned over those years.

Tobacconomics

So let me start with the economics of tobacco, what I call tobacconomics. I want to show you how the models that we developed to look at these issues have changed and evolved over time, and talk about some of the kinds of conclusions that we can draw from this research, and how I think the work has actually made a very important difference. The work has been a team effort; it’s not anything I could have done by myself.

These are my current colleagues: Xiulan Zhang, who was with us at UCSF for many years and is now the dean of the newly established School of Social Welfare at Beijing Normal University; Dorothy Rice, who got me into the work in the first place; Len Miller, econometrician extraordinaire, who challenged me to work on a number of models that really were written in Greek, not in English; Yanling Shi (who is over there with her daughter, who’s noticing her mother’s picture), our intrepid statistician and data analyst, who willingly reruns analyses 500 times if need be; and finally, Hai-Yen Sung, a fellow economist and the best colleague, collaborator and friend that one could hope to work with. These are the people that I work with on the tobacco projects these days.

How did I get interested in this area? Let me give you a little bit of history. It all really started with Proposition 99, the Tobacco Tax and Health Protection Act that, you may recall, was passed by the voters of California in 1988. What this proposition did was to raise the tax per pack of cigarettes by 25 cents. Of the revenues generated by this tax, 5 percent of them were put into a research account and used to establish a wonderful program, the California Tobacco-Related Disease Research Program, TRDRP for short.

Dorothy, who had recently published a seminal work on the cost of smoking in the United States, said to me, “Now that TRDRP has been established, they’ve put out their first call for research proposals. Why don’t we see if they would be interested in funding some research on the cost of smoking in California?” So we applied during the first cycle, succeeded in getting a grant, and began work on some of these smoking issues in California.

Economic Impact of Tobacco on Health

I will step back for just a minute. What do I mean when I talk about the economic impact of tobacco on health? There are a couple of questions here. Why is it interesting? Well, we’ve known for many, many years that smoking has a negative impact on health, but our understanding of the economic implications of that is much more recent. Being able to quantify the cost of the health effects of smoking has been very, very helpful in terms of shepherding resources to reduce the burden of tobacco-related disease, and in determining which interventions make the most sense. In this era of highly constrained resources, being able to attach a dollar value to the cost of this behavior really presents a very cogent argument for action.

What do we include when we talk about the economic impact of tobacco on health? Economists like to talk about three types of costs. We talk first about direct costs. Those are the dollars that are actually spent for health care, including hospital care, medications, nursing home care and so forth. Beyond the dollars that are actually spent to treat and diagnose people with disease, there are what we call indirect costs, including the value of time lost if someone is unable to carry out their work, their housekeeping, their volunteer activities or their education. There is a real economic cost associated with lost time, and one way to assign a value is to look at forgone earnings and an imputed value for other types of activities. That is a very conservative way to value time, and it is one of the approaches that we’ve been using over the years.

People die prematurely from tobacco-related disease, and again, it’s important to be able to assign a value to these lost lives. And so we look at what people would have done. Again, this gives you a value of their lost productivity, so it’s a conservative way of valuing the lives that are lost. But I think it’s really important to include these indirect costs in any type of a study because these are real economic burdens that result from, in this case, smoking.

Estimating the Costs of Smoking

How do we go about estimating these costs of smoking? For most illnesses and injuries, it is more straightforward. If you want to know the cost of malaria, you can add up all the days that people spend in hospitals and figure out what the cost of those days is. You could then add up all of the other health care services that people use, and then you could look at the days that they lose from activities, the number of lives lost, and you add them all up. But you can’t really do that for smoking because smoking is a behavior. It’s not a diagnosis. Nobody gets admitted to a hospital with a diagnosis of smoking. They might get admitted with a diagnosis of lung cancer or heart disease that results from smoking.

So you have to do some extra analysis here. First, you have to figure out what illnesses are caused by smoking, and then figure out what the cost of those illnesses is. Then finally, you need to figure out what proportion of those costs you can attribute to smoking. So even for something like lung cancer, for which at least 80 percent of cases are probably caused by smoking, you still have to figure out what part of those costs are a result of smoking.

