Control case studies

Question 1

What is a spurious correlation

Answer 1

Spurious correlation, or spuriousness, occurs when two factors appear casually related to one another but are not. The appearance of a causal relationship is often due to similar movement on a chart that turns out to be coincidental or caused by a third “confounding” factor.

Question 2

What is an odds ratio and example

Answer 2

Ratio of two odds

• Likelihood of an outcome
• Measure of association between exposure and outcome…

Question 3

Cross sectional vs cohort stufy

Answer 3

• Interested in exposure and outcome
• Through some type of observational study
• Observe, not manipulateMight observe - risk, group frequency, treatment
• But we do not change or alter this.
• Association -> difficult to show causality
• Cross-sectional studies
  
  • Analyse data from the population at a specific point in time
• Cohort study
  
  • Subset of defined population who are, have been, or will be exposed “analysed” at specific time intervals – cross-sections

Question 4

Case control study - why? and the epidemiological designs

Answer 4

Observation of 2 existing groups:

• Differ in outcome
• Compare those who have the condition (ie cases) AGAINST
• Those who do not have the conition/disease but otherwise similar (ie controls)

Question 5

How to interpret an odds ratio?

Answer 5

> 1 = more association

< 1 = less association

~1 = equally likely

Association is correlation and correlation isnt alwyas ewual to causations

Question 6

WHat are the strengths and weaknesses of a case control study?

Answer 6

Pros:

• Cases already exist – quick
• Cheap – can manage smaller sample sizes
• Flexibility in how many exposures you consider
• Great for uncommon/rare diseases - Or preliminary research
• Tend to have greater statistical power than alternatives i.e. cohort study- Not waiting for people to develop disease

Weakness:

• Observational – less causative in design
• Finding a result doesnt always reflect reality - comes form good design.
• Easily confounded – depends on what exposures you ask for. Influences both exposure and outcome. Should control for ensure exposure causes outcome and not oconfounder.
• Colliders - exposure and outcome independenlty influence third variable. Can obscure real or reveal false assoocations/
• Small numbers = difficult to study rare exposures
• Exposures depend on participant recall accuracy
• Difficult to match controls
• Prone to selection bias
• Difficulty in determining timeline of exposure to disease outcome

Question 7

WHo do we select in a case control study

Answer 7

• Sample from limitless population
• Much like case definition - Clear definition of the outcome
• Focus on incidence cases than prevalent cases - Diagnostic criteria change all the time
• Appropriate controls
  
  • Free from outcome i.e., disease of study
  • Should have potential to become cases
• Matched case-control - Clear definition of demographics
  
  • i.e., our cases need to be as like our controls as possible
  • Demographics are potential confounders

Question 8

Population studies and their weaknesses

Answer 8

• In the various fields of healthcare, a population study is a study of a group of individuals taken from the general population who share a common characteristic, such as age, sex, or health condition.
• Selected from the population /Assured cases and controls come from same population.Usually a lot of people to choose from
• but time ocnsuming to entine participants, therefore expensive. Cooperation is difficult to achieve and maintain.

Question 9

Hospital based studies and their weaknesses

Answer 9

• Cases and controls selected from admission in hospital
  
  • Likely similar across cases and controls - Same factors led both to hospital
  • Easy to identify and access- In hospital with contact information. Less expensive and time consuming as a result
  • Generally, more willing to participate
• However…
  
  • Actual patients and therefore ill - May not represent exposure history accurately, Confounders
  • Hospital catchment areas vary

Question 11

Confounder and collider

Answer 11

Question 12

Bias in case controls?

Answer 12

• At all stages, there is risk of bias
  
  • Something in your collection, analysis, interpretation, publication or review of data leads to conclusions that are systematically different from the truth
  • Known or unknown
• We talked about recruitment
  
  • Well, that can lead to selection bias as discussed
  • No random sampling
  • Noncompliance
  • Who volunteers for research? Do you?
• Information/ observation bias:
  
  • Retroscpective measurement of exposure. How accurate, have they estimate dcorrectly, memory fills gaps with plausible data. Spuriou sassociation.
  • Meausrmeent error -difference between measured value and true value
  • Confirmaiton bias - see what we expect to see

Question 13

Odds ratio calculation

Answer 13

Question 14

What type of study design is a case-control study?

Answer 14

Observational. Case-control studies are a form of observational design used in epidemiological and clinical research. They are “observational” in the sense that investigators DO NOT allocate the exposures, risk factors, or other forms of interventions under investigation to participants. Investigators evaluate the effects of observed exposures that happen to people during the usual course of their lives and healthcare. They are used to investigate the causes of disease; specifically they can be used to tell us the strength of the relationship between one or more risk factors and a health outcome of interest. Classic example are studies in the 1950s that established the link between lung cancer and smoking. Modern case-control studies include genome-wide association studies (GWAS) that look to investigate which genetic factors may be associated with a particular disease.

Question 15

Summarise the key characteristics of the case-control study design

Answer 15

Cases who have experienced the outcome of interest (e.g. with incident disease) and controls who have not experienced the outcome of interest (e.g. without incident disease) are compared with respect to their level of exposure to a suspected risk factor

 Unlike cohort studies, case-control studies START WITH THE DISEASE OUTCOME and LOOK BACK AT PRIORHISTORY OF EXPOSURES (“all effects are already produced before the investigation begins”)

 The association between exposure and outcome is usually summarised as an ODDS RATIO

Question 16

Under what circumstances are case-control studies generally most useful?

Answer 16

They can be an efficient way of looking at LOTS OF DIFFERENT RISK FACTORS and is often QUICKER and LESS COSTLY than starting up a new cohort of people and following them over years to see who develops the disease of interest

 They are particularly useful for relatively RARE DISEASES

Question 17

What are the main possible weaknesses to consider in the case-control design?

Answer 17

 There can be problems with RECALL BIAS when investigators rely on asking cases and controls to remember past exposures

 It is important that controls come from the same underlying population as the cases. If not, there is a risk of SELECTION BIAS

Question 18

What is blinding

Answer 18

• Randomly allocating groups but want to hide who is in which group
• Single blind = participants unaware which treatment they are getting
• DOuble blind = Experimenter and particpants unaware
• If those involved are unaware then difficult to be biased.

Question 19

Types randomised control trials

Answer 19

▸Trial type

▹Superiority

▹Noninferiority

▹Equivalence

▸In essence, RCTS might consider

▹The superiority of new treatment

▹The noninferiority of the new treatment

▹The equivalence of the two treatments

▸Placebo

▹If a proven treatment already exists, avoid a placebo