f, as many health professionals believe, our greatest public health challenge lies in stemming the rising human and financial burden of chronic disease, then developing a better way to identify and manage such illnesses is crucial.
Eduardo J. Simoes, professor and chair of MU’s Department of Health Management and Informatics, has spent much of his career working to further this elusive goal, developing “prioritization models” that are helping chronic-disease epidemiologists — those who investigate the causes and ways of controlling non-infectious illnesses — to plan and deploy interventions designed to keep us all healthier.
“This method of prioritizing public health programs, especially in prevention and health promotion, was created back in 1998,” says Simoes, who, while holding joint appointments as an adjunct professor at MU and associate professor at Saint Louis University, also served as chief epidemiologist for the Missouri Department of Health and Senior Services. “At that point, there was a lot of demand on the states to do better in working with local public health agencies.”
This task was anything but simple. Missouri has 115 independent public health agencies, one in each county. The cities of St. Louis, Kansas City and Joplin also have health departments, offices whose responsibilities often overlap with those of counties in their metro areas. Each of these agencies is charged with protecting its citizens from a dizzying array of health concerns and doing it with shrinking financial support from budget-challenged state coffers.
“As the chronic health epidemiologist working with the state on a whole host of chronic health issues, I felt the need to create something that would help them,” Simoes says.
imoes was well equipped to do the job. Born in Brazil, he earned a doctor of medicine degree at the University of Pernambuco in Recife before leaving his homeland to enroll in a graduate program in England. There he completed a master’s degree in community health for developing countries and a diploma in Tropical Medicine & Hygiene (DTM&H) at the prestigious London School of Hygiene and Tropical Medicine. Simoes later moved on to the School of Public Health at Emory University in Atlanta — an institution with close ties to the Centers for Disease Control and Prevention. While there, he added a master’s degree and a doctorate coursework in epidemiology to his long list of credentials.
Simoes began his career at the Missouri Department of Health and Senior Services in the mid 1990s. At the time, Bert Malone worked as the director of the Division of Chronic Disease Prevention and Health Promotion. Malone says Simoes’ research was already prioritizing the risk factors that burden a community with chronic disease.
“It allowed us to sharpen our focus on those areas where there was the greatest burden of disease or anticipated disease based on risk factors associated with those conditions,” Malone says. “And that’s the first time we really had any kind of a model that we could use to prioritize our limited funding. Chronic disease issues are definitely the future of public health.”
The CDC describes chronic disease as “among the most common, costly and preventable of all health problems in the U.S.” Such conditions include heart disease, stroke, cancer, diabetes and arthritis.
“These are all huge problems, and there’s no way that the resources that we have available to address these in a state like Missouri — or even in the nation — can drastically reduce the burden of those diseases unless we prioritize the top risk behaviors associated with those conditions and prevent disease rather than just treat the disease once it develops,” says Simoes.
Epidemiological studies seek to accomplish four key things: to identify the disease or health event; assess its rates of occurrence (who has been diagnosed with the condition, or has acquired the risk factors?); determine its prevalence (what percentage of a given population is suffering from or carrying it?); and, finally, its potential for mortality (among those affected or at risk for being affected, how many are dying or becoming sick?). Gathering these data not only allow epidemiologists to identify and contain outbreaks of infectious diseases — the 2002 outbreak of Severe Acute Respiratory Syndrome (SARS) for example — but also to gather detailed information on health risks related to more persistent and predictable factors in chronic illness: think tobacco use, unprotected sex, obesity, exposure to potentially harmful chemicals, and so on.
During his time working with the Missouri Department of Health and Senior Services, Simoes noticed a problem. When a county was confronted with an infectious outbreak, for instance, pertussis at an elementary school or norovirus at a church supper, both local and state officials moved quickly and efficiently to marshal the resources they needed to combat its spread. But when it came to addressing chronic illnesses and conditions, health officials were less likely to use the available data to maximize their disease-fighting resources.
Clearly, Simoes thought, something needed to be tweaked. The problem, he says, was not that the traditional way of identifying particular chronic disease problem wasn’t working. It just didn’t go far enough.
“It gives you a sense of the priority but doesn’t really give you the full flavor of priorities, because the comparisons make you look for a target that is perhaps not realistic for your county,” says Simoes. His model, on the other hand, uses six specific criteria — and a mountain of available epidemiological data on each one — to help local agencies size up their own situations: these include an evaluation of the size of the problem, its severity, the urgency of intervening and the likelihood that the intervention will have the desired effect, its cost, how much community support officials might expect, and, finally, whether there are disparities among those who suffer from the condition in race or other socioeconomic characteristics.
he model has now been in place in Missouri for close to a decade. It works in practice thanks to a simple digital interface. At a website hosted by the state, users of Priority MICA [Missouri Information for Community Assessment] enter locally relevant, health-prioritization information via an electronic form. When completed, they simply click “compute priorities” and, thanks to a database packed with a wide variety of surveillance system information, a ready-made roster of risk-factor and disease-priority rankings pops up.
Cash-strapped agencies thus not only gain a quantitative insight into their communities’ health needs, they can also initiate action even if statistical and epidemiological support isn’t readily available.
For Bert Malone, now public health administrator at the Kansas City Health Department, the system has been just what the doctor ordered.
“We use that model of risk behaviors to help us target initiatives that need to be implemented in our community,” he says. “For example, tobacco. Instead of looking just at the rates of cancer or cardiovascular disease, we now, using this model and the data that [Simoes] helped us have more readily available, focus on risk behaviors associated with those conditions, like smoking, or sedentary lifestyles or inadequate nutrition. We can focus our efforts on ultimately preventing those diseases and intervening earlier in the onset of those late-stage conditions once they do, in fact, develop.”
