Week 5 Flashcards
What are types of Observational studies?
- Case reports and case series
- Ecologic studies
- Cross-sectional studies
- Case-control studies
- Cohort studies
What is an observational study?
Epidemiologic studies where the investigators collect, record and analyze data on subjects without controlling exposure status or conditions of the study
What is a Case reports and case series study?
An unusual health event or series of events can identify a new problem and – sometimes – suggest a potential cause. (ex: covid) (people with similar symptoms)
Case report is a detailed description about an individual patient (specific to 1 person - 1 case)
Case series is an extension of the case report that describes the characteristics of a group or cluster of individuals with the same condition. (a few dozens of cases)
What are Ecologic studies?
- Studying a group’s characteristics to determine an association (association only on the characteristics of the group)
- Overall or summary measure of exposure and outcome
- Inexpensive, relatively quick, establish preliminary hypothesis at the group level, likely based on existing data resource
What are Cross-sectional studies?
examines the relationship between exposure and disease prevalence at a single point in time (ex: For diabetes, cancer, etc or positive mental health outcomes)
descriptively and analytically
- descriptively: characterize the prevalence of a health problem in a specified population
- analytically: compare the proportion of exposed persons who are diseased with the proportion of unexposed persons who are diseased.
How does sampling work for cross-sectional studies?
Target population: people you want your results to apply to (ex: population of Ottawa, you can gather from data base)
Source population or sampling frame work: people from whom the population is selected (random sampling of stratification from Ottawa citizens)
Sample: people approached to take part in the survey
Study population: people who actually took part in the survey
When are cross-sectional studies used?
- Estimate the occurrence of risk factors in segments of the population characterized by social determinants
- Use for planning or administering preventive or health care services, surveillance programs, surveys and polls
- Establish preliminary evidence for a causal relationship between an exposure and an outcome
What is a cohort study?
Study population
Exposure -> Outcome
No exposure -> Outcome
What are types of cohort studies?
- Prospective/longitudinal cohort studies (Study in the present and outcome in the future)
- Retrospective/historical cohort studies (Study in the past and outcome is in the present)
- Mixed cohort studies (ambidirectional cohort studies) (study from the past and now and then look at outcome in the future)
What can you calculate with an open cohort?
Open = dynamic - different people in the city, people die everyday so what you observe is different everyday
Incidence rate
What can you calculate with a closed cohort?
(assumption of no loss to follow-up and no losses due to competing risks - no changes)
Cumulative incidence or incidence rate
What can you calculate with a fixed cohort?
(when loss to follow-up occurs for closed cohort)
Incidence rate
What can you calculate in a contingency table?
Total of exposed
Total of non-exposed
Total of cases
Total of non-cases
What are Indicators for cohort study?
Incidence rate (Absolute risk)
- Exposed group: A/Texposed
- Nonexposed group: C/Tnonexposed
Risk ratios (when exposed how likely will you get the disease)
- RR=(A/Texposed)/(C/Tnonexposed)
Risk difference
- RD=(A/Texposed)-(C/Tnonexposed)
Attributable risk
- AR=[(A/TExposed)- (C/Tnonexposed)]/ A/TExposed
Population - attributable risk (Incidence of disease in a population, associated with the prevalence of a risk factor)
- PAR=AR*P
- P=prevalence of exposure to a risk factor=(A+B)/Total
Population-attributable fraction (fraction of disease in a population is attributable to exposure to a risk factor)
- PAR%= PAR/ [(A+C)/Total]*100
How to interprate the risk ratio (or relative risk)?
(a/a+b) / (c/c+d)
Ratios > 1.0 indicate risk is higher among exposed than unexposed (risque)
Ratios = 1.0 indicate no association
Ratios < 1.0 indicate risk is lower among exposed than unexposed (protecteur)
how to calculate the rate ratio?
Only when in an open or fixed cohort:
a/Person years exposed / c/person years unexposed
What is the Population attributable risk?
- Measure the excess incidence of disease in a community that is associated with a risk factor;
Identify the relative important risk factors; - Direct health policy into the more important or the relative weak but prevalent common risk factors.
Choose one you think has a higher risk for stroke.
A 35- year man who smokes
A 60- year man who does not smoke
See slide 27
The relative risk decreases with age. However, absolute risk increases with age, mainly because of stroke is more common in elderly regardless the status of smoking.
Age contributes more than smoking in the occurrence of stroke.
What is the difference between Cohort study and case-control study?
Cohort study
- Exposure at the beginning of study
- Disease after follow-up of all individuals
Case-control study
- Case status from registry
- Exposure on cases and controls
- First identification of persons with the disease and a suitable control of persons without the disease
- Key: comparison between disease + vs. disease -
How does a case-control study work?
Study population
Outcome + -> Exposure?
Outcome - -> Exposure?
How to get suitable controls?
