Chapter 1: Epidemiology past and present
Review questions
R.1.1 epi = “upon”; demos = “the people”; ology = “to speak of” or “to study.”
R.1.2 A personal response is requested.
R.1.3 There are several differences between epidemiology and clinical medicine. One difference is their primary unit of concern. The primary unit of concern in epidemiology is the group (an “aggregate of individuals”). In clinical medicine, the primary unit of concern is the individual. With respect to epidemiology and public health: epidemiology is primarily a “study of,” while public health is an “activity” requiring social participation.
R.1.4 Elements of the 1948 WHO definition of health and well-being: (1) physical, (2) mental, (3) social.
R.1.5 See Table 1.1.
R.1.6 The epidemiologist must communicate their findings in order to effectively participate with other disciplines and sectors in deciding and implementing public health practices and interventions.
R.1.7 Morris’s seven uses of epidemiology: (1) historical study; (2) community diagnosis; (3) workings of health services; (4) individual chances; (5) complete the clinical picture; (6) identify syndromes; (7) search for causes.
R.1.8 Community diagnosis determines the incidence and prevalence of disease and disease determinants in communities and community subgroups.
R.1.9 Key elements of demographic transition of the 20th century: increased longevity, decreased fertility, aging of the population.
R.1.10 Key elements of the epidemiologic transition of the 20th century: decreases in acute and contagious diseases; increases in chronic, noninfectious, lifestyle diseases.
R.1.11 Steep declines have been seen in cardiovascular disease (heart attacks), cerebrovascular disease (strokes), and pneumonia and influenza.
R.1.12 The population pyramid has “squared off” (i.e., became less of a pyramid) over time, with a larger percentage of the population shifted toward the older age groups.
R.1.13 Examples of modifiable risk factors: tobacco use, alcohol use, diet, high blood pressure, high risk sexual practices, exposure to sunlight and other forms of radiation.
R.1.14 False. While it is true that death due to some cancers increased during the 20th century (e.g., lung cancer), others have declined. In addition, much of apparent increases have been due to the aging of the population. Thus, after age-adjustment, the ups and downs more-or-less balanced out resulting in a fairly flat cancer mortality rate (see Figure 1.2).
R.1.15 False. Age-adjusted cardiovascular mortality rates continue to decline.
R.1.16 Heart disease; cancer; stroke.
R.1.17 True.
R.1.18 Epidemiology became a recognized discipline in the 19th century with the creation of the Epidemiological Society of London (established 1850).
R.1.19 Hippocrates (400 BCE).
R.1.20 Measuring, sequencing, classifying, grouping, confirming, observing, formulating, questioning, identifying, generalizing, experimenting, and testing.
R.1.21 Matching: A = Syndenham; B = Pott; C = Graunt; D = Fracastoro; E = Salmon; F = Pinel; G = Louis; H = Farr; I = Snow.
Chapter 2: Causal concepts
Exercises
2.1 No answer provided.
2.2 d
2.3 c
2.4 b
2.5 a
2.6 c
2.7 b
2.8 c
2.9 b
2.10 a
2.11 Matching:
a = Virulence
b = Sufficient constellation
c = Non-necessary Component Cause
d = Infectivity
e = Causal web
f = Pathogenicity
g = Necessary cause
2.12 Matching:
a = Coherence
b = Biologic gradient
c = Plausibility
d = Temporality
e = Experimentation
f = Analogy
g = Specificity
h = Consistency
i = Strength
2.13 Matching:
a = Biologic gradient
b = Plausibility
c = Strength
d = Consistency
e = Analogy
e = Temporality
2.14 Consistency implies that studies are consistent in the estimation of association. If an exposure consistently causes disease in a given person, it is said to be sufficient. Sufficiency is not one of Hill’s criteria for causality.
Review questions
R.2.1 Stages of disease: susceptibility, preclinical, clinical, resolution (recovery, disability, or death).
R.2.2 The onset of the preclinical stage is exposure to the agent. The onset of the clinical stage is marked by first symptoms. The onset of the resolution is recovery, disability or death.