Throughout the United States, Simoes adds, similar opportunities exist for local agencies to target their efforts.
Because every state, county and community juggles different trends and needs in public health, Simoes says, local health officials are engaged in a balancing act with the health needs of a particular community and the interventions available. “By reviewing localized data and previous research that examined promotion and preventive interventions, public health officials can create and fund programs that target the most important issues in their communities.”
The model looks at which risk factors are important now and which ones will be an issue in the future. It takes the criteria laid out in the model to determine the severity and can weigh criteria differently depending on the needs of the local community.
“We have all these data for the states and data for the majority of the counties,” Simoes says. He adds that the CDC and state departments of health have made an effort to have risk factor data available to the counties based on the region or the Behavioral Risk Fact Surveillance System — an on-going telephone health survey system that, since 1984, has tracked certain health conditions and risk behaviors in the United States. Also, data on disease morbidity, hospitalization and mortality is available for all cities and counties in the nation.
“So prioritizing public health action based on risk factors or diseases using this model is doable for every one of the 3000-some counties in the United States,” Simoes says.
ollowing the successful implementation of the Priority MICA, Simoes was asked to help the Italian Ministry of Health develop something similar for that country. The result, details of which were published recently in the International Journal of Public Health, reflected the greater complexity of undertaking a prioritization system for an entire nation.
For the Italian project, Simoes and his hosts identified 15 specific risk factors then calculated priorities using a slightly modified set of criteria. These included: severity, the proportion of deaths attributable to the risk factors; magnitude, the prevalence of people engaging in the at-risk behavior; urgency, the increase or decrease in risk factor prevalence; disparity, the differences in risk-factor prevalence among different populations; effectiveness, how well are treatments and interventions working; and, finally, the anticipated cost of intervention.
“The system balances the influence of the six criteria to determine the top health priorities by modulating the 'severity’ criterion effect with that of the other five criteria. It also can accommodate more than six criteria and weight each criterion differently, if necessary,” says Simoes.
In his study in Italy, for example, Simoes determined that physical inactivity, cigarette smoking and hypertension should be the highest priorities for intervention programs in Italy. With this information, he says Italian public health officials can begin to develop prevention programs before a risk factor or condition becomes widespread. Already, Simoes says, Italian public health officials have used the model to identify geographical “hot spots” for alcohol prevention.
Implementation of his model in Italy has generated interest from public health agencies around the world, Simoes says. Among these is his birthplace, Brazil, where Simoes says plans are already in place to begin work with the Ministério da Saúde, the national Ministry of Health.
Simoes would like to see greater use of the prioritization approach because it may simplify the decision-making process. He cites prostate cancer as an example of how it would work.
“Back in 1998, I used prostate cancer to make the case with the Missouri Department of Health to say, listen, prostate cancer is a very important disease when you look at the incidence of death among men in the state of Missouri, especially if you’re older than 65,” Simoes says.
“Nevertheless, there’s no public health intervention that’s recommended to this day. So the priority ranking of prostate cancer will be lowest possible for the intervention effectiveness amongst all chronic diseases and conditions.” Why? Because, Simoes says, evidence-based public health data did not — and still doesn’t — indicate that doing mass screening for prostate cancer is beneficial.
“This model of prioritization and how it works is so simple that it hurts,” Simoes says. “Let’s look at how big the issue is and how fast it’s moving.”
Cardiovascular disease has been the chief killer of Americans for the past 40 or 50 years, he says. “But the trend is down for most cardiovascular-attributed deaths in past 10 years. Though its priority is still high when you look at the size of the disease, its urgency should not be as high because the death-related trend is going down.”
Urgency of action may also depend on levels of community support, or the lack thereof. Simoes cites smoking and gun violence to make the point.
Knowing that tobacco use is major cause of chronic illness, and that banning smoking inside restaurants and bars keeps people healthier, is not necessarily enough to recommend that a given public health agency devote a lot of time and money lobbying for a ban. If the available data indicate community members and their elected representatives in, say, a tobacco-producing area, are intractably opposed to tobacco restrictions, resources might best be directed elsewhere — at least in the short term.
The issue of deaths and injuries related to gun violence is even more fraught. “If a community doesn’t like some ideas about an intervention around gun control, for example, because most members support the National Rifle Association, there’s no way in the world I can get support from the mayor and community leadership in adopting any intervention that relates to … creating a policy that restrains guns,” Simoes says.
Before Simoes introduced his behavioral risk prioritization model, other models on risk behavior existed. These used national and international data on risk behaviors but, according to Bert Malone, Simoes’ version made it much easier to access and process local data.
“This was the first time we, in Missouri, could more easily access this data and get a better reflection of what the true burden of disease would be in our communities,” Malone says. In the past, the issues Simoes explored have been examined in isolation or, at best, using a couple of criteria. Simoes’ model, Malone says, does the job of pulling all relevant criteria together.
Simoes compares it to a self-exam. “It allows a county, state, or region to do their own prioritization without concern of another county or state’s status,” he says. “It’s like looking at yourself in the mirror and doing a self-examination of your neck or breast for potential breast cancer. This doesn’t concern itself with a comparison to anyone else. The model allows you to be different.”
In the end, he adds, this is what it all boils down to: Helping health administrators at every level use their own resources more wisely.
“The model allows you to identify hot areas where a risk factor priority score is highest,” he says. “It allows a county in Missouri to tell the CDC and the DHSS, ‘I think you need to allow us to take our scarce resources and put them into this area.’ By doing that we lower the local rates and reach our goals for the whole state, as opposed to spreading our resources across 115 different counties with minimal or no impact.”