Food stamps

First, a look at food stamps, or the Supplemental Nutrition Assistance Program (SNAP). Currently one version of the Farm Bill has proposed cutting the program by about $20 billion. A fundamental question is: does SNAP work? Does it actually manage to reduce hunger? Two classical studies were done on this question—one survey of primarily rural Southern and Appalachian towns in 1968, and another in the late 1970’s—finding reductions in malnutrition after the introduction of food stamps. A more recent comprehensive 2004 review of primary studies found that the program does indeed appear to reduce the likelihood of severe hunger and malnutrition, doing so more effectively than just cash benefits would (given that SNAP money is restricted away from tobacco and alcohol, for example). However, there are some caveats: that while SNAP does increase overall food expenditure, many people probably cannot receive sufficient nutrition through SNAP (estimates of SNAP benefit levels vary around ~$4 per person per day), and a “food stamp cycle” appears to occur, in which people buy cheaper calorie-dense (read: unhealthy) foods early in the month immediately after receiving SNAP provisions, in anticipation of future hunger–which unfortunately may contribute to their risk of obesity and chronic disease.

Has SNAP generated massive dependency with subsequent negative economic consequences, as some have alleged? It’s true that as the recession progressed, about 1 in 7 Americans ended up on food stamps. That’s expected expansion due to automatic stabilizers (safety net budget expansion during periods of higher demand, to buffer fluctuations in real GDP—fulfilling the definition of a safety net program), complemented by some modest further program expansion. But a major question is: will all these Americans end up being permanently on welfare programs? The data suggest, to the contrary, that as economic recovery occurs, people do return off of the safety net programs to financial independence, with safety net program participation following the poverty cycle (duh, that’s the point).

Indeed, a few careful studies of SNAP’s economic effects answered a related question: can we afford this safety net program? And what are its downstream economic effects? The first study analyzed how SNAP users spent their income and what poverty impact (and therefore downstream tax revenue increase) the program had, including analysis of rate, depth and severity of poverty during the 2001 and 2007-2009 recessions, using a well-studied nationally representative sample of households. After considering a number of control variables and using a well-accepted standard statistical analysis strategy, the researchers observed that SNAP led to an average annual decline of 4.4% in the prevalence of poverty from 2000 to 2009, as well as a larger reduction in the depth and severity of poverty during that time, as families were able to make use of their earnings for other expenditures. This alludes to the “heat or eat” dilemma faced by many families who have to choose between paying for other crucial life necessities (e.g., heating costs in winter) or food. A second study involved calculations of how much the shift in income expenditure from food to nonfood items (as participants can spend on heating, etc., instead of just one food) helps stimulate economic growth during recessionary periods. It was observed that about every $1 in food stamp program funding that is deficit spending contributes to nearly $2 in subsequent economic growth and future deficit reduction due to increased job growth and consumer expenditure even in nonfood areas as people are able to make purchases like utilities that they otherwise wouldn’t be able to afford—a high fiscal multiplier, meaning that the program is stimulatory to the economy during recession rather than a net debt-expander.

Farm subsidies

Conversely, a major component of prior Farm Bills have been the large welfare disbursements in the form of agricultural subsidies to corporate crop firms, which do not appear to have the same stimulus, poverty reduction or hunger benefits as SNAP. This year’s Farm Bill contains changes to farm subsidies; current sponsors of the Farm Bill have advertised reducing or eliminating direct subsidies to farming businesses. But upon close inspection of the Farm Bill draft, these direct subsidies appear to be offset by the expansion of crop insurance programs. The Congressional Budget Office calculated that most of the money saved by cutting direct subsidies would be used to boost such insurance programs and their related components. While insurance sounds like a good idea, given risky weather events, in fact this particular insurance provision involves taxpayer subsidization of both the premiums on the insurance as well as rather generous payouts that some have argued may lead to more risky farming practices ecologically, and essentially constitute a crafty form of continuing subsidies in sheepish insurers’ clothing.

What are the data on current subsidies and crop insurance programs to agricultural businesses? First, at present, it does appear that a vast majority of subsidies go toward large agricultural firms rather than small-scale farmers. Second, far from being a truly new proposal, it appears that—over time—crop insurance has been an increasingly large redistributive mechanism from taxpayers to larger agricultural firms during the last several years, already exceeding direct subsidy payments even before this current Farm Bill draft (see graph below). Similar to direct subsidy programs, while some farms annually collect more than $1 million in crop insurance premium support, the bottom 80% of policyholders annually collect about $5,000.

A more extensive set of reviews of subsidies and their relationship to global hunger and energy policies is available here. The direct subsidy data are available for public download and analysis here. The deep irony in this data is that it appears less debt-reducing to cut SNAP programs than to alter the crop insurance and subsidy proposals for large agribusinesses.

Food aid reform

A final issue that may make it to center stage is the White House proposal to shift foreign food aid toward purchases from nations near hunger zones rather than American-grown food (although at the time of this writing, it appears that the House is voting against even discussing this issue). American vegetable oil, wheat and other crop commodities are sometimes “dumped” in poor countries, with adverse effects on local farmers as crop prices are destabilized; the food itself also appears to be less than ideally matched to actual hunger needs, and more matched to surplus supply levels among large agribusiness producers.

