Chapter 7- Observational studies

Question 1

Case report

Answer 1

An individual-level observation. Physicians can do this to describe a particular phenomenon in one patient or a series of patients with a similar problem. Along with case series, they are considered the simplest of study designs, or “pre-study” designs

Question 2

Main objective of case reports and case series

Answer 2

To provide a comprehensive and detailed description of the cases under observation. This allows other physicians to identify and potentially report similar cases from their practice, especially when they share geographic or specific clinical characteristics. Case reports and case series are descriptive- they have no reference group to compare the observations to. However, the Brazilian case series regarding the association between Zika virus and microcephaly was important for developing CDC guidelines

Question 3

Advantages and disadvantages of case reports and case series

Answer 3

Advantages- simple, inexpensive, and easy to conduct in a busy clinical setting
Disadvantages- lack of a major comparison group, generalizability is limited due to biased selection of cases. Any observed association is also prone to confounding

Question 4

Ecologic studies

Answer 4

A study of group characteristics- generates the average values of two groups. These studies do not take into account any variation between the groups We may be ascribing to members of a group a characteristic that they do not possess as individuals. Often, researchers will use data that is easy to obtain and does not require following up or direct contact with the participants

Question 5

Advantages and disadvantages of ecological studies

Answer 5

Advantage- can be used to suggest avenues of research that may be promising in casting light on etiologic relationships. Can be used to study the relationship between an individual and the community in which they live- some relationships are only strongly apparent in ecological studies
Disadvantage- ascribes characteristics to individuals that only apply to groups, does not demonstrate conclusively that a true association exists

Question 6

Cross-sectional studies

Answer 6

Both exposure and disease outcome are determined simultaneously for each participant. It’s like we were viewing a snapshot of the population at a certain point in time. If looking at the relationship between CHD and cholesterol, patients would all be screened for high cholesterol and CHD at the same time. This approach identifies prevalent cases of disease- we know they existed at the time of the study but not their duration. Therefore, it’s not possible to establish a temporal sequence between the events- may cause temporal bias. Political polls and sample surveys are examples and are often the first studies conducted before moving on to more valid study designs

Question 7

How can survival (selection) bias occur in cross-sectional studies

Answer 7

Occurs when the exposure is related to the duration of the disease. If exposure-induced incident cases have a shorter survival than unexposed incident cases, prevalent cases (survivors) may have a lower proportion of past exposure than those that would have been observed if incident cases had been included

Question 8

Prevalence-incidence bias

Answer 8

Survival of patients with smoking-induced emphysema is worse than patients whose emphysema is induced by other causes. Therefore, past smoking will be observed less frequently in prevalent than incident cases

Question 9

What are the results of a cross-sectional study used for?

Answer 9

To generate hypotheses that can then be evaluated using a study design that includes incident cases and allows establishing the temporal sequence of the exposure and then the outcome.

Question 10

4 possible groups for each participant in a cross-sectional study

Answer 10

1. Exposed and has the disease
2. Exposed but does not have the disease
3. Not exposed and has the disease
4. Not exposed and does not have the disease

Question 11

How to determine if there is evidence of an association between exposure and disease from a cross-sectional study

Answer 11

You can calculate the prevalence of disease in people with the exposure and compare it with the prevalence of disease in people without the exposure. OR, you can compare the prevalence of exposure in persons with disease to the prevalence of exposure in persons without the disease

Question 12

Meaning of the results of a cross-sectional study

Answer 12

In a cross-sectional study, you are identifying prevalent rather than incident cases, so the prevalent cases may not represent all cases of the disease in the population. The association may just represent survival after the disease rather than the risk of developing the disease. It’s also not possible to demonstrate a temporal relationship between exposure and outcome since both variables were measured at the same time- it may be reverse causality. However, the findings could reflect the need to develop another, more valid study. Serial cross-sectional studies can also be useful to evaluate trends in disease prevalence over time, in order to inform healthcare policy and planning

Question 13

Case-control studies

Answer 13

To examine the possible relation of an exposure to a certain disease, we identify a group of individuals with that disease (cases) and a group of people without that disease (controls). Then, we determine what proportion of the cases and of the controls was exposed and what proportion was not. If there is an association of an exposure with a disease, the prevalence of history of exposure should be higher in cases than the controls. The number of cases and controls selected is up to the researcher

Question 14

Dichotomous exposure

Answer 14

When the exposure either has occurred or has not occurred. In this case, participants can be broken down into 4 groups. Cases are broken down into exposed or not exposed, and same with controls

Question 15

Selection bias in case-control studies

Answer 15

Cases can be selected from hospital patients or people who seek healthcare at clinics. If cases are selected from a single hospital, any risk factors that are unique to that hospital could introduce bias and mean that the results aren’t generalizable other patients with the disease. Researchers also must consider whether to use incident or prevalent cases. With incident cases, you must wait for the cases to be diagnosed. It’s generally preferable to use incident cases because using prevalent cases can introduce incidence-prevalence bias

Question 16

Selection of controls in case-control studies

Answer 16

Controls can be selected from nonhospitalized persons living in the community, from outpatient clinics, or from hospitalized patients admitted for diseases other than that for which the cases were admitted. In terms of nonhospitalized people, they could be recruited from certain neighborhoods, voter registration lists, or other sources.

