Public HLT

CHAPTER 3

Public Health Data and Communications

© Orhan Cam/ShutterStock, Inc.

LEARNING OBJECTIVES

By the end of this chapter, the student will be able to:

•   identify six basic types of public health data.

•   explain the meaning, use, and limitations of the infant mortality rate and life expectancy measurements.

•   explain the meanings and uses of HALEs and DALYs.

•   identify criteria for evaluating the quality of information presented on a website.

•   explain ways that perceptions affect how people interpret information.

•   explain the roles of probabilities, utilities, and the timing of events in combining public health data.

•   explain the basic principles for the construction of decision trees and their uses.

•   explain how attitudes, such as risk-taking attitudes, may affect decision making.

•   identify three different approaches to clinical decision making and their advantages and disadvantages.

You read that the rate of use of cocaine among teenagers has fallen by 50% in the last decade. You wonder where that information might come from.

You hear that life expectancy in the United States is now approximately 80 years. You wonder what that implies about how long you will live and what that means for your grandmother, who is 82 and in good health.

You hear on the news the gruesome description of a shark attack on a young boy from another state and decide to keep your son away from the beach. While playing at a friend’s house, your son nearly drowns after falling into the backyard pool. You ask why so many people think that drowning in a backyard pool is unusual when it is far more common than shark attacks.

“Balancing the harms and benefits is essential to making decisions,” your clinician says. The treatment you are considering has an 80% chance of working, but there is also a 20% chance of side effects. “What do I need to consider when balancing the harms and the benefits?” you ask.

You are faced with a decision to have a medical procedure. One physician tells you there’s no other choice and you must undergo the procedure, another tells you about the harms and benefits and advises you to go ahead, and the third lays out the options and tells you it’s your decision. Why are there such different approaches to making decisions these days?

 

© Kvadrat/ShutterStock, Inc.

These are the types of issues and questions that we will address as we look at health data and communications.

WHAT IS THE SCOPE OF HEALTH COMMUNICATIONS?

The term health communications deals with the methods for collecting, compiling, and presenting health information. It also addresses how we perceive information, combine information, and use information to make decisions. Thus, health communication is about information, from its collection to its use. Figure 3-1 displays how these parts of the process fit into a continuous flow of information.

The field of health communications has been growing at the speed of the Internet. This field has implications for most, if not all, aspects of public health, as well as health care. Therefore, we will focus on key issues in each of the above-mentioned components of this burgeoning field. We will look at the following aspects of health communications and ask the following questions:

•   Collecting data: Where does public health data come from?

•   Compiling information: How is public health information compiled or put together to measure the health of a population?

•   Presenting information: How can we evaluate the quality of the presentation of public health information?

•   Perceiving information: What factors affect how we perceive public health information?

FIGURE 3-1 Public Health Data, Health Communication, and the Flow of Information

 

•   Combining information: What types of information need to be combined to make health decisions?

•   Decision making: How do we utilize information to make health decisions?

We can only highlight key issues in this complex and evolving field of health communications. To do this, we will use the above questions and provide frameworks and approaches to explore possible answers.

WHERE DOES PUBLIC HEALTH DATA COME FROM?

Public health data is collected in a wide variety of ways. a These methods are often referred to as public health surveillance or surveillance. Data from public health surveillance is collected, published, and distributed without identifying specific individuals. Data of this type come from a growing variety of sources. It is helpful, however, to classify these sources according to the way they are collected. Table 3-1 outlines common types of quantitative public health data, provides examples of each type, and indicates important uses, as well as the advantages and disadvantages of each type of data.

Data from different sources is increasingly being combined to create integrated health data systems or databases that can be rapidly and flexibly accessed by computers to address a wide range of questions. These systems have great potential to provide useful information to contribute to evidence-based public health. This information can help describe problems, examine etiology, assist with evidence-based recommendations, and examine the options for implementation, as well as help evaluate the outcomes. Despite their great potential, integrated databases also create the potential for abuse of the most intimate health information. Thus, protecting the privacy of data and ensuring its anonymous collection and distribution is now of great concern as part of the development of integrated databases.

Public health data has traditionally been numerical, or quantitative. In recent years, the importance of nonnumerical, or qualitative, data has received much greater attention in public health. Box 3-1 discusses some of the important uses of qualitative data.

Data can be used for a wide range of purposes in public health and health care. One particularly important use is the compilation of data to generate summary measurements of the health of a group or population. Let us take a look at how we compile this data.

HOW IS PUBLIC HEALTH INFORMATION COMPILED TO MEASURE THE HEALTH OF A POPULATION?

Measurements that summarize the health of populations are called population health status measures. For over a century, public health professionals have focused on how to summarize the health status of large populations, such as countries and large groups within countries—for example, males and females or large racial groups of a particular nation. In the 1900s, two measurements became standard for summarizing the health status of populations: the infant mortality rate and life expectancy. These measurements rely on death and birth certificate data, as well as census data. Toward the latter part of the 1900s, these sources of data became widely available and quite accurate in most parts of the world.