At risk, of course, are major government contracts to U.S. for-profit firms involved in the aid business (not just agribusiness, but also consulting firms, shippers and processors). There have been numerous commentaries about this proposal of so-called “local and regional purchasing” to improve food aid, but most interesting and less accessed are the actual data on government contracts to companies who are in the current food aid business. Here’s a look at the most recent publically-available data on which companies get contracts, for which food items, and for which countries:

The report, commissioned by Oxfam International, involved three related, year-long quantitative and qualitative analyses of the food price spikes: (1) the analysis of key global food security indicators, including availability of food for consumption (amount, type, and quality); access (to the required type, quality, and quantity of food in terms of affordability, adequacy of allocation mechanisms, and meeting social and other preferences); and utilization (the ability to benefit from consumed food, which is dependent on the nutritional content, the social value, and the safety of available and affordable food) [for those who would like to analyze this data themselves, it's publicly available here]; (2) twenty-three qualitative community case studies in  10 countries facing spikes, including interviews in both urban and rural areas to identify how people coped with price spikes volatility; and (3) an iterative ‘Q-squared’ process that involves both qualitative and quantitative analysis, in which the qualitative research informed lines of quantitative enquiry, whose findings in turn suggested issues to probe in future rounds of qualitative work.

Theory versus data

The findings from these analyses were profound. Indeed, as predicted by classical theories, many people are earning more money, but at great cost, and not necessarily with great benefits to them. Wage rises did not typically match rises in the cost of living. Rather, people had to cut back on food and other expenditures, substitute for inferior quality food, spend significant time shopping around, and spend more time and money growing and gathering food. The impacts are felt in homes, relationships, communities, and work-places, causing both stress and inter-personal conflicts.

Directly linked to the food price spikes, the researchers observed a major social transformation in several communities: that “crisis mode” for families became routine. Being stressed and desperate for food became a normal part of daily life, and money itself and its pursuit become so obsessively prevalent in the researcher’s longitudinal observations that it was notably ruining social status, relationships, love, and values. Everyone’s priorities and relationships seemed to be profoundly affected by the food price volatility.

One of the major findings was that while much of our conversations around food and public health are centered on the question of cost, the volatility itself–not just the absolute price–seems to have its own major impact. Furthermore, contrary to classical theory, food price rises don’t seem to have attracted young people into farming; agriculture is simply unappealing because of unpredictable returns in the context of volatility, high input costs, and difficult living. No one in their right mind wants to put up with the difficulties of being a farmer, essentially. And there’s no sign that the price rises are going back down at this point. Projections suggest continued high prices in the most-affected regions, even though other commodity prices have decreased (reducing wages) and volatility has decreased around the stably-high price levels.

Many of the non-farmer laborers also had difficulties coping with the cost of living issues presented by the food price spikes. Some formal sector groups mobilized successfully to demand higher pensions (Bolivia), or to increase wages (Bangladesh and Indonesia), but they also entered into riskier work situations: gold mining; sex work; and fishing in dangerous conditions. Migration also increased in the context of searching for higher-wage work. In most of the 23 research sites, while wages rose, this increase has not kept pace with the rising costs of fuel, rent and agricultural inputs, which climbed again in 2012, in the wake of five years’ worth of price rises.

Food substitution

While people did start substituting for their usual foods with less expensive ones, this substitution appeared to present several important public health conundrums. People ended up consuming less than their nutritional requirements, risking micronutrient deficiencies. Surprisingly, even some better-off urban populations were unable to afford basic foods, eating less diverse diets and substituting for foods they dislike using ‘hunger recipes’ that are designed to stretch meals. While this is classically described in the hunger literature, more surprising and atypical was that processed foods and sauces marketed by multinational food companies also became popular ways of making plain food a bit more palatable during this food price spike, and these cheap flavorings contain high quantities of substances associated with chronic cardio-metabolic disease. Food safety also became an issue as people relied more on foods that were of questionable quality, and included potentially-contaminated fish, poultry, and meat and broken eggs.

What to do

The researchers asked the people most affected what they thought would be most helpful in this situation, and supplemented that qualitative analysis with the statistical study of what seems to have worked well in the past. A few key recommendations emerged out of the report: (1) since it is too late to start developing price-buffering schemes when a price spike strikes, communities and governments need to design social assistance policies aimed at protecting against spikes in the form of temporary cash or food transfers, or by providing subsidies that are automatically triggered by price rises; these should be adjusted to real changes in needs by linking social protection to inflation; (2) food security policies need to buffer not just against cost but also against price volatility through appropriate management of food reserves, and regulation of anti-competitive behaviour in the grain trade (hoarding); and (3) future farmers need to be supported, as volatility and low returns to small-scale farming is turning people away from a life on farms.

But despite these explicit recommendations, one of the most striking aspects of this report is understated: that it goes into far more depth, with more careful and longitudinal qualitative ethnographic detail, that many prescriptive or analytic reports on the food price spikes. Indeed, the care with which this report was conducted reveals that another deficit in our conversations on food policy is actual monitoring of the experiences of those most affected. Rather than theorizing about equilibrium prices, the report’s authors reveal that directly monitoring how people are actually paying for food and other basic living costs needs to be conducted to correct often outdated theories, and monitoring real wages and earnings may help us better understand vulnerability, resilience and the impact of global price shocks on nutrition and livelihoods.

That question, it turns out, can be answered through data and careful research on recessions both past and present. My colleague David Stuckler and I are today releasing our peer-reviewed book, entitled The Body Economic, in which we boil down over a century of data from across the globe to answer the question of what policies actually improve both our economies and our public health during and after economic recessions.

Is bad health inevitable during an economic downturn?

Here’s a potentially puzzling fact: during recessions, overall death rates often decline. At least, that’s what appears to be true when you just look at average death rates and average GDP. Average death rates often decrease during economic downturns.

But there’s a problem with this analysis: average death rates are… well… averages. Using more detailed, disaggregated data, we’ve found that hidden beneath this average trend are a lot of variations—with some people doing very well, even becoming healthier during recessions, and others falling into serious illness. For example, much of the decline in death rates during both The Great Depression and our current recession is due to reduced traffic accidents. As people can’t afford to pay for gas, they drive less.