R.2.3 The objective of primary prevention is to prevent new occurrences. The objective of secondary prevention is to delay the onset of disease or decrease its severity. The objective of tertiary prevention is to slow progression or minimize the progression of disease.
R.2.4 The agent multiplies within the host during the incubation of an infectious disease.
R.2.5 Synonyms for incubation period: latent period, empirical induction period (roughly).
R.2.6 Tuberculosis, AIDS, leprosy.
R.2.7 Mammography is a form of secondary prevention; because it detects disease after it has been initiated but before it becomes clinical.
R.2.8 There are many reasons it is important to understand the natural history of HIV for its effective control. As an example, we must be aware of the period between infection and detection of antibodies to understand that diagnosis of the disease will be delayed until some time after exposure. In addition, one must be aware of the long incubation period during which the host is symptom free but is still contagious.
R.2.9 The spectrum of disease. (For infectious diseases: the gradient of infection.)
R.2.10 This describes the epidemiologic iceberg.
R.2.11 A cause is any predecessor without which the effect would not have occurred or would have occurred at a later time.
R.2.12 A causal interaction is the bringing about of an effect by two or more factors act together.
R.2.13 Measles virus is necessary but not sufficient to cause measles. (The agent will not cause disease in an immune individual.)
R.2.14 A causal mechanism is completed when the outcome becomes inevitable.
R.2.15 The causal complement is (E + F).
R.2.16 The causal complement is E.
R.2.17 Phenylketonuria is both a genetic disease and an environmental disease. The genetic disorder involves the deficiency in the enzyme needed to metabolize phenylalanine. The environmental component is the presence of phenylalanine in the diet.
R.2.18 Contributing/component causes of hip fractures in the elderly: Low calcium diet, osteoporosis, genetic susceptibility to osteoporosis, female sex, weakness, poor balance, sedation, slippery surface, lack of hand rails, etc.
R.2.19 A direct cause is close to the pathogenic mechanism. An indirect cause is connected to the pathogenic through other factors.
R.2.20 Types of pathogenic agents: biological, physical, and chemical.
R.2.21 Types of chemical pathogenic agents: nutritive excesses and deficiencies, toxins, drugs, allergens.
R.2.22 Types of physical pathogenic agents: heat, light, radiation, noise, vibration, and objects that cause trauma.
R.2.23 Infectivity = ability to infect; pathogenicity = ability to cause disease; virulence = ability to cause severe disease.
R.2.24 Epidemiologic homeostasis occurs when agent, host, and environmental causes of disease are balanced in such a way as to maintain the current rate of disease in the population.
R.2.25 Causal inference is the process of deriving cause-and-effect conclusions from fact and knowledge.
R.2.26 We base preventive measures on knowledge of causal mechanisms to increase their efficacy. False knowledge can have a contrary effect.
R.2.27 There are times when discovery of effective preventive measures pre-date identification of the causal mechanism
R.2.28 See Table 2.3.
R.2.29 This initial Surgeon General’s Report on Smoking and Health was published in 1964.
R.2.30 Because there are always alternative explanations for associations.
R.2.31 False. Statistical methods cannot by themselves establish proof.
R.2.32 Ratios of incidences (relative risks) are generally considered to be the most direct measure of the strength of an epidemiologic relationship.
R.2.33 Strong associations are less likely to be “explained away” by confounding.
R.2.34 No. Weak associations are just more difficult to “prove.”
R.2.35 No, since multiple studies may be consistently incorrect, especially if they share or exhibit multiple flaws.
R.2.36 True. This is the sine quo non. The cause must always precede the effect.
R.2.37 Coherence holds that all sources of evidence “stick together.” Plausibility holds that relations can be explained by current knowledge.
R.2.38 This is an example of an analogy.
R.2.39 Biological gradient.
R.2.40 Consistency, strength, specificity, temporality, biological gradient, plausibility, coherence, experimentation, analogy.