And the methods that we use to determine the proportion that we can attribute to smoking have evolved over time. We have much better data now than we had 20 or 30 years ago. We have models that are much more sophisticated. We have much better computing capability. We can crank out models and, in a few hours or a few minutes, do what used to take days and days with a calculator and paper to accomplish. So we’ve really gotten much better at doing this over time.

We need several types of information to understand the economic impact of smoking. First, we need to know smoking prevalence. We need to know how many people smoke. Then we need to have some way of figuring out how smokers compare to nonsmokers. What is the relative risk of smoking? How much more likely are smokers to die of tobacco-related disease, or to be diagnosed with these diseases, or to incur expenditures than people who don’t smoke? We need all of this information.

The Prevalence of Smoking

Here are some current figures on smoking prevalence. There are a number of surveys that are done that allow us to figure out how many people are current smokers, how many people are former smokers and how much people smoke. Here are some data from the Behavioral Risk Factor Surveillance System, which is a survey that’s conducted by the Centers for Disease Control and Prevention every year in every state. We have good data going back a number of years.

I’ll point out two things about these data here. First of all, you notice that both in the United States, the blue bars, and in California, the green bars, men smoke at higher rates than women. That’s pretty much true everywhere. In fact, in some parts of the world, the extremes are huge. In China, two-thirds of men smoke and 3 percent of women smoke. Here, the differentials between men and women aren’t quite as large.

The other thing I’ll point out is that both for men and women and for all adults, Californians smoke at lower rates than other Americans. We have a very strong tobacco control program here, and I think you see the results of that. In fact, California has the second-lowest smoking prevalence of any state. Do you know what state has the lowest prevalence? (Utah. Very good. Very good. My students never get that.)

I mentioned that the approaches that we use to estimate these costs have evolved over time, and I’ll walk you through how those models have evolved. The earliest approach was what we call the mortality ratio approach, which is basically drawing from epidemiology. What we assume is that the proportion of health expenditures attributed to smoking is the same as the proportion of deaths that can be attributed to smoking. Although we have much more sophisticated data now, this approach still has some application today.

The Mortality Ratio Approach

A number of years ago, I was involved in the tobacco litigation when the attorneys general of 46 states decided to sue the tobacco industry for Medicaid-related costs – a very, very exciting effort to be involved with. My role was to present the costs. This was the damages phase of the case. So after the jury had been convinced that the tobacco industry was responsible for these costs having been incurred, my job was to come in and help them figure out what the value of those damages was, and I actually used this mortality ratio approach.

I remember when I started working with these attorneys, they said, “Well, when you present to a jury, you should assume maybe a ninth-grade education. Most jurors aren’t going to be able to understand a sophisticated econometric model. You want to tell them something that they’ll be able to understand.”

So I was kind of the warm-up act. I would come in and I would walk through this mortality ratio approach, which is fairly intuitive. If 80 percent of deaths from lung cancer are due to smoking, let’s figure that 80 percent of the cost of lung cancer can be attributed to smoking. And the jury would nod their heads and say, “Yeah, that sounds reasonable.” And then another colleague would come in and would say, “Well, of course, we know that there are many more things going on here and we want to control for this, this and this.” He would put up a big, black box and say, “But even when we control for all of these other factors in a very sophisticated econometric effort, we come up with results that are somewhat similar to this.”

So we used the mortality ratio approach to convince them that what we were doing was reasonable. Then we would say, “Yes, of course, we did take into account all these other differences.” And it was very effective. You may recall that the tobacco industry reached a settlement that required them to pay more than $200 billion over 25 years. This was part of that effort.

Another way in which this mortality ratio approach is still used today is in developing countries. Oftentimes, developing countries don’t have a lot of the data that we would like for the models we run here. But I would bet if they have any health-related data, they probably have data on deaths, and they probably can figure at what rates smokers die compared to people who don’t smoke. And in fact, that’s exactly how it’s been used.

We had a postdoc here a few years ago from India who wanted to look at the economic implication of smoking there. We tried to do all sorts of other approaches, and we ended up using this approach because this was the one that we could justify with the data. It’s been used in China and Indonesia and a number of other parts of the world where these are the data that one can get their hands on. So this was the earliest strategy for estimating costs of smoking.