But hidden beneath this decline, at least in the United States, Britain, and several other countries, is a dramatic rise in suicide rates and alcoholism (spikes in deaths that get “washed out” or masked by the large drop in traffic accident deaths). Overall alcohol sales decrease during recessions, but that average is again deceptive—many people buy one less “two buck Chuck” at Trader Joe’s, but a significant population of over 700,000 men in the United States (mostly single young men who lost work) started binging during the economic recession.

Here’s the good news: these suicides and alcoholism cases aren’t inevitable. Correlation is not causation, and that’s precisely our point: in Sweden, for example, massive spikes in unemployment were not accompanied by increases in suicides:

We can compare the Swedish experience to that of Spain and many other countries, where suicides did spike in parallel with unemployment:

These increases in suicide rates also affected the United States, where about 5,000 “excess” suicides have been observed during the recession (that is, suicides above and beyond pre-existing trends), which are significantly related to local unemployment rates:

What did Sweden do differently from Spain, the United States, and many other countries? It turns out that Sweden started a very specific program that helped people get back to work quickly, simultaneously boosting the economy and reducing rates of depression and suicide. That program wasn’t a simply unemployment check or “hand out”, but it involved a case management program that worked with both the newly-unemployed and their firms to generate new jobs. It was ‘part carrot’ and ‘part stick’: people received help, but they had to participate in the program to get unemployment benefits. To test out whether the program really worked, randomized controlled trials were performed in a variety of different settings and populations, which confirmed the program’s impact on the economy and on health outcomes including depression rates (so it’s not just “being Swedish” that’s special).

Conversely, there’s some very bad health news from our current recession, and from previous recessions: austerity programs (which involve sweeping budget cuts in an attempt to reduce short-term deficits) appear to be counter-productive for both economic recovery and our health. This is evident in the East Asian crisis of the 1990′s, for example, where some countries decided to chop budgets through a sequester-like program that ended up ironically prolonging their economic recessions and increasing malnutrition and other prevalent public health problems; other countries underwent deficit spending and actually stimulated their way out of deficits and recessions, into longer-term economic booms. The stimulus countries avoided major catastrophes like a rise in infectious disease rates that their austerity-promoting neighbors experienced as vaccination rates and public health programs faltered from budget cuts.

The same patterns of austerity versus stimulus were observed in the mid-1990′s in Eastern Europe (and, interestingly, the defenders of austerity and mass privatization in Russia omitted data from their analysis and made serious coding errors, just as in the recent Reinhart and Rogoff affair). These cases are “natural experiments” in which similar populations experience a common shock (in this case, an economic recession), and experience the effects of different policy decisions, allowing us to study which policies actually benefited both the economy and public health (through instrumental variables and related techniques).

During our current recession, Greece has emerged as a particularly bad public health disaster related to austerity. In Greece, major public health statistics were already looking bad due to the effect of the recession on poverty rates and associated outcomes like depression and alcoholism, but austerity itself led to serious spikes in death rates that were avoidable. Austerity measures chopped a mosquito-spraying budget, for example, that led to the first malaria outbreak in decades—an outbreak more costly to control than to prevent. Similarly, an HIV outbreak started after young people facing a 50% unemployment rate turned to injection drug use just as the HIV prevention budget was slashed. Economic and epidemiological projections reveal that the rates of these diseases would have been much lower had a better stimulus been enacted in Greece, helping the country grow out of the slump and preserve key disease-prevention budgets.

Can we afford safety nets in a recession?

In discussing Greece and other countries facing austerity, a major—almost religious—belief seems to be that while public health programs might be useful, we simply can’t afford them. It’s “common sense” to believe that deficits must be reduced to prevent inflation and avoid all sorts of down-stream problems from too much government spending. But it turns out that this philosophical belief is based on shaky economic logic, and has dire results for public health.

Whether to cut budgets, or take on more stimulus spending at any given time can actually be answered through data and scientific approaches, rather than simply through ideologically-driven theories. As we discuss in the book, a key calculation called the “fiscal multiplier” can help us determine when it’s a smart time to cut or stimulate government budgets for public health or any other sector—and, conversely, which budgets are particularly helpful to cut or boost at any given time. The multiplier is a calculation that describes how many dollars of economic growth we get for each $1 of government spending on a particular program at a particular time.

During the ongoing recession, many people assumed that the multiplier was less than 1–meaning that each $1 of government spending would result in less than $1 of economic growth (a net loss) due to crowding-out of the private sector. But now multiple different research groups have actually calculated the multiplier from real data sources (rather than assuming its value)–and found that its value is actually much larger than 1. What this means is that sequester and austerity at this point in time will tend to produce more long-term costs than outweigh short-term budgetary gains, because we need stimulus to promote more employment and consumer spending. The multiplier for public health programs is particularly large because of the large workforce in the health sector, and the downstream effects of a healthy workforce on the economy.

The good news is that some people did take this advice to heart. Iceland, for example, actually experienced a larger crisis than Greece (with debt at 800% of its GDP during the recession), due to bad investments among some of its key banks, which had invested in the US mortgage-backed securities that generated our housing crisis. But Iceland managed to take on a strategy that led to its rapid economic recovery; the country subjected its economic policy to a democratic vote, and decided to pay off its bankers’ debts slowly rather than all at once. It chose not to pay for the banker’s mistakes by chopping public sector budgets. Iceland remained one of the happiest, and healthiest, countries in the world despite experiencing what The Economist called the largest banking crisis in world history relative to the size of its economy. And looking through the history books, it turns out that Iceland is not an isolated case—rather, the Great Depression and The New Deal provide some other key examples of how making data-driven public health decisions during and after an economic crisis can avert serious disaster for both our economy and our health.

