CPH Study Guide Epi Questions – Flashcards

1. What would be the best source of information on adult and infant mortality? (B) Vital statistics Data are routinely collected on all deaths that occur in the US through death certificates. Death certificates record the age at death, and are reported through the vital statistics registration system. References: Friis and Sellers; pp. 190-191 | Last et al.; p. 187

Which of the following terms is expressed as a ratio (as distinguished from a proportion)? (C) Male Births / Female Births A ratio is defined as "the value obtained by dividing one quantity by another; a general term of which rate, proportion, percentage, etc. are subsets." References: Last et al.; p. 152 | Friis and Sellers; p. 89

Mortality rates by sex in the United States generally show the following sex differences: (A) Males greater than females Males generally have higher all-cause age-specific mortality rates than femalesfrom birth to age 85 and older; in 2003, the ratio of male to female age-specific deaths was 1.4 to 1. References: Hebel and McCarter; p. 20 | Friis and Sellers; pp. 134-138

Morbidity rates by sex in the United States show the following sex differences: (B) Females greater than males This is known as the female paradox: females have lower mortality but higher morbidity rates for acute and chronic conditions. Reference: Friis and Sellers; p. 136

Case-control studies are among the best observational designs to study diseases of: (D) Low prevalence The case control design is desirable when disease occurrence is rare, becasue if a cohort study were used in such a circumstance, a very large number of people would have to be followed to generate enoigh people with the disease for study. References: Hebel and McCarter; pp. 106-107, 118 | Friis and Sellers; pp. 243-247 | Kleinbaum et al.; pp. 31-32

Comparison of mortality rates due to cancer of the uterus in users and non-users of supplemental estrogen revealed the following mortality rates per 100,000: Age Users of estrogen Non-users of estrogen 45-54 3.0 1.0 55-70 17.0 6.0 A valid conclusion derived from the above data concerning mortality among estrogen users is: (A) The mortality rates for cancer of the uterus are higher in estrogen users than non-users in both age groups studied A causal relationship cannot be demonstrated based on the results of a single study; the mortality rate is higher in users for both age groups. References: Friis and Sellers; pp. 110-112, 132-134 | Hebel and McCarter; pp. 17-22 | Kleinbaum et al.; pp. 64-66

In a large case-control study of pancreatic cancer, 20% of the cases were found to be diabetic at the time of diagnosis, compared to 5% of a control group (matched by age, sex, ethnic group, and several other characteristics) that was examined for diabetes at the same time that the cases were diagnosed. The authors concluded that diabetes played a causal role in the development of pancreatic cancer. This conclusion: (C) May be incorrect because of failure to establish the time sequence between the onset of diabetes and pancreatic cancer Case control studies cannot demonstrate the temporal relationship between an exposure and a disease. This is a requirement for concluding causality. References: Kleinbaum et al.; pp. 22-24 | Hebel and McCarter; pp. 140-141 | Friis and Sellers; p. 72

Cohort study is to risk ratio as: (B) Case-control study is to odds ratio The risk ratio is the measure of association for a cohort study; the odds ratio is the meausre of association for a case-control study. References: Kleinbaum et al.; pp. 74-75 | Friis and Sellers; pp. 240-241 | Last et al.; p. 128; Hebel and McCarter; p. 111

It is essential in cohort studies of the role of a suspected factor in the etiology of a disease that: (C) The exposed and non-exposed groups under study should be as similar as possible with regard to possible confounders. This helps ensure comparability of the two groups, increasing the inetrnal validity of the findings. All other answers are false. Reference: Friis and Sellers; pp. 266-267

A researcher is interested in identifying potential risk factors for a rare form of bone cancer in children. Which of the following study designs would be best suited to investigate the risk factors for this type of cancer? (B) Case-control Case-control studies are best suited for rare diseases. References: Hebel and McCarter; pp. 106-107, 118 | Friis and Sellers; pp. 243-247 | Kleinbaum et al.; pp. 31-32