Health Service Utilization Ratios

The next improvement to the methodology was to use health service utilization ratios instead of mortality ratios. This is the work that Dorothy Rice had done looking in the US. Instead of assuming that the proportion of deaths is the same as the proportion of expenditures, we could actually look at health care services used for diseases. We would look at how many days people were hospitalized for lung cancer and heart disease and breast cancer and so forth for smokers, compared to those who didn’t smoke or those who used to smoke.

This is getting closer to the expenditures that we’re really trying to estimate. And it became possible because we have some very good data at the national level and, in some cases, at the state level as well, that allow us to look at these utilization rates. We used the National Health Interview Survey in the early studies.

The CDC developed a computer package called SAMMEC based on this approach. SAMMEC stands for Smoking-Attributable Morbidity, Mortality and Economic Costs. They basically took Dorothy’s work and put it into a computer program, which was then distributed to all the states. For the first time, the states could come up with their own estimates of the cost of smoking. And one thing I’ve learned is that people want data that’s specific to them. They’ll say, “Enough of this national stuff, tell me what’s going on in this district of San Francisco.” Well, it was a big improvement to be able to estimate these costs at the state level, and that’s what the SAMMEC program did. It came with a big notebook and all kinds of documentation, and there were data in there. If you didn’t have the data for your state, there were data you could use instead, and it was widely used.

In fact, Dorothy and I used SAMMEC as the basis of the first study we did, looking at the cost of smoking in California. We actually estimated the cost for each of California’s 58 counties. We produced this wonderful report (that doesn’t photograph very well, but that’s the report), and in this report, in addition to some data for the whole state, we had a two-page spread for each of California’s 58 counties. Each county could go to their pages and they could see what the smoking costs were, smoking prevalence, population and so forth.

It was interesting because at the time, we had to make some assumptions in order to come up with estimates for 58 counties. The samples get kind of small when you look at some of the rural counties that we have in California, and I remember going back and forth with Dorothy. Should we do county estimates? Should we make the necessary assumptions? And Dorothy prevailed. She said, “Yes, we should definitely have estimates for each county.”

And she was so right because we know how often this has been used. We’re still hearing people tell us that “I used your report, and do you have anything more recent?” That’s always the second part of that question. We did this study using SAMMEC. I was the SAMMEC queen. I think at that point, I probably ran SAMMEC more than any other person in the world because we ran it for 58 counties, and for each county, of course, you have to run it a couple of times because it’s never quite right the first time. So I was really an expert on SAMMEC.

Comparing Costs: Smokers and Nonsmokers

We know these approaches aren’t perfect. There are some weaknesses. What we’re trying to do is figure out what the cost of smoking is, what people spend, what health care expenditures are related to smoking. The first model uses deaths as a proxy for expenditures. The second model uses utilization rates as a proxy for expenditures. We had to do that because of the available data. But of course, we’d really like to be able to look at expenditures directly and compare those for smokers and nonsmokers.

And in fact, we now have data that allow us to do that. We also know that smokers differ from nonsmokers in ways other than just their smoking behavior. Smokers are less likely to have health insurance. They’re more likely to engage in other risk behaviors like drinking. You want to have a model that allows you to control for these factors. Then you can say that these differences aren’t due to drinking. They’re due to smoking because we already controlled for their drinking behavior. You want to be able to do that in the models.

Another weakness of the first approach, the mortality ratio approach, is that you only look at certain smoking-related diseases – the diseases that there’s been a lot of research on. But sometimes, smokers have higher costs that don’t necessarily work through disease pathways. As a result of all these weaknesses, we began developing econometric approaches to estimate the costs of smoking.

Len Miller joined our team at that point. We made use of his econometric expertise – econometrics being the application of quantitative or statistical methods to study economic principles. We used these econometric methods to assess economic theories about smoking and health expenditures and health, and the data that we had allowed us to control for many of the other ways in which smokers differ from nonsmokers.