The infant mortality rate estimates the rate of death in the first year of life. For many years, it has been used as the primary measurement of child health. Life expectancy has been used to measure the overall death experience of the population, incorporating the probability of dying at each year of life. 1 ,2 These measures were the mainstay of 20th century population health measurements. Let us look at each of these measures and see why additional health status measurements are needed for the 2000s.

At the turn of the 20th century, infant mortality rates were high even in today’s developed countries, such as the United States. It was not unusual for over 100 of every 1,000 newborns to die in the first year of life. In many parts of the world, infant mortality far exceeded the death rate in any later years of childhood. For this reason, the infant mortality rate was often used as a surrogate or substitute measure for overall rates of childhood death. In the first half of the 1900s, however, great improvements in infant mortality occurred in what are today’s developed countries. During the second half of the century, many developing countries also saw greatly reduced infant mortality rates. Today, many countries have achieved infant mortality rates below 10 per 1,000, and a growing number of nations have achieved rates below 5 per 1,000. b

The degree of success in reducing mortality among children aged 2 to 5 has not been as great. 3 Malnutrition and old and new infectious diseases continue to kill young children. In addition, improvements in the care of severely ill newborns have extended the lives of many children—only to have them die after the first year of life. Children with HIV/AIDS often die not in the first year of life, but in the second, third, or fourth year. Once a child survives to age 5, he or she has a very high probability of surviving into adulthood in most countries. Thus, a new measurement known as under-5 mortality has now become the standard health status measure used by the World Health Organization (WHO) to summarize the health of children.

TABLE 3-1 The 6 S’s of Quantitative Sources of Public Health Data

Let us take a look at the second traditional measure of population health status—life expectancy. Life expectancy is a snapshot of a population incorporating the probability of dying at each age of life in a particular year. Life expectancy tells us how well a country is doing in terms of deaths in a particular year. As an example, life expectancy at birth in a developed country may be 80 years. Perhaps in 1900, life expectancy at birth in that same country was only 50 years. In 2020, life expectancy may be 85 years. Thus, this metric allows us to make comparisons between countries and within a single country over time.

BOX 3-1 Qualitative Data and Its Importance to Public Health

Qualitative data can serve a variety of functions in public health. It can generate ideas or hypotheses for further study, provide key information on the reasons for success or failure of an intervention, and provide explanations for findings that cannot be derived from quantitative research. Thus, qualitative research can contribute to the full range of health communications, from data collection to decision making. Qualitative and quantitative research should be seen as complementary, not competitive, as they can work together to provide greater insight and understanding. Quantitative research often includes a large sample and focuses on numbers, whereas qualitative research often looks in depth at a small sample, producing descriptions and allowing for a thorough exploration of the phenomenon of interest.

Focus groups and interviews are increasingly important forms of qualitative research that are being used to gain insight into how and why people come to conclusions or hold opinions. The opinions examined increasingly go beyond commercial products and politicians to include the use of health services, acceptance of new and existing technology, and speculation about the reasons for diseases and health outcomes. The ideas put forward by these types of investigations may generate new hypotheses to be examined using quantitative studies. They may also help assess barriers to implementation and suggest new approaches. Focus groups and interviews are useful when the behavior, thoughts, and interpretations cannot be observed directly but the participants can provide pertinent information on the topic of interest.

Document review and observations are another important method used in qualitative research. Document review includes collection and analysis of existing documents in order to provide context for the issue being explored. Because the documents were developed for purposes other than research, they tend to be objective, shedding light on the culture of the population being studied and revealing intentions and behaviors of the population. Observation can be particularly useful when it is important to understand the real-time behavior of a population in its own environment.

These and other approaches to qualitative research allow us to understand, describe, and explore public health issues in a flexible way to produce richly descriptive and comprehensive findings.

Despite its name, life expectancy cannot be used to accurately predict the future even at the population level—that would require assuming that nothing will change. That is, it assumes that the death rates at all ages will remain the same in future years. We have seen increases in life expectancy in most countries over the last century, but declines occurred in sub-Saharan Africa and countries of the former Soviet Union in the late 1900s. c

Life expectancy tells us only part of what we want to know. It reflects the impact of dying, but not the impact of disabilities. When considering the health status of a population in the 2000s, we need to consider disability, as well as death.

Today, the World Health Organization (WHO) uses a measurement known as the health-adjusted life expectancy (HALE) to summarize the health of populations. 4 The HALE measurement starts with life expectancy and then incorporates measurements of the quality of health. The WHO utilizes survey data to obtain a country’s overall measurement of quality of health. This measurement incorporates key components, including:d

•   Mobility—the ability to walk without assistance

•   Cognition—mental function, including memory

•   Self-care—activities of daily living, including dressing, eating, bathing, and use of the toilet

•   Pain—regular pain that limits function

•   Mood—alteration in mood that limits function

•   Sensory organ function—impairment in vision or hearing that impairs function

From these measurements, an overall quality of health score is obtained. In most countries, these range from 85 to 90%. We may consider a score of less than 85% as poor and greater than 90% as very good. A quality of health measurement of 90% indicates that the average person in the country loses 10% of his or her full health over his or her lifetime to one or more disabilities.

