1 Epidemiology in Sum

1.1 What Exactly Is Epidemiology?

Epidemiology is the study of the distribution and determinants of

health-related states or events (including disease) and the application

of this study to the control of diseases and other health problems.

—World Health Organization^[1],^[2]

Epidemiology is the foundational science of public health and population health. Epidemiologic studies are applied to help control health problems in populations. Examples of populations are patients in a single hospital, patients seen by a particular practice, people living in a particular town or group of towns, and people with a particular disease. Epidemiology is the study of the distribution and determinants of health-related states or events (including disease) and the application of this study to the control of diseases and other health problems.^[3],^[4] Very simply put, epidemiology is the study of how many people got sick, how they got sick, and why they got sick. For our purposes, epidemiology is important for clinical decision making (e.g., diagnostics) and for continuing education on the who, what, where, when, why, and how of health.

• Distribution – The frequency of disease occurrence. This may vary from one population group to another. An example is the “Top 10 Causes of Death in the United States,”^[5] published annually by the National Center for Health Statistics,^[6] a part of the Centers for Disease Control and Prevention^[7] (figure 1.1).

• Determinants – Factors that are capable of bringing about change in health. Examples include things that are CBRE (chemical, biological, radiological, or explosive), environment, stress, and social determinants of health. (See section 1.2.)
• Health-related states or events – These include but are certainly not limited to disease, behavior, utilization of health services, drugs, and health outcomes. Examples include infectious diseases, disabilities, hospital bed capacity, life expectancy, mental health, chronic disease, and injury.
• Application – Epidemiology has its own particular measures and also applies biostatistics to identify and solve problems.
• Control – Epidemiology has four aims: to describe the health status of populations, to explain the etiology of disease, to predict the occurrence of disease, and to control the occurrence of disease. Prevention is the ultimate solution. There are three types of prevention (figures 1.2 and 1.4): primary, secondary, and tertiary prevention.

1.1.1 Primary, Secondary, and Tertiary Prevention

Figure 1.2: Primary, secondary, and tertiary prevention.

Much of what we do in public health and healthcare is primary prevention. Our goal is to keep someone from getting disease or having an injury or some other health issue. Not everyone can get every health problem, but in the people that can (those that are susceptible) we want to create targeted messaging, use clinical markers, and work on behavioral change to reduce risks. For example, with COVID-19, our preventative messaging and actions would be to encourage vaccination, encourage mask wearing, and to stay home when sick. With ovarian cancer, however, we would not have the same discussion with every patient we might come across. If patients do not have ovaries, they are not susceptible and we do not focus on them for this particular prevention.

At some point, patients may become exposed, meaning that they may actually come in contact with an infectious agent or carcinogen or simply have a behavior that does something to cause pathological changes in the body. When this happens, it is not always immediately clinically apparent. For example, if a person is exposed to HPV, it takes time for that infection to lead to throat cancer. Until the person shows outward signs and symptoms, the disease may only be picked up via screening, a technique used in secondary prevention. Screenings help us find disease that otherwise would not be found until it reached the clinical stage (signs and symptoms). Patients may not be seen until the disease is clinically apparent and a stage too late for intervention, a phenomenon related to the social determinants of health. This is a reason to encourage all patients to have an annual physical exam including blood work and any additional screenings recommended by the US Preventive Services Task Force (USPSTF).^[8] Our goal is to diagnose patients as often as possible when the disease is subclinical. This is called lead time. We’ll discuss lead time and related bias in section 4.5. Screening tools do not exist for all health conditions.

Tertiary prevention is used when the disease is already clinically apparent and we need to minimize the long-term effects of it. For example, patients with cataracts may need surgery to see better and improve their ability to perform activities of daily living. Patients with a broken femur may need surgery and rehabilitation to walk again. The ultimate goal is to prevent long-term disability and death.

