Study Designs - Chapter 7

Question 1

Observational study

Answer 1

does not involve any intervention, experiment or otherwise. Such a study may be one in which nature is allowed to take its course, with changes in one characteristic being studied in relation to changes in other characteristics

Question 2

Experimental

Answer 2

exposure conditions are under the direct control of the investigator

Question 3

Two types of Observational studies: Descriptive

Answer 3

**Descriptive: **
* distribution of disease by person, place and time
* **Little information available **on phenomena
* NO prior hypothesis about exposure (E) →outcome (O) relationships
* Hypothesis generating
* Designed only to describe existing distribution of variables; who, what, when, why, where
* Study association not causation; things are correlated,

Question 4

Two types of Observational studies: Analytic

Answer 4

determinants of a disease by testing the hypotheses formulated from descriptive studies, with the ultimate goal of judging whether a particular exposure causes or prevents disease [association]
Testing hypothesis about exposure outcome relationship

Question 5

D - Case reports

Answer 5

In-depth, textual description of a single patient
Describe somethign unusual to alert others
Limitations
- cases are not randomly selected
- no comparision population
- cannot test hypothesis

Question 6

D - Case series

Answer 6

Textual or statistical analysis of cases

Question 7

D - Descriptive studies based on rates

Answer 7

• Combine data on **population based set of cases with denominator **data
• Quantify burden of disease
  ** Hypothesis generating

Question 8

D - Ecologoical studies

Answer 8

• Studies of groups
• Correlation between group measures of exposure and outcome
• unit of analysis is group
• geographical comparision
• time trend analysis
  Limitations
• cannot adjust well for confounders
• prone to ecological fallacy
• wide variability in summary of exposure

Question 9

Analytic - Cross-sectional

Answer 9

• at one point in time
• snapshot
  * study of prevelance
• temporal sequence cannot be determined
• lenght bias (people with long duration of disease overrepresented, people with rare short duration underrepresented)
• exposure outcome relationship studied at the same time
• Has a comparision group
• impossible to ascertain temporal sequence as E & O is being studied at the same time
  Limitations
• short duration disease get under represented
• long duration disease get over represented
• Prevalence-incidence bias
• unsuitable to study rare outcome

Question 10

Analytic - Case-Control

Answer 10

• has a comparision group
• outcome comes first and then we go back to find out what exposure was

Limitations
- Risk of the disease cannot be obtained because cases and control are determined by the investigator it is not a random selection, not generalizable

Question 11

Analytic - Cohort

Answer 11

• select a cohort (a group of people)
• observe who is exposed and who is unexposed
• “wait” for the outcome of interest
• Compare the outcome among the exposed and unexposed

Question 12

Hierarchy of populations

Answer 12

1. Target pop: Largest component; results may be generalized ; iv drug users in canada
2. Source pop: Sampling frame; where do we see the iv drug users, hospitals, clinics etc.
   3. Eligible pop: Intended sample, random sample of hospital and clinic where they are found
   4.** study pop:** those who are eligible and said yes

Question 13

Absolute must to take Incident cases in Case-Control study Why?

Answer 13

**Incident cases **= Researchers can be **more confident about the temporal sequence **as the new cases got exposed to something and now we see the outcome

If we select **Prevelant cases **= they already have the disease, we dont know how long ago or what is the duration, there is recall bias introduced here

The reason is that any risk
factors we may identify in a study using prevalent cases may be
related more to survival with the disease than to the development of the disease (incidence).

Question 14

Primary Base for Case-Control study

Answer 14

Defined by the population investigator wishes to target. all cases assumed to be identifiable. e.g. population of Canada

Question 15

Secondary Base for Case-Control study

Answer 15

if we cannot identify all the cases in the population, we first identify the cases and then define the study base, or where they came from. e.g. cases = patients treated with melanoma in UAH ; Secondary base = all patients who would have been treated in UAH if they had melanoma

Question 16

Benefits of cross-sectional study and cohort study

Answer 16

Benefits of the repeated cross-sectional study design
* Not impacted by aging
* Better estimate of prevalence

Benefits of the **cohort study **design over the repeated crosssectional
study?
* Allows for studying temporal associations
* Allows for studying incidence

Question 17

Cases

Answer 17

subjects who have developed the outcome of interest, have the disease

Question 18

Controls

Answer 18

** Do not have the disease; **subjects who reflect the exposure pattern for the source population (population at risk) from which the cases arose

Question 19

Neighbourhood controls

Answer 19

interviewers are instructed to identify
the home of a case as a starting point, and from there walk past a
specified number of houses in a specified direction and seek the
first household that contains an eligible control.

Question 20

Selecting controls using Random Digit Dialing

Answer 20

sampling of residential phone numbers; limitations gives us the random sample of phone number not the controls not all controls have phone numebrs.