Philosophical considerations
R.2.41 Ultimate proof in empirical sciences is not possible. However, statements of proof can be very strong, and even overwhelming.
R.2.42 Decisions having to do with scientific hypotheses (type 1) require rigorous skepticism. The latter having to do with public health interventions (type 2) may require making a reasonable choice based on available information.
R.2.43 The Problem of Induction is the philosophical quandary that observed sequences of occurrence do prove cause and effect (post hoc propter hoc).
R.2.44 True. Refutationists believe that a theory is not scientific unless it is falsifiable.
R.2.45 One can never fully prove that all swans are white because the next swan that comes long may be black or even light gray.
R.2.46 No amount of observations can prove a hypothesis. In contrast, one strong disproof can dispel a theory.
Chapter 3: Epidemiologic measures
Exercises
3.1 Point prevalences, period prevalence, and risk
3.2 Hypertension in a cohort of men
year−1Without the actuarial adjustment: 
year−1
The adjustment made little difference because the outcome was uncommon.
3.3 Vital statistics
All rates are “per 1000.”
3.4 Prevalence in an open population
3.5 More vital statistics
= 853
= 8.503.6 Effect of a Treatment An effective treatment that increases survival but does not result in a permanent cure would have no influence on incidence, but would increase prevalence over time.
3.7 Fatalities associated with travel
miles travels3.8 Accidents in hospitals The author’s interpretation is incorrect. The data represent incidence counts; no “denominator data” are presented. Such data cannot be used for statements about rates or risks. It is possible that there are many more patients in the 62 and over age group than in any other age group, and the higher number of accidents simply reflects this great number of people initially at risk.
3.9 Stationary? The size of the US population is increasing and its age distribution is shifting. Therefore, US demographics are not stationary.
Mortality rate and life expectancy
3.11 Comparing prevalences We can not conclude population A has twice the risk because prevalence depends on both incidence and mean duration of disease. If the cases in population A survived twice as long as the cases in population B, it could have the same incidence and double the prevalence.
3.12 Risk and rate of Breast cancer
3.13 Coronary heart disease
3.14 Driving errors
= ½ mistake per mile
= 0.0025
3.15 N = 6
= 0.143 per person-year3.16 Cohort study
3.17 More population based rates
3.18 Actuarial adjustment of person-Time (18 persons × 1 year) + (82 persons × 1 years) = 182 person-years
3.19 Framingham men RR = R1/R0 = 0.1203/0.0352 = 3.4. Here is how the original paper (Kannel et al., 1961, p. 39) reported the results: “Analysis of these groups reveals a gradient of risk of developing CHD, such that those with serum cholesterol over 244 mg per 100 ml have more than three times the incidence of CHD as those with cholesterol levels less than 210 mg per 100 ml.”
3.20 Framingham women RR = R1/R0 = 0.0435/0.0180 = 2.4. The high cholesterol women had more than double the risk of CHD.
3.21 Restenosis
3.22 Primary cardiac arrest and vigorous exercise
3.23 California mortality
3.24 Arkansas mortality
3.25 Egyptian mortality
Review questions
R.3.1 The numerator of incidences includes only cases that had onsets during the period of observation. Prevalence counts all cases, old and new.
R.3.2 Not necessarily. New York has a larger population. The greater number of deaths may merely reflect its large population size.
R.3.3 Size, time of observation, age, and other and characteristics
R.3.4 A ratio is a combination of two numbers that show their relative sizes. It is one number divided by another.
R.3.5 Cohort.
R.3.6 Average risk, risk, cumulative incidence.
R.3.7 Numerator = no. of disease onsets; denominator = size of cohort at risk.
R.3.8 Because the objective of an incidence proportion is to estimate the probability of developing the disease.
R.3.9 (a) The period length of observation. (b) The age distribution of the group.
R.3.10 40.
R.3.11 Incidence density, average hazard, person-time rate.
R.3.12 1 person observed for a year; 2 people observed for half a year each; 3 people observed for one-third of a year each; etc.
R.3.13 68 person-hours.