Example: Physical activity as a preventative measure

For a more complete example, look at figure 1.4. If we want to prevent noncommunicable diseases (e.g., diabetes), we can use physical activity as a preventative measure. Of course, it is not the only component in preventing noncommunicable diseases but is a known tool to reduce risks. In this figure we see something called primordial prevention appear before primary prevention. When we think of the difference between the two, often we think about what we can do to prevent problems in our healthiest of patients. If we take a public health perspective, this would include making green space available to more of the population to encourage physical activity. From a clinical perspective, we want to encourage physical activity among all of our patients. When our patients begin to have risk factors for chronic disease, such as high blood pressure, being physically inactive, or high central adiposity (body fat), we may want to encourage higher levels of physical activity and work with patients and specialists to target their particular barriers to achieving more physical activity and exercise (primary prevention). If patients develop prediabetes, for example, we have now entered the secondary prevention area. We need to try even more tactics to improve their physical activity level and reduce their risk for development of Type II diabetes. We want to involve specialists such as nutritionists and life coaches in addition to exercise specialists to stop and reverse the course the patient is on. If none of these things work or the biologic component (e.g., family history, malabsorption of vitamins and minerals, or other conditions that preclude more activity or the benefit from activity), the patient may progress to Type II diabetes (tertiary prevention). We must continue working with this patient and specialists to minimize the long-term effects of their diabetes and encourage a long and healthy life.

As we examine topics in this book, we will use a subspecialty of injury epidemiology—sports and recreation injury epidemiology (SRI)—for examples. Other subfields of epidemiology can be found in section 1.3.

1.2 Causality and Disease Prevention

1.2.1 Causality

Disease (“when something is wrong with a bodily function”^[9]) is a complicated relationship^[10] between factors of the agent, the host, and the environment (figure 1.5). Different diseases require different balances and interactions of these three components. Development of appropriate, practical, and effective public health measures to control or prevent disease usually requires assessment of all three components and their interactions.

An agent is the thing that causes disease or injury. An agent can be chemical, biological, radioactive, or environmental. Most often we think of infectious agents such as viruses. However, for many diseases, just having the presence of an agent is not sufficient enough to cause disease. We must also consider pathogenicity (ability to cause disease) and dose (“amount of a substance available for interaction with metabolic processes or biologically significant receptors”^[11]). Some diseases, especially chronic diseases, have multiple potential causes that must be evaluated on a case-by-case basis.

The host is the human or creature that can get the disease. Intrinsic factors about the host influence their exposure, their susceptibility to the disease, and their response to whatever causes the disease or injury. Our exposures are linked to our behaviors and social determinants of health, such as our sexual practices, hygiene, diet, physical activity patterns, occupation, and personal choices. Whether or not an exposure affects us is linked to how susceptible we are and our body’s response to the agent. Immunocompromised patients or patients with particular genetic features may respond differently to certain conditions, while a person’s anatomic structure may make them more susceptible to other conditions. There are a number of different factors about the host that influence susceptibility, including their medications and disease.

The environment in the epidemiological triangle refers to the extrinsic factors about the host and factors that support the agent. Social determinants of health such as neighborhood flood risk, air quality, sanitation, clean water, and access to health care are signs that the environment may encourage disease or injury. Poor drainage or the presence of invasive species may encourage a poor climate that encourages insects such as mosquitoes that can transmit disease.

While epidemiology often helps prove that one thing has an influence on or directly causes another thing, the idea of causality (causing an effect) is complex. The Bradford Hill Criteria^[12] (figure 1.6) are a good starting place when deciding whether a particular something (a risk factor, an action, etc.) is the thing or a part of things that cause a disease or health problem. These criteria are not requirements to prove causality, but we can consider them sometimes as strong suggestions. If we remove an actual cause of disease, we expect that there is then a lower risk of the occurrence of the outcome (e.g., disease or injury). There can be more than one cause of any health outcome (see below for more on multicausality). The more of these items in the Bradford Hill Criteria that are true, the more believable the possibility that the “factor” causes the outcome of interest. However, one study result does not prove causality no matter how many of the Bradford Hill Criteria are met. It takes a mountain of evidence, a solid combination of study types, and a variety of populations being examined to have strong evidence of causality.

Figure 1.6: What are the Bradford Hill Criteria for causation?

Sometimes we simply do not know why answers are different between studies. We always need to acknowledge that we have varying amounts of uncertainty when examining relationships and causality. Our research to find certainty (or causality), leads us to identify factors that are associated with or create risk for disease or a health-event. To be a risk factor,

• The exposure must precede the onset of disease,
• The frequency of disease must vary by the value of the exposure, and
• The observed association must not be due to error.