What to do?

A major section of our book describes the data on what we can do next—how to maintain our public health systems in both bad times and after our recovery. The first, and perhaps most important, principle is a simple one: “first do no harm”. It’s often stated that the majority of factors that affect our public health have nothing to do with which medical school our doctors went to; health doesn’t start on an exam table or in an intensive care unit. It starts at home, with the quality of the air we breathe, the food we eat, and whether we smoke or drink too much. Our economic policies have a profound influence on these factors, as evidenced by decades of research.

If our social and economic policies were tested as rigorously as we test pills in a clinical trial, we would have stopped our ongoing austerity experiment long ago, given the profound evidence of its lack of benefits and deadly side-effects. We should test our social and economic policies with the same evidence-based rigor that we test other interventions that affect our health. But that requires a willingness to follow the data—and to use evidence to guide how our governments will protect their electorate in times of great vulnerability to poverty and illness.

This is a guest post by the computational epidemiologist

Most of us are aware of the “big data” revolution fueled by electronic information. It has been suggested that big data, along with hypothesis-free methods popularized by films such as Moneyball, will allow for an unprecedented growth of knowledge across disciplines, including epidemiology and preventive medicine. While I am a bit more circumspect in expectations (there is no substitute for survey data in many cases), I do believe that electronic data collected for a fraction of the cost of survey data can work hand-in-hand with research derived from more traditional sources.

Our study, published this month in the American Journal of Preventive Medicine, is a great example of the complementarity of big data approaches to mental health research. Previous work had identified mood symptoms as varying in many individuals suffering from depression. Using Google search as a proxy for changing patterns in mental illness, we sought to better understand seasonality in mental health.

For the analysis, all Google mental health queries were monitored in the U.S. and Australia from 2006 to 2010. Additionally, queries were subdivided among those including the terms ADHD (attention deficit-hyperactivity disorder); anxiety; bipolar; depression; anorexia or bulimia (eating disorders); OCD (obsessive-compulsive disorder); schizophrenia; and suicide. A wavelet phase analysis was used to isolate seasonal components in the trends, and based on this model, the mean search volume in winter was compared with that in summer.

While some conditions, such as seasonal affective disorder are known to be associated with seasonal weather patterns, the connections between seasons and a number of other major disorders was surprising. We found eating disorder searches were down 37 percent in summer versus winter in the U.S., and 42 percent in Australia. Schizophrenia searches decreased 37 percent during U.S. summers and by 36 percent in Australia. Bipolar searches were down 16 percent during U.S. summers and 17 percent during Australian summers; ADHD searches decreased by 28 percent in the U.S and 31 percent in Australia during summertime. OCD searches were down 18 percent and 15 percent, and bipolar searches decreased by 18 percent and 16 percent, in the U.S. and Australia respectively. Searches for suicide declined 24 and 29 percent during U.S. and Australian summers and anxiety searches had the smallest seasonal change — down 7 percent during U.S. summers and 15 percent during Australian summers.

Typically, telephone surveys are used to assess population mental health, but this approach has a large margin of error because respondents may be reluctant to give honest answers about their mental health. This approach also has high material costs and as a result, investigators are not able to collect as much data as they need to assess seasonal patterns, especially for rare mental illness. Data availability (or the lack there of) has tremendous consequences on theoretical and subsequently clinical developments in mental health. For example, we saw strong seasonal patterns for schizophrenia, a disease for which symptom severity had not been closely associated with seasonal patterns. In contrast, tremendous attention had been given to seasonal birth patterns in schizophrenia. Why? Population surveys readily collect birth date without any added planning, concerns with sensitivity/reliability/validity, or additional budgeting. Since all theories are based on some data, our approach can provide the beginning data stream for theoretical development in global mental health seasonality.

Clearly, these results are not intended to be definitive. Further research is needed, especially for understanding the link between search patterns and symptomatology. However, intuition suggests that these results are reflective of an important link between the seasons and mental health that goes beyond our previous understanding of these conditions. This kind of work can continue to cost effectively inform the field on a variety of vital health topics, and ours are just the beginning steps.

Disclosure: Dr. Ayers holds an equity stake in Directing Medicine LLC that advises hospitals, allied health groups and industry on data mining strategies.

Or, more precisely, they invested in a new Drugs for Neglected Diseases Initiative (DNDi) that sought to develop an alternative model for the research and development (R&D) of new drugs for neglected diseases. By “neglected” we mean diseases that are only caught by people too poor to pay for medications: illnesses like malaria, visceral leishmaniasis (VL), sleeping sickness (human African trypanosomiasis, HAT), and Chagas disease. These sicknesses are currently treated by medications that are  too expensive, no longer produced, highly toxic, or ineffective.

Over a decade later, DNDi and other initiatives have highlighted some stark failures in the R&D process. Many of us were under the impression that R&D initiatives needed to further fuel private sector incentives for drug development. We were surprised to see the data on how the private sector patent-based pharmaceutical industry did indeed have high profits as a percentage of revenue (typically far higher than the average Fortune 500 industry) and per their tax records were not putting that revenue into R&D as much as they were funding marketing, executive salary and the like:

Looking further, an inquiry into the  inventors of 252 new drugs approved by the US Food and Drug Administration from 1998 to 2007 revealed that between a quarter and a third of did their innovations through university labs rather than industry; and if you narrowed the list of drugs to those that weren’t “me too” minor derivatives but truly novel therapeutically-significant innovations, the fraction attributable to university labs was even higher.

So it was appropriate that recently the group Universities Allied for Essential Medicines, a student-led organization run across several college campuses internationally, has looked into strategies to align university-based pharmaceutical research with population needs (e.g., rather than focusing on developing highly-profitable drugs like Viagra and hair-loss treatments that have little to do with public health…as much as I might need the latter). [Full disclosure: I worked with the organization in the past].