TABLE 3-2 Life Expectancy and Health-Adjusted Life Expectancy for a Range of Large Countries

Country	Life expectancy	Health-adjusted life expectancy (HALE)
Nigeria	49	42
India	64	56
Russian     Federation	66	60
Brazil	73	64
China	74	66
United States	78	70
United Kingdom	80	72
Canada	81	73
Japan	83	76

Data from World Health Organization. Healthy life expectancy (HALE) at birth (years). World health report 2009. http://www.who.int/whosis/whostat/EN_WHS09_Full.pdf . Accessed July 17, 2013.

The quality of health measurement is multiplied by the life expectancy to obtain the HALE. Thus, a country that has achieved a life expectancy at birth of 80 years and an overall quality of health score of 90% can claim a HALE of 80.00 × 0.90 = 72.00. Table 3-2 displays WHO data on life expectancy and HALEs at birth for a variety of large countries. e

Today, the under-5 mortality and HALEs are used by the WHO as the standard measures reflecting child health and the overall health of a population. An additional measure, known as the disability-adjusted life year (DALY), has been developed and used by the WHO to allow for comparisons and changes based on categories of diseases and conditions. 5 Box 3-2 describes DALYs and some of the data and conclusions that have come from using this measurement. Table 3-3 displays DALYs according to these categories of diseases and conditions for the same large countries for which HALEs are displayed in Table 3-2 .

The Global Burden of Disease (GBD) project has produced a number of important conclusions using DALYs, including:

•   Depression is a major contributor to most nations’ DALYs and may become the number one contributor in the next few decades in developing, as well as developed, countries.

•   Chronic disabling diseases, including hookworm, malaria, and HIV, affect the young and working-age population and are the greatest contributors to the burden of disease in many developing countries.

•   Cancers, such as breast cancer, hepatomas (primary liver cancer), and colon cancer—which affect the working-age population and are common in many developing countries—have an important impact on the burden of disease as expressed in DALYs.

•   Motor vehicle, occupational, and other forms of unintentional injuries have a disproportionate impact on the burden of disease compared to merely measuring deaths because these injuries produce long-term disabilities, as well as death at young ages.

•   Obesity is rapidly overtaking malnutrition as a burden of disease in developing countries as early onset diabetes, heart disease, and strokes become major causes of death and disability among younger populations.

We have now looked at important sources of public health data and examined one key way that data is compiled to generate population health status measurements. Now, let us look at a third issue: the presentation of public health information.

HOW CAN WE EVALUATE THE QUALITY OF THE PRESENTATION OF HEALTH INFORMATION?

Having information is not enough. A key role and essential tool of public health is to effectively present the information in ways that serve as a basis for understanding and decision making. Issues of information presentation are increasingly important and increasingly complex. They require the study of a range of disciplines, from mass media, to computer graphics, to statistics. f Public health information is often presented as graphics. Graphics create a picture in our mind of what is going on, and a picture is truly worth a thousand words. Graphical presentations can accurately inform, but they can also mislead us in a wide variety of ways. The accurate presentation of visual information has become an art as well as a science that deserves attention from all those who use information. 6

BOX 3-2 DALYs

Disability-adjusted life years (DALYs) are designed to examine the impacts that specific diseases and risk factors have on populations, as well as provide an overall measure of population health status. They allow comparisons between countries or within countries over time, based not only on overall summary numbers, such as life expectancy and HALEs, but also on specific diseases and risk factors.

The DALY compares a country’s performance to the country with the longest life expectancy, which is currently Japan. Japan has a life expectancy that is approximately 83 years. In a country with zero DALYs, the average person would live approximately 83 years without any disability and would then die suddenly. Of course, this does not occur even in Japan, so all countries have DALYs of greater than zero. The measurement is usually presented as DALYs per 1,000 population in a particular country. a

Calculations of DALYs require much more data on specific diseases and disabilities than other measurements, such as life expectancy or HALEs. However, the WHO’s Global Burden of Disease (GBD) project has made considerable progress in obtaining worldwide data collected using a consistent approach. 5 Data is often not available on the disability produced by a disease. The WHO then uses expert opinion to estimate the impact.

The GBD project presents data on DALYs divided into the following categories:

•   Communicable disease; maternal, neonatal, and nutritional conditions

•   Noncommunicable diseases

•   Injuries

Data is also available on specific diseases and risk factors, such as the impact of cigarette smoking, alcohol use, or depression.