Which of the following statements regarding disease measures is inaccurate? (D) Incidence density is a proportion The numerator for incidence density is "number of cases" and the denominator is "person-time." This represents a rate and not a proportion. References: Last et al.; p. 92 | Hebel and McCarter; p. 25 | Friis and Sellers; p. 98 | Kleinbaum et al.; pp. 53-54

In a population of 5,000 people, 100 ate spinach contaminated with E. coli (O157:H7) and became ill. Of the ill, 15 died. What was the case fatality rate? (C) 150 per 1,000 The numerator for case fatality is "number of individuals dying during a specified period of time from the disease of interest" and the denominator is "number of persons with the disease of interest." In this question, the numerator is 15 and the denominator is 100. 15/100 = 0.15 X 1000 = 150/1000 persons. References: Friis and Sellers; pp. 414-415 | Last et al.; p. 24 | Hebel and McCarter; p. 20 | Kleinbaum et al.; p. 63

A case-control study comparing ovarian cancer cases with community controls found an odds ratio of 2.0 in relation to exposure to radiation. Which is the correct interpretation of the measure of association? (C) Ovarian cancer cases had 2.0 times the odds of exposure to radiation when compared to controls. In a case control study, we are comparing the prevalence of exposure given disease status, therefore, we compare the prevalence of radiation among those with cancer to the prevalence of radiation among those without cancer. We cannot use "risk" to describe this association because risk implies probability and we are calculating odds. References: Friis and Sellers; pp. 240-241 | Last et al.; p. 128 | Kleinbaum et al.; pp. 74-75

Double-blinded (masked) studies are an important way to: (B) Avoid observer and interviewee bias Double masking or blinding a study means that neither the patient nor the individuals involved in evaluating the outcome know the exact exposure status of any individual. This decreases the risk of observer bias because the observer does not know the true exposure. It also decreases the risk of interviewee bias because the participant doesn't know which exposure group he/she is in. References: Friis and Sellers; pp. 304-305 | Last et al.; p. 18 | Kleinbaum et al.; p. 26 | Hebel and McCarter; p. 125

The difference between primary and secondary prevention of disease is: (C) Primary prevention focuses on control of causal factors, while secondary prevention focuses on early detection and treatment of disease Primary prevention suggest that the "incidence" of disease is prevented. In general, this means that the prevention has decreased the prevalence of risk factors or increased the prevalence of preventive factors thus changing the incidence of disease. In secondary prevention, disease has already occurred and the prevention program is designed to identify disease early so that treatment can be started early. References: Friis and Sellers; p. 80 | Last et al.; pp. 141-142 | Hebel and McCarter; p. 59

A screening test is used in the same way in two similar populations, but the proportion of false-positive results among those who test positive in population B is higher than that among those who test positive in population A. What is the most likely explanation for this finding? (C) The prevalence of disease is higher in population A The number of false positives increases as the prevalence of the outcome of interest decreases. Communities with lower prevalence of outcome will have higher false-positive rates. I community B has higher false- positive rates, then it must have lower prevalence compared to population A. Therefore, population A has a higher prevalence of disease than community B. References: Friis and Sellers; pp. 382-385 | Hebel and McCarter; pp. 62-63 | Kleinbaum et al.; pp. 152-154 | Last et al.; p. 140

The relative risk of dying for smokers compared to non-smokers is: The death rate per 100,000 for lung cancer is 7 among non-smokers and 71 among smokers. The death rate per 100,000 for coronary thrombosis is 422 among non-smokers and 599 among smokers. The prevalence of smoking in the population is 55%. (D) 10.1 for lung cancer and 1.4 for coronary thrombosis Prevalence is not relevant to this question - so ignore the prevalence of smoking statement. The relative risk of dying from lung cancer for smokers compared to non-smokers is 71/7 = 10.1. The risk of dying from coronary thrombosis comparing smokers is 599/422 = 1.4. References: Friis and Sellers; p. 276 | Kleinbaum et al.; pp. 72-73 | Hebel and McCarter; pp. 29, 118 | Last et al.; p. 156