Assessing Economic Theories About Smoking and Health

Here is the conceptual framework that’s behind the econometric models we developed. Smoking leads to smoking-related diseases. People with smoking-related diseases are in poor health, and being in poor health results in incurring health expenditures. In addition to that, though, we know that smoking leads to poor health, not through this mechanism of specific diseases. For example, a smoker with a broken leg might be in the hospital longer than a nonsmoker because they heal more slowly, but broken legs aren’t considered a smoking-caused disease.

And it’s even possible that smoking may lead to health care expenditures without really affecting health. For example, a pregnant woman wouldn’t consider herself to be in poor health or to have a disease. However, a pregnant smoker might have higher costs than a pregnant woman who didn’t smoke. So we allowed for all of these different ways in which smoking can lead to health care expenditures.

We used an excess cost approach. The idea is that you want to compare smokers and nonsmokers who are the same in every way except for their smoking behavior. Then the difference in their health care costs can be attributed to smoking if you controlled for everything else. For example, the excess hospital cost of smoking would be the hospital cost for a smoker minus the hospital cost for someone who never smoked and is the same in every other way. Statistically, you can make them the same in every other way. So that’s the basic strategy behind the econometric models that we have developed.

The early models were primarily descriptive. We quantified the cost of smoking for each of the 50 states; we did it again for each of California’s 58 counties. There was another report that actually looked very similar, also had the two-page spreads for each county, but in fact, it was based on completely different methods. Instead of using the SAMMEC program, which uses utilization ratios, the second report was based on econometric models, which control for many other factors.

And then we extended the research over the years to look at how different groups of people were impacted by smoking because there are different impacts on different subgroups. We looked at Medicaid and Medicare recipients, and this was about the time that the attorneys general were suing the tobacco industry. They were very interested in our Medicaid costs, and that’s how a number of us became involved as expert witnesses working on that litigation.

A Place to Cut Costs

We also extended the model in California. We looked at communities of color: African Americans and Hispanics. Let me show you a few results (I could go on forever, but I won’t). Here are some of our early results looking at the impact of smoking on Medicare and Medicaid programs. And you see that of all Medicaid expenditures, 12 percent of them are a result of smoking-related disease. Of all the Medicare expenditures, 9 percent result from smoking-related disease and illness. And when you look at all health care expenditures, about 12 percent of health care expenditures can be attributed to smoking-related disease.

If you want to look at a place to cut health care costs, you don’t have to look much further than this. And of course, that’s what tobacco control folks have been doing for many, many years.

I mentioned that we did a study looking at the cost of smoking for communities of color. Actually, Ruth Malone convinced me to do this study. We had been thinking about how to extend our model, and Ruth called me up one day and she said, “You know, Valerie Yerger and I have been doing all this research looking at how the tobacco industry ingratiates itself into African American communities, but nobody’s looking at the costs that those communities ultimately bear. The industry provides scholarships and they support events, but what about the negative impact on the community? Somebody needs to do that study and you’re the person who should do it.” And you don’t argue with Ruth.

So we ended up writing up a proposal and getting funding from the Tobacco-Related Disease Research Program and doing just the study that Ruth suggested. It’s been very well received. You were right, Ruth. The people out in the community are very grateful for these results.

Findings in Communities of Color

Here are some of the findings from that study. I guess that what I would point out here is that we look at both medical care and then the value of the lives lost, the mortality costs. For both of those types of costs, the costs per African American smoker are much higher than the costs per anyone else. In fact, the costs per Hispanic smoker are relatively low in California.

We conclude that African Americans bear a disproportionate share of the cost of smoking. They have the highest adult smoking prevalence rates – at about 19 percent. They have the highest total cost of smoking per smoker. They represent 6 percent of the population, but in fact, 10 percent of the cost of smoking, 8 percent of the health care costs and fully 13 percent of the value of the lives lost. So clearly, smoking is posing a disproportionate burden on this community.

For the Hispanic community, the findings were a little bit less clear and took us awhile to figure out how to interpret. In fact, Hispanics at this point have the lowest adult smoking prevalence rates. However, because California has a very large Hispanic population – almost a third of Californians are Hispanic – we’re talking about a lot of smokers. There are almost a million Hispanic smokers in California. One in every four adult smokers in California is Hispanic, and we did see that the mortality cost per death was very high. What that means is that Hispanic smokers are dying at younger ages than the average Californian.