It is also important to recognize that DALYs require a number of policy decisions that are hidden in the numbers. For instance, it was decided to emphasize the importance of death and disability among those of working age by giving them greater weight or importance in the calculation of this measurement. Working age was defined as age 16–60, which reflects a concept of working age more often used in developing countries. Death and disability at ages greater than the approximately 83 years of life expectancy of Japan do not add to the DALYs. In calculating DALYs, separate maximum life expectancies are used for males (~80 years) and females (~85 years), implying that loss of life among females is given slightly greater importance. These issues illustrate that in order to understand quantitative measures such as DALYs, you need to recognize that policy decisions are often subtly integrated into quantitative measurements. We need to appreciate the policy and sometimes the ethical decisions that are part of what appears to be objective measurements.

_______________

a The DALY is a complex and technical measurement. If, in a country with 0 DALYs, 1,000 newborns suddenly died, there would be a loss of as much as 83,000 DALYs from the death of these 1,000 newborns. Thus, the total DALYs a country can lose in a particular year can range from 0 to approximately 83,000 per 1,000 persons. This somewhat overstates the possible loss due to the discounting and weighing that occurs in the calculation of DALYs. Nonetheless, when interpreting a country’s total DALYs, it may be useful to compare the number of DALYs to this maximum possible loss.

TABLE 3-3 DALYs Lost by Disease Categories and Total of All Categories Per 1,000 Population

Data from World Health Organization. Global Burden of Disease Project 2004. Geneva: World Health Organization; 2004.

Issues of quality are key to the presentation of information. The Internet is increasingly the primary source of public health information for the user. Thus, when we address issues of quality, we need to have a set of criteria for judging the quality of information presented on the Internet. Before relying on a website for health information, you should ask yourself key questions. 7 These questions are summarized in Table 3-4 . Try these out the next time that you view a health information website.

The presentation of health information also requires taking into account the audience who will be using the material. Understanding the degree of health literacy of the intended audience is so important that a national movement has developed to address these issues. Health literacy is more than the ability to read. It refers to the degree to which individuals have the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions. 8

Even the most accurate data presentation does not tell us how the user will perceive the data. Let us take a look at the rapidly growing component of health communications that deals with how we perceive information.

WHAT FACTORS AFFECT HOW WE PERCEIVE PUBLIC HEALTH INFORMATION?

Regardless of how accurately information is presented, communication also needs to consider how the recipient perceives the information. Therefore, we also need to look at factors known to affect the perception of information or the subjective interpretation of what the information means for an individual.

At least three types of effects can greatly influence our perceptions of potential harms and benefits. 9 We will call them the dread effect, the unfamiliarity effect, and the uncontrollability effect.

TABLE 3-4 Quality Standards for Health Information on the Internet

Criteria	Questions to ask
Overall site quality	•   Is the purpose of the site clear?
	•   Is the site easy to navigate?
	•   Are the site’s sponsors clearly identified?
	•   Are advertising and sales separated from health information?
Authors	•   Are the authors of the information clearly identified?
	•   Do the authors have health credentials?
	•   Is contact information provided?
Information	•   Does the site get its information from reliable sources?
	•   Is the information useful and easy to understand?
	•   Is it easy to tell the difference between fact and opinion?
Relevance	•   Are there answers to your specific questions?
Timeliness	•   Can you tell when the information was written?
	•   Is it current?
Links	•   Do the internal links work?
	•   Are there links to related sites for more information?
Privacy	•   Is your privacy protected?
	•   Can you search for information without providing information about yourself?

Data from American Public Health Association. Criteria for Assessing the Quality of Health Information on the Internet. Available at http://www.apha.org/NR/exeres/412DDD11-42A0-46CB-8705-F9D47A073AF1.htm Accessed July 17, 2013.

The dread effect is present with hazards that easily produce very visual and feared consequences. It explains why we often fear shark attacks more than drowning in a swimming pool. The dread effect may also be elicited by the potential for catastrophic events, ranging from nuclear meltdowns to a poisoning of the water supply.

Our degree of familiarity with a potential harm or a potential benefit can greatly influence how we perceive data and translate it for our own situation. Knowing a friend or relative who died of lung cancer may influence how we perceive the information on the hazards of smoking or the presence of radon. It also may explain why we often see the danger of sun exposure as low and food irradiation as high, despite the fact that the data indicate that the degree of harm is the other way around.

Finally, the uncontrollability effect may have a major impact on our perceptions and actions. We often consider hazards that we perceive as in our control as less threatening than ones that we perceive as out of our control. Automobile collisions, for instance, are often seen as less hazardous than commercial airplane crashes, despite the fact that statistics show that commercial air travel is far safer than travel by automobile.

Perception of bad outcomes (or harms) and good outcomes (or benefits) needs to be considered along with the numbers if we are going to understand the ways information is used to make decisions. Not everyone perceives harms and benefits the same way. The selection of accurate and effective methods for conveying data is key to health communications. g

Understanding how we perceive information can help us design effective health messages. Box 3-3 discusses the SUCCESs approach to developing messages that stick.