The group has done some interesting work, such as designing new patenting and licensing protocols for universities so that taxpayer-funded, university-based research can actually be licensed to pharmaceutical companies but ultimately still have legal provisions that allow low-income country populations to make use of them.

But their recent work also includes investigating whether the last few years of efforts to increase research into neglected diseases has actually sparked progress on university campuses.

Sadly, the answer–for most universities–is no. In their recent release of a “report card” on university research, the group finds that few universities are doing much of anything on neglected disease research.

The universities were graded on a number of criteria (take a look at their detailed methodology), including whether they invest in medical research that addresses neglected diseases, and whether they license their health technologies to for-profit companies in ways that ensure treatments reach developing world patients at affordable prices. The organization looked at the 60 highest-funded universities based on FY 2011 funding figures from the NIH and CIHR’s publicly available funding databases (RePORTER and Funded Research Information, respectively, as well as a few other public sources). They also evaluated 14 specific metrics of where research funds were going, based on self-reporting from university officials in response to standardized survey questionnaires. The metrics were standardized, and normalized with respect to degree of institutional funding, in order to produce fair comparisons between differently-sized universities.

Overall, the University of British Columbia came out on top (scoring an A- on its report card), particularly as it’s engaged in a number of licensing efforts to keep its inventions accessible to the poorest groups. Sadly, my university (Stanford) got a “C” (tying with Berkeley…boo).

From the university researcher point of view, a key challenge in “soft money” schools is to gain grants to conduct research on these diseases. And while the Gates Foundation and a few other initiatives provide venues to do so, the universities’ excuse is that there aren’t sufficient funds available, the NIH excuse for not funding these diseases is that they don’t affect Americans (and their foreign budget is dwindling), and the private funders are left to shore-up the remaining budgetary space. So we have proposed some alternative strategies. Of course, low funding levels don’t forbid universities from enacting open licensing strategies when they conduct research that’s of potential benefit to those who can’t afford to pay for medications–a key reason why alternative licensing mechanisms shouldn’t have a low barrier to adoption, since they affect neither university finances nor make meaningful dents in the revenues of for-profit manufacturers (since the poorest aren’t part of the market share anyway).

First, we looked at the industry’s data to examine how much sales in low- and middle-income countries even made up significant business for soda companies. To our surprise, the majority of soft drink sales are indeed outside of North America and Europe, and the rate of increase in these sales is highest in low- and middle-income countries:

Of course, merely correlating a rise in per-capita soda consumption to a rise in obesity or diabetes would be silly—there are many other changes taking place at the same time in low- and middle-income countries, such as urbanization and changes in the work environment that are associated with lower physical activity, changes in a number of other foods being consumed (like higher meat consumption, and higher overall calorie intake as incomes rise), and aging, among others. So instead of merely doing rough correlations, we looked at age-standardized estimates of overweight, obesity and diabetes, and corrected for other types of foods (e.g., other carbohydrates, fruits, vegetables, meats, fats, oils, and total calories), as well as aging, income and urbanization. After these factors were controlled for, soft drink consumption still was significantly related to a rise in overweight, obesity and diabetes prevalence across the globe, and most prominent in low- and middle-income nations. Even more concerning, only a small increase in per-capita consumption (gallons per person per year) was associated with a large increase in weight, after controlling for the other factors—it didn’t require much increase in consumption to show up as a major increase in overweight prevalence.

As a “natural control group”, it turns out that bottled water consumption has also increased with economic development in many low- and middle-income countries. So if there were “unobserved confounders” that we didn’t think of, including bottled water consumption in the statistical analyses, and analyzing the relationship between weight/diabetes and bottled water consumption, should give an indication that the soda-weight/diabetes relationships were spurious. But in fact, the natural control of bottled water consumption had no relationship to overweight, obesity and diabetes prevalence. Adjusting BMI thresholds for Asian countries (since a high BMI may be a poor indicator of disease-associated obesity among Asians) didn’t make any difference either. We also isolated ourselves to analyzing carbonated full-calorie soda rather than all sugar-sweetened beverages, to reduce potential confounding in industry data sources in which juices and other products that are sometimes natural and sometimes have added sugars may be mixed together.

What was particularly interesting was that income did not inevitably result in higher soda consumption. That is, there was great variation between countries in how much soda they consumed even as they experienced economic development. This is the good news: that there may be social policies and practices out there that may allow people to experience economic development without inevitably leading to dietary changes that appear to contribute to obesity and diabetes.

Since the start of the recession in December 2007, alcohol-related emergency room visits and hospitalizations increased in a few locales. Some health departments therefore anticipated that alcohol-related health problems would require increased medical and public health attention during this period. Binge drinking has been found to be correlated with economic downturns in the past, even among those who remain employed. Yet some other agencies have proposed reducing or eliminating substance abuse programs, especially in the context of budget deficits. Some studies suggest that economic downturns are associated with significant declines in overall alcohol use and other health risks, and have therefore been cited as justification for such cuts.

This conflicting evidence have led to a variety of hypotheses about what might be going on: an ‘uncovering’ hypothesis suggesting that potentially abusive drinkers will be “frightened out of drinking” by the threat of job loss if they continue to drink; an ‘income-effect’ hypothesis suggesting that less income with which to purchase alcohol will lead to less drinking during the recession; and a ‘provocation’ hypothesis reasoning that people will cope with insecurity and stress related to real or threatened job loss by drinking more. Depending on which hypothesis you subscribe to, you might expect decreased or increased drinking during the Great Recession, and therefore recommend a different policy response in terms of funding alcohol abuse programs.