So the way I interpret this is that the potential for high future cost is definitely here, and I think that this is a great time to be involved to prevent more Hispanic kids from taking up smoking. We know that the tobacco industry is very effective at targeting particular groups of people, and they have already done this with brands and marketing strategies that appeal to the Hispanic population. Now is the time to get involved. And even though the impact is less than it is on other communities, we’re still talking about a cost of smoking of $2 billion. So it’s not a small amount here.

Higher Costs for Former Smokers

One pattern that emerged from all these studies that was intriguing was that the costs were higher for former smokers than for current smokers. I remember the first time I saw that, I kind of scratched my head. Does that mean that you shouldn’t quit smoking because your costs will go up? Well, no, I hope it wouldn’t be interpreted that way. What is happening is that people don’t quit smoking until they get really sick. So if you just compare people at a period in time, the former smokers look like their costs are higher.

The real question is, what happens to those people after they quit? You really need to follow them and see what happens to smoking costs for individuals when they change their behavior over their lifetime. And in fact, that led to our next study. We developed a dynamic model of smoking that allowed us to follow people, a cohort of individuals, over their lifetime.

I have to say that this study was really Len Miller’s baby. Len spent years working on these econometric models and making them better and better. At one point, he wanted to buy a new computer. And I remember our computer person saying, “Do you really need a computer with that much power? I think that the computer you’re asking for is about the same kind of computer that the Department of Defense uses to track all troop movements during the course of a year. Do you need that much computing ability?” Len insisted that he did, and we got him what he needed to be able to follow the cohort of people. We followed them for up to 90 years.

And then we used this model to look at the impact of the California Tobacco Control Program – all of the programs that have been developed with the tobacco tax money to reduce smoking in the state – to look at the impact of the first 10 years of the program on the people who were impacted by the program. We found that, in fact, the program resulted in higher health care costs for some people.

When you get people to quit smoking or cut back on smoking, they’re healthier and they live longer. And when people live longer, their health care costs may go up. However, this was more than outweighed by the value of those additional years of life and by the value of being in better health for a number of years.

Tobacco Control Program a Raging Success

When we assigned a value to these years of better health, it was clear that the program had been a raging success. Our estimates were the program had saved something like $22 billion to $100 billion, depending on how you valued life. There was no question that a lot of the value here was in terms of people living longer and people being healthier.

The recent work we’re doing is to take these models and apply them to a new area, the economic impact of secondhand smoke exposure. We have two grants right now, one from TRDRP, which has been a wonderful funder for me over the years, as you see. We’re looking at secondhand smoke exposure in California, and we’re looking at very specific subcommunities within the state.

I’m also part of a center that Neal Benowitz runs that’s funded by the Flight Attendant Medical Research Institute (FAMRI), and we have a study within that center looking at how to model secondhand smoke exposure in the US. What we’ve done is to try to apply the models we’ve been developing all these years for active smoking to secondhand exposure.

Secondhand Smoke and ADHD

The first step was to compare disease rates among those who are exposed and those who are not exposed to secondhand smoke. We controlled for a number of other factors, and a couple of very interesting things came out. We found that among young children, kids who were exposed to secondhand smoke had higher rates of attention deficit hyperactivity disorder and stuttering than those who weren’t exposed. And this is after controlling for prenatal exposure and a number of other things.

We haven’t seen this anywhere in the literature, so we’re quickly writing it up, and hopefully we’ll send this out fairly soon. These are clearly conditions that may impact kids’ ability to learn and to succeed, particularly kids in that age group.

The Contribution of Economics Research

What is the conclusion from 20-something years of doing tobacco economics research? I think the overall conclusion is that tobacco control is a major public health success story. I like to think that economics has played an important role in our understanding here, and has helped us to define the problem and come up with effective interventions.

The economic research has helped to document the impact of smoking – and also secondhand smoke exposure – on health and on health care costs. It has helped us figure out who to target and what kinds of programs to design, not to mention that it helped us to hold the tobacco industry accountable for what they have been doing to the tune of $200 billion, and that’s just for the Medicaid program.