Which measure of mortality would you calculate to determine the proportion of all deaths that is caused by heart disease? (D) Proportionate mortality ratio PMR is calculated as the number of deaths within a population due to a specific disease or cause divided by the total number of deaths in the population. Reference: Friis, page 118

The staff at the state health department Birth Defects Registry systematically visit the labor and delivery units of all hospitals in the state on an on-going basis and review their records to identify major congenital malformations. This is an example of: (B) Active surveillance Active surveillance refers to a system in which project staff make periodic visits to clinics and hospitals to identify new cases of disease that have occurred (case finding). Reference: Gordis, page 54

Which of the following statements is true about surveillance data? (D) It can be used to monitor temporal trends in disease Monitoring temporal trends in disease is a primary function of surveillance systems. While surveillance historically, and perhaps more frequently, is conducted for infectious diseases, in recent years it has become increasingly important in monitoring changes in other types of conditions, such as cancer, congenital malformations, asthma, occupational exposures and diseases, and injuries and illness after natural disasters. Reference: Gordis page 54

The following are characteristics of a cross-sectional study except: (D) Patients are followed for a sufficient period of time for outcome events to occur Patients are not followed in a cross-sectional study.

All of the following statements are true about both prospective and retrospective cohort studies except: (C) They require the manipulation of the exposure of interest by the investigator There is no manipulation of the exposure in a cohort study.

In a cross-sectional study of peptic ulcer among men and women aged 35 to 59, 70 per 100,000 men and 95 per 100,000 women met the criteria for symptomatic peptic ulcer disease. The inference that in this age group, women are at greater risk of developing peptic ulcer is: (B) Incorrect because of failure to distinguish between incidence and prevalence Prevalence does not represent risk.

A prevalence study of depression drew a sample of the residents of Houston by visiting all residence in randomly selected census tracts. A screening test with a sensitivity of 99% and a specificity of 50% was used to identify individuals as "likely depressed." The study reported a prevalence of 28%. This prevalence is: (B) Higher than expected because of the specificity of the screening test Higher than expected because the specificity is so low there will be many false positives.

The answer correct is 0 (meaning of epidemiology). Epidemiology refers to the study of the frequency, patterns, causes, and control and prevention of diseases, disabilities, and injuries in the human population. There are two main aspects of epidemiology: (a) distribution-epidemiology evaluates the distribution of health states by age, gender, race, geography, time, etc., and (b) epidemiology seeks to examine the association and causal association between the exposure and the health-related states or events (Last, J.M., A Dictionary of Epidemiology, 3rd ed., 1995; MacMahan, B, & Trichopoulos, D., Epidemiology, Principles & Methods, 2nd ed., 1996, p. 1).

The correct answer isE(factors influencing the prevalence of disease).Prevalence proportion does not measure disease onset but the disease status. Whereas incidence is a measure of risk, prevalence is not because it does not take into account the duration of disease. But prevalence is related to the duration of the disease and incidence of disease. Prevalence = Incidence x Disease duration. The number of people living with a disease (prevalence) will be influenced by the rate at which a new disease develops as well asthe rate at which they are eliminated through cure (prognosis) or death (mortality). Therefore, the prevalence will increase if incidence increasesand if there is improvement in treatment but not cure, and it will decrease if there is dramatic cure (quick recovery) (Gordis, L., Epidemiology, 3rd ed., 2004, p. 40; Aschengrau, A.,&Seage, G.R., Essentials ofEpidemiology, 2003, pp. 48-49).

The correct answer is E (descriptive epidemiology).

The correct answer is B (primary attack rate). Attack rate, which is really a proportion and not a rate, is useful for comparing the risk of disease in groups with different exposures. This measure is calculated by Number of people at risk in whom a certain illness developsfTotal number of people at risk. Substituting, 35/80 x (multiplier) = 35/80 x 100 -+ 43.75 per 100 (Gordis, L., Epidemiology, 3rd ed., 2004, p. 25).