A recent epidemiological analysis by Jacob Bor at Harvard (full disclosure: I’m a co-author on the study) newly informs the debate between these alternative hypotheses. The analysis uses data on drinking patterns during the Great Recession using a nationally representative sample of over 2 million U.S. adults who were subjected to a standardized alcohol drinking survey. The study assessed drinking behaviors along several dimensions: drinking participation, drinking frequency (number of drinking days), drinking intensity (number of drinks on drinking days, maximum number of drinks in a single episode), total alcohol consumption, and frequency of binge drinking. In contrast to industry reports that measure total consumption of different alcoholic beverages, this analysis was able to assess changes across the full distribution of drinking behaviors, ranging from abstinence to frequent binge drinking.

Here’s the surprising result: during the Great Recession, rates of abstinence from alcohol increased among U.S. adults, but total alcohol consumption also increased. What does that mean? It means the country was diverging: most people drank a little less Charles Shaw wine at dinner because of the pinch on their pocket-books, but there was a smaller population about the size of a small city—about  770,000 adults—who started binging (a statistically significant rise in the prevalence of frequent binge drinking of 7.2% relative to baseline levels). In other words, there was simultaneously a rise in the number of moderate and heavy drinkers and a decline in the number of light drinkers.

Who were these bingers? During the recession, frequent binge drinking was highest among non-Black, unmarried men under 30 years who were unemployed for less than one year. This finding is consistent with existing literature. Binging rose the most among people 25-34 and 55-59 years, age groups that are the most vulnerable to job scarcity and job insecurity (either not finding a job after college/in the early years of work or being forced into early retirement).

This polarization of drinking behaviors supports a few of the theories at the same time: an increase in abstinence from drinking (the “income-effect” hypothesis) among primarily young people with less than a college education (who may be more sensitive to changes in expected income), but a rise in frequent binging and total alcohol consumption (the “provocation” hypothesis) suggesting job insecurity, the threat of loss of a home or life savings, or other recession-linked exposures may have led to greater use of alcohol as a coping mechanism among a subpopulation. However this study found no evidence of an “uncovering” effect, in which higher-risk drinkers reduced consumption due to threat of job loss during the recession.

Although this analysis is descriptive, it allows us to document some of the potential human costs of The Great Recession, and provide evidence that can guide clinicians and policy makers to target resources towards populations at highest risk. It also suggests that common analyses of “counter-cyclical” phenomena during recessions–that is, looking at how average rates of substance abuse or mortality decline during recessions–is likely to be a bit foolish: the averages disguise a striking variation within the country. Vulnerable people appear to be hidden behind the average statistic.

Why did we do this study?

Our study sought to address this question: could sugars affect the risk of diabetes, even independent of their role in affecting the risk of obesity? Laboratory studies and experimental data suggest that the relationship between obesity and diabetes pathogenesis is not entirely clear. For example, several studies have estimated that about 20% of obese individuals appear to have normal insulin regulation and no signs of the metabolic syndrome (no indication of diabetes); such individuals also have normal longevity, such that focusing on their obesity may be obsessing about an imperfect marker of metabolic dysfunction. By contrast, up to 40% of normal weight people manifest aspects of the metabolic syndrome, so clearly something else is going on with metabolism besides being fat. There is really no doubt that obesity is a statistical risk factor for diabetes; our study was not designed to rebut that idea at all. Rather, it was designed to investigate an additional possibility that the availability of sugars may also have an independent role in diabetes, even aside from contributing to weight or total calories consumed. This could explain some puzzling findings about why diabetes rates among some populations have escalated independent of changes in obesity rates, as we discuss in the paper.

In addition to sugar leading directly to obesity, there appear to be other aspects of diabetes besides obesity that sugars contribute to. These include hepatic de novo lipogenesis and reduced fatty acid oxidation, which results from the liver’s metabolism of some sugars in the fed state in a manner that generates lipogenic substrates in an unregulated fashion. This process forms excessive liver fat and inflammation that inactivates the insulin signaling pathway, leading to hepatic insulin resistance. Sugary foods also appear to contribute to the development of insulin resistance in laboratory-based studies. Reactive oxygen species are produced by the Maillard reaction, damaging pancreatic beta cells, and leading to a subcellular stress response (the “unfolded protein response” in the endoplasmic reticulum) that drives insulin inadequacy. In concert, insulin resistance and reduced insulin secretion lead to overt diabetes. But all of these laboratory studies cannot tell us how important these findings are at a population level—that is, whether these relationships actually matter in the real world and could be contributing to epidemic-level diabetes rates.

What was our approach?

In the ideal scientific world, we would take two groups of people and give them very carefully calculated diets, in which one diet had higher sugars but otherwise identical total calories, and the other diet had low sugars and otherwise identical total calories. We would have the two groups exercise the same amount and maintain the same weights, then see if the high-sugar group got diabetes more or less than the low-sugar group.

Of course, such a study would not be very ethical. Hence, we had to do some fancier statistics with the data we have available. The problem is that typical medical correlation studies are fairly weak—they look at just one point in time, and—as we ourselves have pointed out several times—correlation does not imply causation because third variables can explain point-in-time correlations (e.g., Christmas cards do not cause Christmas to happen, but the two are highly correlated in time). Even having good control variables does not always resolve this problem.

So we used some more advanced statistical methods from the field of econometrics in order to move beyond the typical limitations associated with medical correlation studies. First, we looked at longitudinal trends in sugar availability and diabetes prevalence rates in 175 countries. Looking at long term trends allows us to conduct time-series analysis, which can filter out a lot of garbage that plagues common medical correlation studies, and allows us to look at long-term relationships that are inherent to food consumption and diabetes risks. In addition to statistically controlling for overweight, obesity, and other calories (not just sugar, but total calories, as well as calories from fats, proteins, etc), and other risk factors for diabetes like tobacco smoking and alcohol consumption, we did a different type of study design.