We can often identify the specific action that resulted in a concussion in sports, such as heading the ball, running into a wall, falling, or getting hit. In addition to plotting our prevention on a timeline, as seen in figure 1.3, subfields of epidemiology (presented in section 1.3.2) often create and refine tools to help identify where to intervene. In SRI, Van Mechlen’s Injury Prevention Sequence^[13] (figure 1.7) is an example of this. In order to prevent an SRI, we must first know the distribution (“Step 1: Establishing the extent of the injury problem”). We must then know the determinants (“Step 2: Establishing the etiology and mechanisms of the injury”). We can use those two pieces of information to apply some intervention that can allow us to control SRI (“Step 3: Introducing a preventative measure”). After this, we need to evaluate what we did and how it worked (“Step 4: Assessing the effectiveness”). As seen in figure 1.7, this process is cyclical. If we succeed, the extent of the problem should be lessened. We want to eventually make the problem not a problem.

The next figure shows how this works when applied to injuries to volleyball players.

If we extend the idea of the four-step intervention to primary, secondary, and tertiary prevention, we might use an injury tool called Haddon’s Matrix.^[14] It was originally created for designing the prevention of motor vehicle crashes. Figure 1.9 presents an example for the prevention of traumatic brain injury (TBI) in baseball and softball.^[15] In Haddon’s Matrix, primary prevention is called “pre-injury,” secondary prevention is called “injury,” and tertiary prevention is called “post-injury.” Measures for prevention can be applied to the host (the athlete), the agent (the thing that can hurt the athlete), the physical environment (the built environment), and the social/economic environment (nonphysical environmental factors such as social determinants of health [discussed later in this chapter]).

Figure 1.9: Haddon’s Matrix for prevention of TBI in baseball and softball.

1.2.2 Multicausality

No matter what, there is no singular cause of any disease. As seen in the example Haddon’s Matrix in figure 1.9, multiple factors play into a TBI before it happens, when it happens, and after it happens. The same is true for other diseases, including infectious ones. Take COVID-19 for example; people must come in contact with the virus. More importantly, they must come in contact with enough of the virus for a long enough time and be susceptible to getting the virus. People who work particular jobs (e.g., ICU nurses, cashiers at convenience stores, and educators) are at higher risk, or more susceptible, to coming in contact with COVID-19 than people who telework and pick up all of their necessities using curbside pickup. This susceptibility, of course, is also tempered by each individual person’s health status, vaccination status, and age, among other factors.

The factors we consider about the distribution of disease could be summed up as the overlapping factors of who is affected (person), where disease occurs (place), and when disease occurs (time). In descriptive epidemiology we quantify our population and the problem by these three factors:

• Person includes any and all characteristics of the patients affected by the outcome we are interested in. Examples include age, sex, gender, race, ethnicity, religion, education, behaviors, housing status, and occupation. We can use this information to better describe and examine who is affected by particular conditions, who is more likely to have particular risk factors, and overall what those affected have in common.
• Place makes us think not just about geographic or physical locations for those that are affected but the characteristics of those places as well. We want to know where people live, where they got sick, where they sought care, the climate of these places, the type of facilities they live or work in, the places that they eat, and anything else that could inform their susceptibility, response, or exposure to a disease. Examples include the zip code where the event occurred, whether disease occurred in a rural area, and whether or not the climate was arid.
• Time, as we might guess, has to do with when the outcome or exposure occurred. We want to know not just a time as in an hour of the day but a day of the week, time of the year, whether the event occurred after some other event, or if all events happened at the same time. Examples include daybreak, flu season, after a rock slide, at a potluck, or the first day of the work week.

Below is an example of how we might share this descriptive epidemiological information with others. It is a typical “Table 1,” or the first table usually shown in an epidemiological or clinical paper. This table is our first look at a study population and describes each factor about them.

Example: Person and time

In a 2020 paper, Newton et al. describe the characteristics of Medicaid-insured children with sports- and nonsports-related concussion before and after Ohio’s 2013 concussion law.^[16]