One approach to addressing differing perceptions of information is the use of a method known as decision analysis. Decision analysis relies on the vast information-processing ability of computers to formally combine information on benefits and harms to reach quantitative decisions. It provides us with insight into the types of information that need to be combined. Let us look at how we combine information—the next question in our flow of health information.

WHAT TYPE OF INFORMATION NEEDS TO BE COMBINED TO MAKE HEALTH DECISIONS?

Decision analysis focuses on three key types of information that need to be combined as the basis for making decisions. We can better understand these types of information by asking the following questions:

•   How likely?—What is the probability or chance that the particular outcome will occur?

•   How important?—What is the value or importance we place on a good or a bad outcome?

When expressing the chances that an outcome will occur, we often express the results as a percentage from 0 to 100. Probabilities, on the other hand, range from 0 to 1. Percentages and probabilities are often used interchangeably—the probability of 0.10 can be converted to 10% and vice versa. When faced with a percentage or probability, we need to ask: What period of time is being considered? For instance, if you hear that the chances of developing a blood clot while taking high-dose estrogen birth control pills is 5%, what does that mean? Does it mean 5% per cycle, 5% per year, or 5% over the time period that the average user is on the pill?

Outcomes vary from death to disabilities. Some outcomes greatly affect our function and limit our future, while we can learn to live with other outcomes despite the limitations they impose. When dealing with a quantitative approach, we are forced to place numbers on the value or importance of specific outcomes. A scale known as a utility scale is one method to measure and compare the value or importance that different people place on different outcomes. This scale is intended to parallel the scale of probabilities; that is, it extends from 1 to 0 or from 100% to 0%. It defines 1 or 100% as the state of health in which there are no health-related limitations. Zero is defined as immediate death. On the utility scale, there is nothing worse than immediate death. Figure 3-2 displays the utility scale. h Box 3-4 illustrates how we can use the utility scale to assign numbers to specific outcomes.

Utilities are important, especially when we need to combine potential harms with potential benefits. Probabilities alone often do not give us the answers we need when addressing issues of hazards ranging from environmental toxins to unhealthy behaviors. Utilities are also critical when looking at particular interventions, such as prevention or treatment options that include positive benefits, but also involve side effects or harms. Thus, whenever we need to combine or balance benefits and harms, we need to consider the utility of the outcomes along with the chances or probabilities of the outcomes.

BOX 3-3 SUCCESs in Public Health Communications

Effective health communications starts with understanding how information is perceived. In their book Made to Stick: Why Some Ideas Survive and Others Die, 10 Chip and Dan Heath have come up with a memory technique they call SUCCESs, which focuses on the perception of ideas and identifies six principles of highly successful communications. SUCCESs stands for:

Simplicity: This first principle requires a short, memorable statement that captures the core of the message. The golden rule, the authors write, “is the ultimate model of simplicity: a one-sentence statement so profound that an individual could spend a lifetime learning to follow it.” 11 While public health messages cannot be expected to rival the golden rule, some public health messages say it all. The Back-to-Sleep campaign, for instance, was able to convey the core of its message in just three words.

Unexpectedness: Getting and holding people’s attention is often achieved by presenting unexpected facts that are counterintuitive, at least to your audience. Challenging common myths or conventional wisdom may be a good place to start when engaging an audience.

Concreteness: Proverbs often provide specific examples that can be remembered and generalized. For instance, “an apple a day keeps the doctor away” has become a memorable way of conveying the importance of diet in health. Providing concrete, visualizable examples is key. Bad breath and brown teeth may be more convincing reasons for stopping cigarette smoking than the long-term consequences, which are not immediately obvious.

Credibility: Credibility relies not so much on numbers but rather on the source of the information. For instance, news of an epidemic may start with, “Today, the CDC announced…” Credibility is enhanced if people can test out the ideas from their own experience. “Think about the last time you texted while driving. Could it have waited until you stopped?”

Emotions: Connecting with people’s emotions is key not only in getting their attention but also for ensuring they will retain the ideas. Emotions connect people with ideas. For instance, the Heaths write, “It’s difficult to get teenagers to quit smoking by instilling in them a fear of the consequences, but it’s easier to get them to quit by tapping into their resentment of the duplicity of Big Tobacco.” 12

Stories: We remember and relate to stories about real or realistic people. Sharing “war stories” is a classic example of how people relate to the events in each other’s lives. Short vignettes and stories help to make the issues real. Hopefully, the vignettes at the beginning of each chapter of Public Health 101 accomplish this for you.

Putting the SUCCESs principles together as a coherent message is a real but important challenge. As the ideal example, the Heaths cite John F. Kennedy’s famous challenge to “put a man on the moon and return him safely by the end of the decade.” 13 They conclude that the message is simple, unexpected, amazingly concrete, credible because it is from the president of the United States, full of emotion, and a story in miniature.