First, we looked at a “selection model” to examine whether unobserved variables could be an issue—meaning whether having greater sugar around was just an artifact of overall economic development, and since economic development is itself correlated to changes in physical activity and diet, that could contribute to diabetes. Of course, we also used statistical controls for economic development, urbanization, and so on, but the “selection models” determine whether there is unobserved variables that would predispose to a selection bias in the study (the main reason we do randomization in medical trials), and allow us to correct for any such biases we find so that we can closer to an ideal scientific experiment (one reason its inventor won a Nobel Prize).

Second, we used something called a Granger causality test, which looks at precedence—if X leads to Y, then when X increases, Y should increase later, and vice versa. That still doesn’t “prove causality”, it’s just a statistical test that’s commonly used in economics and sociology, but rarely has been applied to medical questions (fyi, its inventor also won the Nobel Prize). We found a so-called “dose-response” relationship between sugar availability and diabetes prevalence, such that increased sugar availability was associated with increased diabetes prevalence in the future, and those countries who reduced their sugar availability also had a proportionate reduction in diabetes prevalence—independent of other changes like changes in economic development, urbanization, physical activity, obesity, and changes in consumption of others foods like meats and fats and total calories. Essentially this is trying to get as close as possible to satisfying an epidemiological equivalent of Koch’s postulates: when you are exposed to an agent, you get the disease, when the agent is removed, the disease goes away, when you re-expose, the disease comes back. This is far stronger than a typical point-in-time medical correlation study. The approach makes use of a so-called “natural experiment”—because it would be unethical to watch people get diabetes after giving them more sugar, we instead look at long-term data from different populations who already increased or decreased their sugar consumption (often due to various changes in economic trade laws), statistically controlling for other factors that differed between these populations to isolate the effect of sugars.

There are, nevertheless, limitations to any statistical study. As we teach our students, we can’t “prove causality” through any amount of statistics—we’re simply halfway between the typical weak medical correlation studies and the ideal case of a randomized controlled trial (which often also can’t prove causality for a variety of reasons, despite common misconceptions). Secondly, our study is “ecological” meaning that we look at lots of countries but this requires that we use large-scale aggregate data, so like any epidemiological study using aggregate data we can suffer from the “ecological fallacy”, which means that when we look at aggregate populations, we can’t be sure that those people eating the greater sugars were the exact same people who experienced more diabetes in that given country. This seems extremely unlikely given the massive amount of data, and the extremely robust nature of the finding when we tested it against two independent datasets over a long duration, but is still worth noting. Third, the data themselves are not perfect—in addition to looking for selection bias and doing “robustness checks” by repeating the analysis while excluding outliers or extreme data points (finding, still, consistent results), we have to acknowledge that food availability data from even the best sources are not perfect, and diabetes surveillance rates (even though we checked them against multiple sources), as well as estimates of overweight, obesity and physical activity in many countries are far from perfect. We just used the best data available to date, given the urgency of this question. Also, body-mass index may not be the best marker of obesity and may mean different things in different countries (especially Asian ones, where the cut-offs between normal weight, overweight and obesity may be less valid than in the United States due to differences in body type), but we did repeat our analysis using alternative cut-offs and didn’t find any difference in the results. Similarly, metabolism is not necessarily the same among all populations; some theorists have hypothesized that South Asians may metabolize calories differently in a way that predisposes to diabetes, but this also is confounded by the fact that body mass index is probably not a very good measure of abdominal adiposity which may be more relevant than body mass per se.

We looked for a variety of confounding factors like food wastage (since food availability is not the same as food consumption), corrected for secular trends in independent and dependent variables, and various time-lags (various alternative durations between the sugars availability and the diabetes outcomes), finding greater exposures to sugars were associated over time with greater diabetes prevalence rates no matter how many of these factors we controlled for. What’s particularly reassuring is that parallel results were obtained by independent scientists at two other universities.

What’s the bottom line?

The bottom line is that this is one of several studies from independent scientific groups that have questioned the old mantra that “a calorie is a calorie”. Some calories may be more metabolically harmful than others, and sugar calories appear to have remarkably potent properties that make us concerned about their long-term metabolic effects. This study also suggests that obesity alone may not be the only issue in diabetes pathogenesis. The study was conducted to understand a statistical theory, using a statistical approach. It doesn’t say anything about any specific person’s diabetes risk or provide any kind of dietary advice. This data cannot distinguish between types of sugars (like high fructose corn syrup versus other types of sugars), nor does it establish more insight into the mechanisms that are at play, which need to be pieced together in laboratory and experimental research studies. This study also can’t inform any specific policies like the New York City ban on large soft drinks, since the real-world effects of specific policies weren’t evaluated in this experiment.

However, the precautionary principle in public health suggests that when there is not scientific consensus on the harms of a substance, the burden of proof falls on those who are declaring it to be safe. What we would do next is conduct a randomized controlled trial of low-sugars versus regular American diets (which are already very high in sugars) and follow people over years to identify whether lowering sugar intakes can lower diabetes risks. But this first study tells us that sugars may be important at a population level, not just an individual or molecular level, and conducting an econometric study is one of the few ways to do that.

Our data-visualization colleagues at Periscopic have released a new report on US gun statistics.

They looked at the FBI’s Unified Crime Report, which describes gun murders from the year 2010 from police precincts across the country, and combined it with an estimate of the “expected life” of each victim based on standard age prediction tools using the UNSD Demographic Statistics database that provides estimates of the probability distribution of life expectancy among various groups in the population.