Question 17

Information bias

Answer 17

Patients may have difficulty recalling past exposure and will report it inaccurately. Since the majority of the information in the study is obtained through interviews, this can present a significant limitation if the information provided is inaccurate

Question 18

Recall bias

Answer 18

Also called rumination bias. Patients or parents of patients with an illness or disability may report minor details that people typically would not remember or report because they want to figure out the cause of the illness if it is unknown. There is often differential recall between cases and controls

Question 19

Matching in case-control studies

Answer 19

Matching is the process of selecting the controls so that they are similar to the cases in certain characteristics. In case-control studies, cases and controls may differ in characteristics or exposures other than the one that has been targeted for study. There is group matching and individual matching

Question 20

Group matching

Answer 20

Selecting the controls in such a manner that the proportion of controls with a certain characteristic is identical to the proportion of cases with the same characteristic. The proportions generally won’t be exact

Question 21

Individual matching

Answer 21

When, for each case is selected for the study, a control is selected who is similar to the case in terms of the specific variable(s) of concern. Often used in cases using hospital controls

Question 22

Multiple controls

Answer 22

In case control studies. multiple controls for each case are frequently used. This increases the statistical power of the study. You can choose multiple controls of the same type (multiple controls for each case) or multiple controls of different types- when there are multiple control groups used in a study

Question 23

When is a case-control study warranted?

Answer 23

Useful as a first step when searching for a cause of an adverse health outcome. When cases and controls are compared, we can explore the possible roles of a variety of exposures or characteristics in causing the disease. If the exposure is associated with the disease, we would expect the proportion of cases who have been exposed to be greater than the proportion of controls who have been exposed. After an association has been discovered, a cohort study is usually conducted

Question 24

Case-crossover design

Answer 24

Used for studying the etiology of acute outcomes like heart attacks or deaths from acute events. In this study, a case is identified and the level of environmental exposure is ascertained for a short time period preceding the event. This level is compared with the level of a control period that is more remote from the event

Question 25

Descriptive observational designs (2)

Answer 25

Case reports & case series

Question 26

Analytic observational designs (4)

Answer 26

1. Ecologic designs
2. Cross-sectional studies
3. Case-control studies
4. Cohort studies

Question 27

Descriptive epidemiology

Answer 27

Analysis of the distribution
of determinants, health-related
events, & outcomes. Asks the questions who, when, and where. How does disease compare across subpopulations? Are there patterns of disease across time? Are there patterns of disease across spatial units?

Question 28

Analytic epidemiology

Answer 28

Study of determinants/factors
as causes of health-related
events & outcomes. We aim to identify whether exposure to a
determinant/factor is a cause of a health-related event or outcome

Question 29

Predictive epidemiology

Answer 29

Use distributions to forecast
outcomes, but not based
solely on causes & effects

Question 30

Observational study

Answer 30

A type of study where an investigator observes a relationship between a factor of
interest (exposure, risk factor, population, etc.) and a health outcome but does
not intervene

Question 31

Case reports and case studies

Answer 31

Published details of a single (report) or multiple
(series) cases that represent findings of clinical
interest. May be referred to as “prestudy designs

Question 32

Limitations/considerations of case reports or case studies

Answer 32

1. No control group
2. Potential selection bias

Question 33

Ecological fallacy

Answer 33

incorrectly ascribing
“group-level” characteristics to individuals- group associations can be misleading. Although Florida has high rates of both ice cream consumption and skin cancer, that does not mean that ice cream can cause skin cancer

Question 34

When are ecological studies useful? (4)

Answer 34

1. Help to generate/explore hypotheses
2. Aggregate level data may be more readily available and less expensive
3. Examining “ecological” exposures: environmental exposures, laws or policies, community-level interventions
4. When exposure differences between populations are larger than within populations
   When are ecologic studies useful?

Question 35

Limitations of cross-sectional studies (2)

Answer 35

1. Temporality + reverse causality
2. Selection bias from “survivors”- relationships may reflect survival after disease rather than risk of disease

Question 36

Reverse causality

Answer 36

If there is an association between multivitamins and metabolic disorders, people might assume that the multivitamin caused the metabolic disorder. However, it’s possible that patients started taking the multivitamin after their diagnosis to help their symptoms

Question 37

Causal pies

Answer 37

A pie where each slice is a factor contributing to the development of a specific disease. Smoking and asbestos exposure could be part of a causal pie for lung cancer

Question 38

Population control recruitment for case-control studies (3)

Answer 38

1. Neighborhood controls
2. Random digit dialing
3. “Best-friends”- ask a participant whether they have a friend that wants to participate

Question 39

Why choose different approaches/methods of selecting controls over others?

Answer 39

1. Convenience
2. Some methods may be regarding as better ascertaining the “source population” or less susceptible to
   bias
3. Some studies may use MULTIPLE sources of controls to mitigate potential bias

Question 40

Cohort studies

Answer 40

Start with general population without disease. At baseline, measure their exposure & any changes in the exposure over time. Follow each person until disease/study ends, then compare disease rates between exposed & unexposed groups