	Pre-law		Post-law		Overall
	n (%)		n (%)		n (%)		P-Valueb
Total	2742 (44.5)		3415 (55.5)		6157 (100)
Sexa							0.982
Male	1751	-64.1	2187	-64	3938	-64
Female	982	-35.9	1228	-36	2210	-36
Age at first visit							< 0.0001
< 5	546	-19.9	421	-12.3	967	-15.7
5–9	484	-17.7	630	-18.5	1114	-18.1
10–14	879	-32.1	1274	-37.3	2153	-35
15–18	833	-30.3	1090	-31.9	1923	-31.2
Sports-related	726 (39.9)		1095 (60.1)		1821 (29.6)
Sex							0.172
Male	583	-80.3	850	-77.6	1433	-78.7
Female	143	-19.7	245	-22.4	388	-21.3
Age at first visit							0.012
< 5	24	-3.3	13	-1.2	37	-2
5–9	97	-13.4	156	-14.2	253	-13.9
10–14	359	-49.4	573	-52.3	932	-51.2
15–18	246	-33.9	353	-32.2	599	-32.9
Nonsports-related	2016 (46.5)		2320 (53.5)		4336 (70.4)
Sexa							0.706
Male	1168	-58.2	1.337	-57.6	2505	-57.9
Female	839	-41.8	983	-42.4	1822	-42.1
Age at first visit							< 0.0001
< 5	522	-25.9	408	-17.6	930	-21.4
5–9	387	-19.2	474	-20.4	861	-19.9
10–14	520	-25.8	701	-30.2	1221	-28.2
15–18	587	-29.1	737	-31.8	1324	-30.5
aThere are nine injuries with missing sex.
bP-values are based on chi-square tests of the distribution of sex and age across the law period.

Figure 1.10: Injury rates by person: Participant demographics by sports- and nonsports-related concussions, sex, and age, from pre- to post-law period, 2008–2016.

In addition to describing the population, Newton also showed the rates of concussion over time before and after the law was enacted.

Example: Person and place

Our next example includes data on person and place from a study of single-sport specialization during youth by present-day professional baseball players.^[17] In table 1.12 we see that people can be described not just by their demographic characteristics. Players are described by their level of professional baseball and the position they play. In figure 1.13, we see a geographic display of where players lived when they first started specializing in one sport.

	n (%)
Level
Rookie	336 (20.7)
A	578 (35.6)
AA	224 (13.8)
AAA	166 (10.2)
MLB	320 (19.7)
Total	1624
Position
Pitcher	902 (55.4)
Starting pitcher	413
Relief pitcher	482
Infield	333 (20.4)
Outfield	249 (15.3)
Catcher	145 (8.9)
Total	1629

Figure 1.12: Level and position of all players: ^aMLB, Major League Baseball.

Once we start looking at the problem by person, place, and time, we also have to identify our best estimate as to how the particular case or cases of disease occurred. Rothman’s Pie Model^[18] (figure 1.14) is an example of how we can consider the multicausality of disease or health-related events.

• Each completed pie (sufficient cause) is a case (person) of a single disease. Figure 1.14 demonstrates that the same disease may have more than one sufficient cause. Consider that two people with a TBI do not have to have the same factors that contributed to their TBI or recovery from it (sufficient cause I, II, III).
• Each individual pie piece (or component cause) is the equivalent of an individual factor that contributes to causing disease (A, B, C, D, E, F, G, H, I, J).
• If a piece of the pie is present in every pie, it is considered a necessary cause. In figure 1.14, piece A is in every pie. We do not know whether every disease or health-related event has a necessary cause. Some, such as COVID-19, do. If a person has every factor in common with a person who has COVID-19 but did not come in contact with the virus, that person will not have COVID-19.

Component causes can be proximal (downstream), medial (midstream), and distal (upstream) to the health-related event (figure 1.14). In both clinical and public health matters it can often be very important to know which particular risk factors or component causes are present. For example, focusing on a patient’s lack of physical activity (risk factor for disease) when they come in for an annual visit without knowing more details about why they are not getting enough physical activity likely will not change the patient’s behavior. However, if you know that the patient works three jobs to feed the family (proximal), you may be able to find small ways to help them increase their activity level. Continue asking why and you’ll eventually get to the distal causes. These component causes are where we should focus our primordial and primary prevention efforts.

What are some more examples? Figure 1.16 shows an example of how component causes contribute to sports injury. Factors about a person such as age, physical fitness, and anatomy (internal risk factors) make an athlete predisposed to injury. Being exposed to external risk factors, such as particular equipment or a hazardous playing environment, influences the predisposed athlete and results in one now susceptible to injury. Finally, something (an inciting event) must occur, such as the player planting a foot on the turf in a particular way to cause an injury.

Social determinants of health (SDOH)^[19] (figure 1.17) are the conditions in the environments where people are born, live, learn, work, play, worship, and age that affect a wide range of health, functioning, and quality-of-life outcomes and risks. Social determinants are issues that affect the people we study; what we ask or measure in our studies; what we consider as solutions; or whose diseases matter the most. Each of these factors holds its own importance in health.

Figure 1.17: Examples of social determinants of health.