There were other cases that went on at the same time. And I think this is the evidence. If you just look at 25 years of smoking prevalence – the green line is the US, the yellow line is California – you see that smoking prevalence has been going down. It’s been going down in California. Many years it went down more quickly than in the US, but we’re always below the US smoking prevalence rates.

Along with that decline, we have lower rates of disease. We have lower health care costs. We have fewer people being exposed to secondhand smoke. These are all wonderful outcomes of all of this research that has gone on, and economics has made an important contribution.

Let me shift gears now. Enough of the tobacco stuff. Let me pass along to you 10 lessons that I’ve learned along the way.

Lessons Learned

First of all, I’ve learned that the economic impact of illness goes beyond just the health care sector.

It’s so important to include things like the value of time, the value of lives. In our work on Alzheimer’s disease, we looked at the value of the time caregivers spent taking care of people with Alzheimer’s. In our partner violence study, we looked at criminal justice costs. But all of these result from health conditions. So I think you really need to go beyond the health care sector to fully appreciate the impact of illness.

Another lesson learned is that people will misinterpret your work.

Right after we published our Alzheimer’s findings, Dorothy received a letter from a woman (I think in Minnesota) and the woman wrote, “I thought you’d be interested in seeing what my mother’s nursing home administrator is saying about your research.” Apparently, this nursing home administrator had sent out a letter saying, “As you know, we haven’t raised our rates in this nursing home for several years. According to a study that just came out yesterday from the University of California, San Francisco, the average cost of nursing home care is $42,000 a year. Our rates are only $32,000, so as of next month, we’re going to be increasing our rates.”

Now, of course, the cost of nursing home care in California should not be used to justify an increase in Minnesota, and they used the wrong number anyway. Our number included things far beyond just the cost of nursing home care. But the lesson is you’ve really got to try to present things as clearly as possible. You won’t be able to control everything people do with your research, and people will misinterpret your work if you’re not careful.

Lesson three is to present your findings in a user-friendly way.

This is my copy of one of those cost-of-smoking reports. It’s falling apart. It’s dog-eared. I think the first study we did was before everything got put on the web, and so people were constantly calling and saying, “Can you fax me the two pages for Mendocino County?” (I don’t know who we got to bind that book, but if I can remember, I will advise you to never use them. The book just fell apart.)

But as I mentioned, we did go back and forth as to whether we should present estimates for each county, and it turned out to be extremely useful to people who wanted this information. In fact, just a couple of weeks ago, I was at a meeting and a woman from the American Lung Association came up to me and said, “When are you going to update that report?” As soon as it comes out, people want it updated. But it’s very gratifying to know that they use the data.

Lesson four: There are trade-offs between primary and secondary data.

The first study that I was involved in was “The Cost of Injury” study, where we used many different sources of secondary data – federal surveys and so forth. And at the end of that study, I said, “Never again am I using secondary data. Next time, I’m designing my own questions. I’m collecting exactly what I want. I’ll know exactly what I have.”

And so the next study I was involved in was our Alzheimer’s study, where we recruited 100 patients in nursing homes, 100 patients in the community. We designed the questionnaires. We collected the data. We cleaned the files. And at the end of that study, I said, “Never again am I going to collect my own data. Next time, I’m going to use data sets that professionals have created.”

The point here is that there are pros and cons to each type of data, and it’s just important that you understand up front what you’re getting into and what the limitations will be.

Fifth lesson: Advocates want your results yesterday and preferably with tomorrow’s data.

But there’s always this kind of tension because, while they understand that your study is more useful to them if they can quote a New England Journal article rather than just some data that they got from a researcher at UCSF, they can’t wait. When you’re doing a study, people immediately start calling you up. As soon as the abstract is published that the study has been funded, you get phone calls saying, “Well, do you have any results yet?” And they don’t like it when you tell them, “Well, it’ll probably be two years till we really can say anything here.”

So I think it’s sort of a back and forth. It has always made me feel an obligation to try to publish stuff as quickly as possible so that people can use it. And I have to say that if I ever get to the point where the advocates don’t want my research, I may stop doing it. So it’s kind of a mixed blessing here.