It is not so easy to put together such memorable messages, but focusing on how information is perceived and using the SUCCESs principles is a good way to start.

Data from Heath C and Heath D. Made to Stick: Why Some Ideas Survive and Others Die. New York, NY: Random House; 2007.

FIGURE 3-2 Scale Used to Measure Utilities

 

Probabilities and utilities (both on a scale of 0 to 1) are often combined by multiplying the probability by the utility to obtain a probability that takes into account the utility or what is called expected utility. Expected utilities are often displayed using graphical methods called decision trees. 14 Box 3-5 discusses the use of decision trees based on expected utilities.

BOX 3-4 Obtaining a Utility Score

Let us see how we can use the utility scale to put numbers on a specific outcome: complete and permanent blindness. Using the scale in Figure 3-2 , place a number on the importance or value that you give to complete and permanent blindness.

In large groups of individuals, the average utility placed on blindness is quite predictable—about 50%. However, the range of values among a group is generally quite wide ranging, from 20 to 80% and sometimes even wider. Predicting an individual’s utility is quite difficult because gender, socioeconomic group, and other predictors have little impact. a

Individuals who place a high utility on complete and permanent blindness usually indicate that they can learn to live with blindness and it will not greatly affect their enjoyment of life. Those who place a low utility on blindness generally say just the opposite. Thus, we need to understand that a utility of 50% is an average, including some with a much higher and some with a much lower utility. Therefore, the best way to know the value or utility that an individual places on a particular outcome such as blindness is to ask him or her.

_______________

a There are at least two predictors that are of some value. Those who have experienced an outcome usually find that they can adapt to it to a certain extent and usually rate its utility as somewhat higher than those who have not experienced the outcome. Second, age does have an impact on the scoring of utility. Younger people generally rate the utility of an outcome as somewhat worse or lower than older people, perhaps due to the longer-term impact the disability has on their future options. The average utility placed on blindness by college students, for instance, is often closer to 40%. Neither of these impacts is large on average, nor can they be used to successfully predict the utility of any one individual.

WHAT OTHER DATA NEEDS TO BE INCLUDED IN DECISION MAKING?

We also need to ask:

How soon?—When, on average, will the particular outcome happen if it is going to happen?

The expected timing of the occurrence of good and bad outcomes can also affect how we view the outcome. Most people view the occurrence of a bad outcome as worse if it occurs in the immediate future compared to years from now. Conversely, we usually view a good outcome as more valuable if it occurs in the immediate future. Thus, whenever we consider harms and benefits and try to combine them, we need to ask: When are the outcomes expected to occur? When both the good and the bad outcomes occur in the immediate future, the timing is not an issue. In public health and medicine, however, this is rarely the case. When dealing with many treatments, the benefits come first while the harms may occur at a later time. When dealing with vaccines and surgery, the pain and side effects often precede the potential gain. The timing of the benefits is rarely the same as the timing of the harms. Thus, we need to take this into account. This process is known as discounting. Discounting is a quantitative process in which we give greater emphasis or weight to events that are expected to occur in the immediate future compared to events that are expected to occur in the distant future. i

We have seen that probabilities, utilities, and timing are key components of health communications that need to be combined when making public health and healthcare decisions. j However, there are other factors that are characteristic not of the data itself, but of the decision maker. A decision maker may be an individual; a health professional; or an organization, such as a nonprofit, a corporation, or a government agency. Let us turn our attention to decision makers and ask how we can go about making decisions. To do this, we need to address issues beyond probability, utility, and timing.

BOX 3-5 Using Decision Trees to Compare Interventions

Decision trees are a visual method for displaying the benefits and harms of two or more options for intervention. They allow us to directly compare the outcomes, incorporating the probability and the utility of each outcome in a process known as decision analysis. Decision trees are made up of two types of nodes, which reflect points in which decisions are made or events occur by chance. Therefore, we speak of choice nodes and chance nodes. As indicated in Figure 3-3 , choice nodes are presented using a square box, while chance nodes are represented using a circle.

Let us see how choice nodes and chance nodes can be put together to develop a decision tree. Figure 3-4 represents a simple decision tree. For each of the outcomes, a probability is included.

Note that the probability of each of the potential outcomes adds to 1 or 100%. For intervention #1, the potential outcomes are cure and die. For intervention #2, there is a third potential outcome: blindness. To compare intervention #1 and intervention #2, we need to know more than the probabilities of each outcome—we need to know the utility of blindness.

Thus, to compare intervention #1 and intervention #2, we will need to assume that die has a utility of 0 and cure has a utility of 1. Blindness is a more subjective utility. As we have seen, on average, it has been found that people regard blindness as having a utility of about 0.5. However, many people will have a utility that is as high as 0.8 or as low as 0.2. Let us start by using a utility of 0.5. In Figure 3-5 , the probabilities and utilities of each intervention have been filled in, and the probability has been multiplied by the utility to produce what is called an expected utility, or a probability that takes into account the utility of the outcomes. We can add together the expected utilities of each potential outcome to produce an overall expected utility. The overall expected utilities allow us to compare one intervention to another.