The calculation provides an estimate of “years of life lost” from gun violence in the US. In total, their estimate suggests about 410,000 years of life lost from gun violence in 2010. See the full visualization, and the breakdown of years of life lost due to type of gun, race, sex, age, region, and type of murder (single victim or multiple) here.

To look at the breakdown by state, the Guardian newspapers “datablog” (one of our favorite data sites) has compiled gun statistics by state through 2011. You can get the full spreadsheet of data here. This data includes breakdowns by state of how much guns are used as murder weapons as compared to other weapons (guns are a vast majority, at about 68%, as compared to knives at 13%), and how gun violence rates have been changing over time.

In 2011, California had the highest number of gun murders–1,790, but given the higher population size that’s about 3.25 per 100,000 people in the state. When you correct for population size, the highest gun murder rate is actually in Washington D.C. (at 12 gun murders per 100,000 people). The rate of gun murders is dropping in most states. However, it’s rising in Indiana, Arkansas, North Carolina, and Louisiana.

It would be interesting to merge this data with gun control law data to examine the relationships. Some preliminary assessments to this effect have been done and suggest both a rise in homicides after a rise in gun availability, and a decline after strict gun control. Unfortunately a lot of folks have been relying less on these systematic assessments than on a website called “justthefacts.com” that claims to have done independent analysis–but we found out that the site is actually run by a conservative/libertarian group that appears a priori opposed to gun control judging from what appear to be rather manipulated charts and data on their website…(they also claim to have published a “highly researched book evidencing factual support for the Bible”!).

We used the NNDB Mapper software (which we previously used to catalog relationships among major global health donors—see our paper in PLoS Medicine) to understand what kinds of linkages exist among prominent institutions, individuals, and major soda companies. The major relationships we looked into and isolated were those that involved corporate board memberships, though it’s also possible to look at direct financial relationships (e.g., donor-recipient) and private political or social clubs as well. The less controlled aspect is that we can’t define individuals included in the NNDB Mapper—institutions or individuals that have been featured in major Internet-archived news outlets are added to the database, and there are no filters on this aspect of the search.

To pick the key soda companies of interest, we used the industry’s own Global Market Information Database to find those companies that were largest in terms of overall worldwide sales of soda per capita. We constructed the next several graphs from that primary data:

For those of you who like to be precise, we’re referring here to “non-diet, non-alcoholic carbonated sugar- or fructose-sweetened non-fruit-juice beverages” (i.e., what those in the Midwest call “pop”). Most of this stuff is now sold outside of the United States, interestingly:

And when we look at the major linkages (financial, political, and social) between the major soda corporations, their boards members/top executives, and other prominent folks in government and the public sphere, we see a rather extensive set of linkages for many companies—that is, they’re highly connected to everything from school boards to the United Nations (perhaps not surprising to the jaded among us):

Some prominent cases are particularly apparent in the Coca-Cola executive board. New York City Schools Chancellor Cathleen Black appears to have a remarkable history of relationships between both the school system and the soda company, as recently reviewed elsewhere. There are a few former Senators, Cabinet Secretaries, and Ambassadors on the diagram above; some prominent members of the global health policy scene (members of the Gates Foundation) and former leaders of Latin American and Eastern European governments; and perhaps most interestingly a lot of active service industry conglomerates who would use Coca-Cola products are also on the board.

A similar pattern was observed even more prominently with PepsiCo, who actually had more sitting members on its board who are also active key executives in sales at major retail outlets or restaurants such a Wal-Mart and Burger King (of course Yum Brands also commands KFC and Pizza Hut itself):

Pepsi has some health-related board members such as the former President of the Ford Foundation and CEO of the Duke Health System on its board. We were somewhat bemused that Caspar Weinberger was on the board until his death (he was a defense man and military contracting expert involved in the Iran-Contra affair…which makes it a bit unclear what role he played on the executive board of a soda company). There are also several media-related members including top executives at Sun-Times and the New York Times.

Overall, the most interesting finding was that a lot of executive board members are…well…professional board members, it seems. Several people seem to be on the board of numerous companies simultaneously, and serve on many public boards as well. The PepsiCo CEO Indra Nooyi, for example, is also on the board of the Federal Reserve Bank of New York. The senior VP Larry Thompson also appears to be on the Board of Directors for the National Center for State Courts. Nearly every name was on at least two other boards, with several people on over 10 boards, many of them public foundations or government organizations. Many individuals were also connected to the recent Romney campaign (via standard donations or political action committees) but a few were related to the Obama campaign as well.

Danone was similar but Nestle was a bit more enigmatic. Very little information was available on its board via the NNDB software, apart from its CEO, who is again on many other prominent company boards:

The board members appeared elusive to standard news archive searches. When we individually searched the backgrounds of the key board members, we found they were also prominent members of other boards, including banks and oil companies, but also interestingly included Ann Veneman, the former Executive Director of UNICEF and former Secretary of the US Department of Agriculture, which is in charge of supplemental nutrition programs like food stamps.

Looking now at the monetary linkages rather than just the direct board members, we observed that the Coca-Cola companies appears to have four of its own political action committees (PACs) for donating money to political campaigns. They seem to lobby the most against tax legislation, spending about $3.8 million in lobbying but having peaked spending around the time of the “Child Nutrition Promotion and School Lunch Protection Act of 2009”, which would have updated national school nutrition standards but was killed in Congress. Their top political recipients can be found here.

Pepsi has slightly smaller lobbying of $2.2 million this year with 3 PACs, but also provided a parallel peak in 2009 around the time of the school lunch bill. Their top political recipients can be found here.

All in all, the picture is not very surprising, but the linkages at both the executive and political donor level appear to be unsubtle and tie the private sector to the public decision-making sector rather tightly as prominent decisions about soft drink regulation continue to be debated and voted on in the public sphere.