The correct answer is A (secondary attack rate). Secondary attack rate yields the index of the spread of a disease within a family, household, dormitory, dwelling unit, etc. This measure is estimated by Number of new cases in a group - Initial case(s) / Number of susceptible persons in a group - Initial case(s). Substituting: 150 - 100 (each household had one primary case and there were 100 households) /500 - 100 (primary case) x 100 -+ 50/400 x 100 = 12.5% (Mausner, J.,& Kramer,S., Epidemiology-An Introductory Text, 2nd ed., 1985, pp. 280-281).

The correct answer is E(study design). A design that begins with exposure(s) of interest and disease-free subjects and aimed to assess the outcome in the future by following the subjects represents a prospective cohort study. This is not a clinical trial or a broad intervention design because the subjects of the FHS were not assigned but observed. The time relation in a prospective cohort study is present and future (continuing). A case cross-over design uses the previous experience of the cases as a substitute for a control series 88 CPH EXAM QUICK REFERENCE REVIEW to estimate the person-time distribution in the source population. Ambidirectional cohort isa design tha t is both retrospective and prospective in its observation of the outcome. Retrospective cohort (historical cohort studies) is a design in which the cohorts are identified from recorded information and the time during which they are at risk for disease occurred before the beginning of the study. Simply, retrospective study is conducted by defining the cohort and collecting information, which applies to past time. Case control is a design in which the groups of individuals are defined in terms of whether they have or have not already exper ienced the outcome under consideration, and the exposure is then measured (Elwood, J.M., Critical Appraisal of Epidemiological Studies in Clinical Trials, 2nd ed., 2003, pp. 14-35; Rothman, K.J., Epidemiology, An Introduction, 2002, p. 89).

The correct answer is E (study design). This is an observational design because it involves natural observation without assignment. A casecontrol is a design in which the outcome had already occurred prior to the commencement of the study, and study participants were selected on the basis of the disease.The direct measure of risk is not possible in case-control study but estimated as odds ratio.

The correct answer is E (observational versus experimental design). Experimental design if feasible is considered the gold standard compared with observational studies. The active manipulation or assignment of the treatment by the investigator is the hallmark that differentiates experimental designs from observational studies. Whereas not all experimental designs utilize blindness to minimize bias, blindness is not a feature of observational designs. Neither in observational design nor experimental ones are investigators required to manipulate the outcome. Experimental designs, like observational ones, are conducted in human as well as animal populations (Aschengrau, A.,& Seage, G.R., EssentialsofEpidemiology, 2003, pp. 163-165).

The correct answer is E (internal validity of study). Confounding is derived from Latin "coniondere", meaning lito pourtogether;' and it indicates the conCHAPTER 2:PRINCIPLESAND METHODSOF EPIDEMIOLOGY 89 fusion of two supposedly causal variables so that part or all of the purported effect of one variable (exposure) is actually due to the other (confounder). Confounding variable is considered an extraneous variable because it competes with the independent variable in explaining the outcome. In a study conducted to assess whether exposure (A) causesdisease (8),Cis a confounding if (a) C (smoking) is associated with A (coffee drinking) in the population that produced the cases 8 (pancreatic cancer), (b) C(smoking) is an independent cause or predictor of 8 (pancreatic cancer), or (c) C (smoking) is not an intermediate step in the causal pathway between coffee drinking (A) and pancreatic cancer (8) (Hennekens, C.H., & Buring, J.E., Epidemiology in Medicine, 1987, pp. 287-293; Gordis, L., Epidemiology, 3rd ed., 2004, pp. 228-229).

The correct answer is C (study validity-bias). Because men with diabetes mellitus are more heavily monitored compared to men who are not, disease ascertainment may be better in the monitored population and may introduce surveillance bias (monitoring of one population [diseased] more than the other [disease-free]). This bias may lead to erroneous estimation of the effect of the disease in the monitored population, thus inflating the point estimate (RR, OR) (Gordis, L., Epidemiology, 3rd ed., 2004, p. 227).