FIGURE 3-3 Choice Node and Chance Node

 

Notice that when we use a utility of 0.5, the overall expected utilities for the two interventions are the same. At least by decision analysis, these two potential interventions are considered a toss-up.

Now let us see what happens when we change the utility of blindness first to 0.8 and then to 0.2. Figure 3-6 displays the decision tree and expected utilities when blindness’s utility is set at 0.8. Figure 3-7 displays the decision tree and expected utilities when blindness’s utility is set at 0.2.

Notice that when the utility is set at 0.8 in Figure 3-6 , the overall expected utility for intervention #2 is greatest. That is, the decision analysis recommends intervention #2 over intervention #1. However, in Figure 3-7 , when the utility of blindness is set at 0.2, the overall utility for intervention #1 is greater than intervention #2. That is, the decision analysis recommends intervention #1 over intervention #2. When a factor, such as the utility we place on blindness, produces a change in the recommended choice of intervention, we say that the decision analysis is sensitive to the factor, such as blindness. a

FIGURE 3-4 Decision Tree

 

FIGURE 3-5 Decision Analysis with Utility of Blindness Equal to 0.5

 

FIGURE 3-6 Decision Analysis with Utility of Blindness Equal to 0.8

 

FIGURE 3-7 Decision Analysis of Utility of Blindness Equal to 0.2

 

Decision trees and decision analysis are increasingly being used to display the options for intervention and to compare them based on probabilities and utilities. Decision trees are increasingly used as part of public health decision making for populations and for health policy decisions that affect large groups. Decision trees are often far more complex in an effort to reflect the realities of decision making. Decision analysis can help us compare interventions and help us recognize why individuals or organizations may come to different conclusions about their preferred intervention. b

_______________

a Decision trees assume that an outcome is measured by its probability multiplied by its utility. Thus, expected utility focuses on outcomes, not on the process of getting there. Therefore, this approach considers an outcome such as death or blindness to be the same whether it occurs suddenly or after a complicated hospitalization with multiple unsuccessful interventions. Also note that the example used does not take into account the timing of the outcome. It is possible in decision analysis to take into account the timing of outcomes through the process of discounting. Decision analysis assumes that when two chance nodes appear one after another, the probability of the second outcome is not affected by the outcome of the first node; that is, the chances of a good or bad outcome after the first and second chance nodes are independent of each other. The decision trees used in this box are also simplified in that only one choice node is presented. It is possible to introduce choice nodes even after a chance node. Decision trees may become very complex and may need to become so in order to realistically reflect the choice and chance situations faced in practice. However, the more complex the decision tree, the more data is needed to utilize it to make recommendations.

b Decision analysis, which compares interventions for two or more conditions, frequently utilizes a measure of outcome known as quality adjusted life-years, or QALYs. QALYs incorporate the increase or decrease in life-expectancy as well as the probabilities and utilities of each outcome. QALYs may be used in decision analysis, but they are expected for cost-effectiveness analysis.

HOW DO WE UTILIZE INFORMATION TO MAKE HEALTH DECISIONS?

There are two key questions that we can ask to gain an understanding of how we use information to make health decisions:

•   How do our risk-taking attitudes affect the way we make decisions?

•   How do we incorporate information into our decisions?

There are a large number of attitudes that can affect the way we make decisions. One of the most important is known as our risk-taking attitudes. k

Let us examine what we mean by this term and see what type of risk-taking attitude you use in making decisions. Attitudes toward risk greatly influence the choices that we all make in the prevention and treatment of disease. 9 Box 3-6 illustrates how you can understand your own attitudes toward risk taking by making some choices. We will assume that you understand what we mean by “utilities” and that you have thought through what a wide range of utilities means to you personally.

Understanding attitudes toward risk is important for analyzing how individuals make decisions about their own lives. It is also key when trying to understand how group decisions are made that require society to balance harms and benefits. Perhaps the most common health decisions that you will make are the decisions related to your health care and that of your family. Therefore, let us complete our examination of health communications by looking at three different approaches that can be used to make clinical healthcare decisions.

HOW CAN WE USE HEALTH INFORMATION TO MAKE HEALTHCARE DECISIONS?

There are three basic approaches to using health information to make healthcare decisions. We will call these approaches inform of decision, informed consent, and shared decision making. Preferences for these types of approaches have changed over time, yet all three are currently part of clinical practice.

The inform of decision approach implies that the clinician has all the essential information and can make decisions that are in the patient’s best interest. The role of the clinician is then merely to inform the patient of what needs to be done and to prescribe the treatment, or write the orders. At one point in time, this type of decision-making approach was standard for practicing clinicians. In the not-too-distant past, clinicians rarely told patients that they had cancer, justifying their silence by arguments that the knowledge might make the patient depressed, which could interfere with their response to the disease and to the treatment. The decision to administer many tests and receive a range of medications is still often done using the inform of decision approach.