By including SDOH in epidemiological work, we can make progress toward achieving health equity (figure 1.18). These factors are all quite important when we consider a person or community’s risk for disease, the probability they have for getting adequate care, and the probability for achieving health professionals’ intended outcome from prevention efforts.

In order to achieve health equity, health disparities must be addressed. Health disparities are not simple differences in health status. They are differences in health outcomes between populations that are tied to race, ethnicity, sex, gender, age, disability, SES, and geography. By describing the health of the population, we can identify what health disparities exist so we can do something about them.

1.2.3 Clinically Apparent Disease

When we consider the universe of disease or health-related states, we must also recognize that some existing diseases are not immediately apparent. This contributes to fluctuation in our understanding of disease and disease patterns. It also contributes to changes in our observed burden of disease or how we decide what resources are needed for our community. For example, a large percentage of people with SRI do not seek care in a clinical system. Many of our estimates of the prevalence of SRI, however, are based on how many appear in an emergency department or other clinical centers. Epidemiologists have a responsibility to make it clear who and what they are considering when discussing disease. Figures 1.19 and 1.20, icebergs of disease and injury respectively, show that though we might see some aspects of disease or injury, most factors about the disease are not visible without looking below the surface.

If we consider the iceberg illustration in figure 1.19, we see that much of the disease that exists is clinically inapparent. This is a good rationale for secondary prevention. We need ways to find and address persistent diseases or symptoms that patients have, find disease that is not yet causing symptoms, and find those who may be carrying disease. When we apply this same concept to the causes of injury (figure 1.20), we see that factors we notice most about disease are ones we can easily “diagnose”—things that are biological, things that are psychological, and things that are caused by behaviors of patients.^[20] However, most injury is driven by a host of other factors, including relationships with family and friends (interpersonal), where people work and play (organizational), societal norms and community resources (community), and policies (society). Different people have different risks.

1.3 The Importance of Epidemiology in Your Research

As seen throughout the first part of this chapter, epidemiology is a broad subject that studies the distribution and determinants of health-related states or events (including disease).^[22],^[23] This section and the rest of the text focus on the application of this study to the control of diseases and other health problems.^[24],^[25] Specifically, there are two major uses of epidemiology: (1) describing the status of the population’s health and the health services of the population and (2) determining or describing the etiology of diseases, conditions, syndromes, and so on. Nearly everything epidemiology is used for in public health and clinical medicine falls within these two uses. We also recognize that when we think about epidemiology, we can consider that we use it to describe the who, what, where, when, why, and how of disease. All of these factors are important for research and for helping patients every day. We conclude this section with a list of some of the subspecialty areas of epidemiology. Each of these areas has very specific methods that are important to helping people and each of them may have a more direct impact on the work you do in clinical settings than others. You do not need to remember each of these areas, but do know that you can find subject-matter experts for nearly any problem you face!

1.3.1 Uses of Epidemiology

There are two major uses of epidemiology:

1. Describing the status of the population’s health and the health services of the population
   • Health services research
     • How do our services work?
     • Are the services in the right place?
     • Is utilization adequate? Who is not using the services?
     • Appropriateness of staffing and facilities
     • Diagnose the health of the community
     • History of the health of populations
   • Policy
     • Create or evaluate the policies that affect health
   • Health promotion/health behavior
   • Social science
   • History
     •  Where have we been?
     • Where are we going?
     • Look at long-term trend
       1. What causes changes in our numbers?
       2. Were there changes in our diagnosis, etiology, or reporting criteria?
     • Identify at-risk populations
     • Note improvements or declines in health
2. Determining or describing the etiology of diseases, conditions, syndromes, etc.
   • Biology
   • Ecology
   • History
   • Laboratory sciences
   • Genetics

If epidemiology is about the who, what, where, when, why, and how of disease:

Descriptive epidemiology is about who, what, where, and when.

Analytic epidemiology is about why and how.

1.3.2 Subfields of Epidemiology

In addition to injury epidemiology, subareas or subfields of epidemiology include:

While this is not an exhaustive list of subareas of epidemiology, they do represent much of the expertise available. Because of the breadth of subspecialties, when searching for literature on any given topic you are likely to find at least some epidemiological papers available. Peer-reviewed epidemiology manuscripts are in a large variety of journals, including many that do not use the word “epidemiology” in the title. Many epidemiology papers can be found using academic library resources (such as those from the University Libraries at Virginia Tech^[26]), PubMed,^[27] or Google Scholar.^[28]

Figure Descriptions

Figure References