R.3.14 Numerator = no. of disease onsets; denominator = amount of “person-time” in population.
R.3.15 0.013 33 person-year−1 × 1000 person-years = 13.3.
R.3.16 When the disease is rare (cumulative incidence <5%) and the period of observation is one year.
R.3.17 Numerator = number of existent cases; denominator = population size.
R.3.18 It considers both new and old cases and involves no follow-up of individuals.
R.3.19 It will increase.
R.3.20 “Exposure” (independent variable) and “disease” (dependent variable).
R.3.21 True.
R.3.22 It means that higher levels of exposure are associated with higher incidences of disease.
R.3.23 Division.
R.3.24 Subtraction.
R.3.25 This is an example of an RD.
R.3.26 This is an RR.
R.3.27 RD quantifies the effect of the exposure in absolute terms.
R.3.28 RR quantifies the effect of the exposure in relative terms.
R.3.29 RR.
R.3.30 It would not be correct to make this statement because the exposure increases risk by only 50%.
R.3.31 False.
R.3.32 True.
R.3.33 It changes to it’s reciprocal. For example, an RR of 2 becomes 1/2 = 0.5.
R.3.34 85%.
R.3.35 AFe.
R.3.36 AFp.
Chapter 4: Descriptive epidemiology
Exercises
4.1 Ecological correlations There are strong positive correlations for cigarettes and bladder cancer and cigarettes and lung cancer; there is a moderately strong correlation for cigarettes and kidney cancer. There are associations between bladder cancer, lung cancer, and kidney cancer. The implication is that lung cancer, kidney cancer, and bladder cancer may have a common underlying cause, perhaps cigarettes or something associated with cigarettes.
4.2 Notifiable conditions No answer provided.
4.3 Le Suicide
Review questions
R.4.1 Descriptive epidemiology explores rates according to person, place, and time variables with the primary intention of generating hypotheses. Analytic epidemiology collects data that has been specifically designed to address hypotheses about specific risk factor.
R.4.2 False. There is no firm demarcation between descriptive epidemiology and analytic epidemiology.
R.4.3 A case series is a description of the history and clinical manifestations of a small number of individuals with a particular disease outcome.
R.4.4 Case series lack the denominator data needed to calculate rates.
R.4.5 True, for example see Illustrative Example 4.1.
R.4.6 Epidemiologic surveillance systems are structures set up to collect and analyze outcome-specific health data for planning, carrying out, and evaluating public health practices.
R.4.7 Active surveillance requires an active seeking out of population-based cases. Passive surveillance relies on doctors, hospitals, and the public to send reports to the appropriate public health surveillance system voluntarily.
R.4.8 The National Center for Health Statistics
R.4.9 place, and time.
R.4.10 Person.
R.4.11 Host factor closely tied to place: cultural practices, occupation, recreational practices, etc.
R.4.12 Environmental factor that is closely tied to “place”: climate, economic development, etc.
R.4.13 Studies show that Japanese–American women develop breast cancer rates that are typical of American women after several generations of acculturation.
R.4.14 (a) Propagating epidemic, (b) point epidemic, (c) endemic, (d) sporadic.
R.4.15 A unit of observation is the level of human aggregation upon which measurements are recorded.
R.4.16 ecological
R.4.17 True.
R.4.18 An ecological correlation is a correlation in which the units of observation are based on group rather than individual characteristics.
R.4.19 Neighborhood crime rate is an integral aggregate-level variable.
R.4.20 True.
R.4.21 An ecological fallacy (also called aggregation bias) occurs when an association seen in aggregate data does not apply to individuals.
R.4.22 A multilevel study incorporates individual- and aggregate-level variables in order to help untangle relationships between direct and indirect causes of disease.
R.4.23 Confounding bias is a spurious association caused by extraneous factors.
R.4.24 (1) Contextual variable, (2) integral variable, (3) contagion variable.
R.4.25 This is an integral group property.
Chapter 5: Introduction to epidemiologic study design
Exercises
5.1 Study types.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