A second type of decision-making approach is called informed consent. It rests on the principle that ultimately, patients need to give their permission or consent before major interventions, such as surgery, radiation, or chemotherapy, can be undertaken. Informed consent may be written, spoken, or implied. Clinically, informed consent implies that individuals have the right to know what will be done, why it will be done, and what the known benefits and harms are. Patients have the right to ask questions, including inquiring about the availability of other options. Informed consent does not mean that all possible options are presented to the patient, but it does imply that a clinician has made a recommendation for a specific intervention.

The third type of decision making is called shared decision making. In this approach, the clinician’s job is to provide information to the patient with which he or she can make a decision. This might include directly giving information to the patient; providing consultations; or referring patients to sources of information, often on the Internet. Shared decision making places a far greater burden on the patient to seek out, understand, and use information. Using this approach, clinicians are not required to provide recommendations on specific interventions, though patients are free to ask for a clinician’s opinion. 15

All three types of decision-making approaches are currently in use today. Table 3-5 outlines the process and roles implied by each of these approaches, as well as some of the potential advantages and disadvantages of each approach.

Health communications provides key tools for population health. We have taken a look at important issues related to each of them. We have asked questions about how public health data and information is collected, compiled, presented, perceived, combined, and used in decision making. Data and information are key public health tools for guiding our decision making. We will find ourselves coming back again and again to these principles as we study the population health approach. Now, let us turn our attention to the utilization of the social and behavioral sciences as key tools of public health.

BOX 3-6 Risk-Taking Attitudes

Review the following situations and write down your decisions.

Situation A

Imagine that you have coronary artery disease and have a reduced quality of life with a utility of 0.80, compared to your previous state of full health with a utility of 1.00. You are offered the following pair of options. You can select only one option. Which of the following options do you prefer?

OPTION #1: A treatment with the following possible outcomes:

50% chance of raising the quality of your health (your utility) from 0.80 to 1.00

50% chance of reducing the quality of your health (your utility) from 0.80 to 0.60

OPTION #2: Refuse the above treatment and accept a quality of your health (your utility) of 0.80

Situation B

Imagine that you have coronary artery disease and have a reduced quality of life that has a utility of 0.20, compared to your previous state of full health that had a utility of 1.00. You are offered the following pair of options. You can select only one option. Which of the following options do you prefer?

OPTION #1: A treatment with the following possible outcomes:

10% chance of raising the quality of your health (your utility) from 0.20 to 1.00

90% chance of reducing the quality of your health (your utility) from 0.20 to 0.11

OPTION #2: Refuse the above treatment and accept a quality of your health (your utility) of 0.20

What was your answer in situation A? Situation B? To understand the meaning of your answers, you need to appreciate that in terms of the probabilities and utilities presented in each situation, these options are a toss-up. That is, taking into account the probabilities and the utilities, there is no difference between these options. To convince yourself of this, draw a decision tree including the two options in situation A and the two options in situation B. You will find that they produce the same overall expected utilities. a

Thus, the information does not determine your choice; it must be your attitude toward taking chances, which is your attitude toward risk taking.

Did you choose option #2 in situation A and option #1 in situation B? Most, but not all, people make these choices. In situation A, we begin with a utility of 0.80. For many people, this is a tolerable situation and they do not want to take any chances of being reduced to a lower, perhaps intolerable utility. Thus, they want to guarantee a tolerable level of health. We can call this the certainty effect. In situation B, we begin with a utility of 0.20. For many people, this is an intolerable situation. Thus, people are usually willing to take their chances of getting even worse in the hopes of a major improvement in their health. When the quality of life is bad enough, most, if not all, people are willing to take their chances and go for it. This risk-taking behavior can be called the long-shot effect. Thus, risk-taking and risk-avoiding choices are both common, defensible, and reasonably predictable. Most of us are risk takers when conditions are intolerable and risk avoiders when conditions are tolerable.

A few people will choose option #1 in both situations A and B. These individuals are willing to take their chances in a range of situations in order to improve their outcome. We call them risk takers. Are you one of them? The only way to know is to ask yourself. Similarly, a few people will choose option #2 in both situations. These individuals seek to avoid taking chances in a range of situations in order to preserve their current state of health. We call them risk avoiders. Are you one of them? Only you can answer that question. b

_______________

a Notice that the outcomes occur in the immediate future so there is no issue of timing or need to discount the benefits or the harms.

b There is a fourth option, which is to choose option #2 in situation A and option #1 in situation B. The small number of individuals who make this choice usually have a very different perception of what utilities mean to them. For instance, they might perceive little difference between a 0.80 and a 0.20 utility.

TABLE 3-5 Types of Individual Decision Making