Question 21

Matching in case-control studies

Answer 21

Matching is defined as the process of selecting the controls so that they are similar to the cases in certain
characteristics, such as age, race, sex, socioeconomic status, and
occupation. Matching may be of two types: (1) group matching (proportion of control is same s proportion of cases) and
(2) individual matching. (e.g. 45 yr old white women case is matched with 45 yr old white women contrl

Question 22

case-control strength vs weakness

Answer 22

Strength
* Efficient for** rare diseases** and diseases with long induction and latent period. [because we already have the data with us, we are going back in time to identify the exposure]

• Can evaluate many risk factors for the same disease [many exposure could have caused the disease]

Weakness
* Inefficient for rare exposures

• Vulnerable to selection bias
• Vulnerable to information bias
• Difficult to infer temporal relationship between exposure and disease

Question 23

Hospital controls

Answer 23

used most often cases selected from a hosptial population.
Illnesses that make ‘good hospital controls’ are those UNRELATEDto the risk factor(s) under study.

Question 24

Selection of control in a case-control study

Answer 24

Cumulative - controls selected at the end of the specified observation period
Case cohort - controls taken at the begining of the observation period; these controls can later become cases.
**Incidence density case-control: **for each case, 1+ controls are randomly sampled from the cohort at risk at the time of an incident case event. Sampled controls can later become cases.

Question 25

types of **exposure** and **outcome** that can be studied in **Cohort study**

Answer 25

**Multiple exposure can be studied at the same time**; any exposure lifestyle, chronic, point, socio demographic factors Any agent - pathological agents, **Any outcome **can be studied mortality, morbidity...

Question 26

Types of cohort

Answer 26

1. **Closed:** enrollment at a closed (fixed) time, no one joins all enrolled at one poitn in time. 2. **Open: **continued enrollment 3. **Dynamic:** participants come and go

Question 27

type of cohort studies

Answer 27

**Prospective: **going forward in time [most common] EXPOSURE OUTCOME ASCERTAINED AS THEY OCCUR IN THE STUDY **Retrospective/historical:** identifying the cohort back in time .e.g. exposure is somewhere in the past ; EXPOSURE ASCERTAINED FROM PAST RECORDS, OUTCOME DETERMINED WHEN STUDY BEGAN

Question 28

when do you use a case-control study?

Answer 28

At an early stage **in our search for an etiology,** we may suspect any one of several exposures, but we may not have evidence, and certainly no strong evidence, to suggest an association of any one of the suspect exposures with the disease in question. Using the case-control design, **we compare people with the disease (cases) and people without the disease**. We can then explore the possible roles of a variety of exposures or characteristics in causing the disease. First step in determining if an exposure is related to the increased risk of the disease. ; **after this we can do a cohort study to further elucidate this relationship.**

Question 29

Repeated cross-sectional studies

Answer 29

Same **information collected from independent samples over time**. examine changes in population over time. e.g. prevalance of a disease

Question 30

Stratification

Answer 30

Dividing your sample into sub-group based on certain characterisitics; it helps to study heteroginty across and within groups

Question 31

Ecological fallacy

Answer 31

“The bias that may occur because an association observed between variables on an **aggregate level does not** necessarily represent the **association** that exists a**t an individual **level.”

Question 32

what do you need to keep in mind when selecting controls in a case-control study?

Answer 32

Make sure tht the cases and controls are coming from the same study population.

Question 33

In case-control studies, risk of the disease CANNOT be obtained because…

Answer 33

…the proportion of cases vs. controls in the study population is determined by the investigator . it does not reflect the true prevalence of the disease in the population

Question 33

When OR is a good(ish) estimate of RR?

Answer 33

When casesare representative, with regard tohistory of exposure, of all people with the disease in the population from which the cases were drawn. * When the controlsare representative,with regard tohistory of exposure, of all people without the diseasein the population from which the cases were drawn. * The risk of the disease is low (rare disease assumption).

Question 34

Why is it better to take incidence cases rather than prevalent cases in case-control study?

Answer 34

1. incidence cases helps in temporal sequence of exposure and outcome 2. Prevalence case can introduce recall bias if disease happened long ago. 3. Prevalence - More likely to include long-term survivors

Question 35

Primary study base

Answer 35

The base is defined by the population experience that the investigator wishes to target * The cases are subjects within the base who develop D * All cases are assumed to be identifiable * Example: population in Canada in 2022

Question 36

Secondary base

Answer 36

If ascertainment of all cases is not possible →we identify cases, and they define the study base * Controls are people who would have been recognized as cases if they had developed disease * Example: cases are patients treated at the UAH for melanoma in 2022; the base would be all patients who would have been treated at UAH if they developed melanoma in 2022