Lesson six: Science is one thing, politics quite another.

Sometimes, you do a rigorous study. You spend years creating these wonderful estimates, and then for some reason totally unrelated to the research, you just can’t get the thing out.

Let me give you two examples. The study that we did on the cost of injury had a couple of controversial things in it. There was something related to handgun regulations, as I recall, and somehow nobody would approve this study to be released at the CDC. The study, I think, outlasted three CDC Injury Center directors. There was nobody there to approve it until finally, somebody came on board and said, “Yes, this is really important work. Let’s get it out the door.” But it was so frustrating because we had done all this research, and then we couldn’t get the study released.

In the same study, there was a chapter that someone else wrote that reviewed some of the injury interventions and whether they were cost-effective or not. Well, it turns out that most injury interventions that this person reviewed were cost-effective except for one, and that was offering driver training programs in high schools. It turns out that, in fact, when you have a school-based program to educate kids about driving, it doesn’t really affect their risk factors all that much. What it does is get a whole lot more 16-year-olds out on the highway, and that’s the age group that’s most likely to be involved in motor vehicle crashes.

So we sent our report around for review, and all they picked up on was that one paragraph. And it was infuriating because we’d done all this wonderful work, and it looked like this little piece was going to delay the report. And it took quite a while to get beyond that, and the report finally came out. But sometimes, it’s not the science that holds things up. It’s the political process that ultimately took precedence.

Lesson seven: Find the people who can make things happen.

My good friend Bill Rankin comes to mind here. Some of you know Bill. He’s established a nonprofit that provides health care services and education to rural villagers in Malawi, training nurses there as part of the program. I’ve been privileged to work with this group. At one point, I remember sending out a couple of articles to the team, related to tobacco, because I believe that tobacco is going to be a major epidemic in Malawi. All the signs are there. It’s the main cash crop and the tobacco industry is very active. I was trying to persuade my group that we ought to include some questions on a survey we were developing that relate to smoking.

Not five minutes later, I received an email back from Bill saying, “I’m going to forward your message to our senior staffers in Malawi and ask that they incorporate no-smoking strategies into the health talks in the villages.” I just sat there and grinned. You get the right information to the right person and something just happens. You don’t have to wait three years for three CDC directors. So the lesson is: Find the Bill Rankins in the world and put the right information in their hands.

Lesson eight: Economics isn’t everything.

This has probably been one of the more difficult lessons for me to learn. As an economist you do a study, you come up with a clear graph that shows your results, and obviously based on those – the sine of the coefficient and the third equation – this is what we should do. Well, decisions are often based on things other than just economic analysis.

I remember a study that Pat Fox and I were involved in years ago. We looked at the cost of heart disease in California, and the conclusion was that the costs were greatest for middle-aged white men. So the conclusion obviously is we should use our resources for preventing heart disease in middle-aged white men, right? Well, not exactly. It’s not just the economics that are important here. We need to consider other factors.

Lesson nine for me was that it’s very important not to lose sight of fundamental values.

One of the key lessons of our dynamic models was that it’s not just the health care costs you need to think about. You also need to think about the value of life. You need to think about what it means for people to be in better health. So I think it’s really important not to lose sight of some of these fundamental goals that we hold dear.

Final lesson: There are tremendous benefits and joys from working with colleagues in other disciplines.

These are some of the people that I’ve worked with over the years, and none of these are economists. I think about a third of them are nurses. I had great fun putting this slide together, and I ran out of slide sooner than I ran out of colleagues. I kept thinking of other people and making them smaller. So forgive me if you’re not up here, but these are all wonderful people that I’ve worked with over two decades.

I couldn’t do my work without the expertise that some of my colleagues bring to the table. But more importantly, it’s just more fun to work with people with different perspectives. It has been tremendously satisfying.

So let me come back to my original question. What is an economist like me doing at the UCSF School of Nursing? Well, the short answer is that I’m trying to change the world. But where else could I find such wonderful and supportive co-workers, such brilliant and insightful colleagues, the freedom to pursue the work on the topics that I think are most important? I really can’t imagine being anywhere else.

Thank you.

 

Add new comment