Make sure you cases and controls are comparable
Would your cases have been selected to be controls had they not gotten the disease?
Would your controls have been selected to be cases had they not gotten the disease?
(we will see in another class)
Can we determine the incidence rate from a case-control study?
Why can’t calculate incidence rate for cases and non-cases groups : cuz we only looked at the people who had the disease – we don’t have the total number of people (denominator)
What can we calculate instead?
Odds ratio = (A/C)/(B/D)
If bigger than 1 = exposure is a danger
If smaller than 1 = its a protector factor
What is something we do with our controls sometimes during our case-control studies?
Matching
Try to match a very similar control to a case
See slide 36
What tests can we use for our case-control studies when we have unmatched cases and controls?
- Chi-square 2 × 2 analysis;
- Mantel-Hanszel statistic;
- Fisher’s Exact test;
- Unconditional logistic regression.
What tests can we use for our case-control studies when we have matched cases and controls?
- Chi-square 2 × 2 analysis;
- Mantel-Hanszel statistic*;
- Fisher’s Exact test;
- Conditional logistic regression
Can you think of more than one reason why a matched case-control study could take longer to complete than an unmatched study?
Find matched controls, sometimes more than one per case; (harder to get the samples or get people involved)
Since only the discordant pairs contribute to the statistical analysis, achieving a desired statistical power depends on obtaining a particular number of discordant pairs.
Why bother with matching if it means a longer case-control study?
- We match to eliminate the possibility of the relationship being confounded by the matching variable because both the case and the control are similar for that variable.
- We don’t want to match on too many variables because it will cause an extreme delay in the completion of the study.
What are the advantages for Case control study?
- Quick and cheap;
- Only feasible method for very rare disorders or those with long lag between exposure and outcome;
- Fewer subjects are needed than cross-sectional studies.
What are the advantages for Cohort study?
- Ethically safe; (cuz no intervention)
- Subjects can be matched;
- Can establish timing and directionality of events;
- Eligibility criteria and outcome assessments can be standardized;
- Administratively easier and cheaper than RCT.
What are the disadvantages for Case control study?
- Reliance on recall or records to determine exposure status;
- Confounders;
- Selection of control groups is difficult;
- Potential bias: recall, selection.
What are the disadvantages for Cohort study?
- Controls may be difficult to identify;
- Exposure may be linked to a hidden confounder;
- Blinding is difficult;
- Randomization is not present;
- For rare diseases, large sample sizes or long follow-up are necessary.
What is a type of Advanced case-control designs?
Nested case-control study
This is a case-control study within a cohort study. At the beginning of the cohort study (t0 or base line), members of the cohort are assessed for risk factors.
Cases and controls are identified subsequently at time t1. The control group is selected from the risk set (cohort members who do not meet the case definition at t1.) Typically, the nested case-control study is less than 20% of the parent cohort.
See slide 43
What are the advantages for a Nested case control study?
Efficient – not all members of parent cohort require diagnostic testing
Flexible – allows testing of hypotheses not anticipated when the cohort was drawn (at t0)
Reduces selection bias – cases and controls sampled from the same population
Reduces information bias – risk factor exposure can be assessed with investigator blind to case status
What are the disadvantages for a Nested case control study?
Reduces power (from parent cohort) because of reduced sample size by: 1/(c+1), where c = number of controls per case
What is a Case-cohort study?
In April of 1986, Ross Prentice published a paper in Biometrika introducing the case-cohort design.
This innovative design uses a sub-sampling technique in survival data for estimating the relative risk of disease in a cohort study without collecting data from the entire cohort.
See slide 47
What are the advantages of a Case cohort study compared to nested case-control study?
Efficient – not all members of parent cohort require diagnostic testing
Flexible – allows testing of hypotheses not anticipated when the cohort was drawn (at t0)
Reduces selection bias – cases and controls sampled from same population
Reduces information bias – risk factor exposure can be assessed with investigator blind t0 cases
What are the disadvantages of a Case cohort study compared to nested case-control study?
Increased potential for information bias
- subcohort may have been established after t0
- exposure information collected at different times (e.g. potential for sample deterioration)
How to communicate risk?
Relative risk> absolute risk, e.g. 20 fold vs. 0.00001; (harder to get the absolute risk and people dont get the same feeling hearing that your risk is 20x higher than definitely 0.00001)
“Loss” > “Gain” framing, e.g. risk is more attractive; (prefer to talk about the outcome)
“Positive” > ”Negative” framing, e.g. persuading patients to take health services; (instead of saying you should quit smoking, we say that the nonsmokers will have a bgetter health outcome so we recommend to do the same. Give a recommendation or changes instead of saying to quit)
More information needed;
Where can we find Vital statistics and mortality data for cohort or case-control studies?
Census data (Canadian survey every 5 years - long form)
Civil registration systems
- Registries of births, deaths and
marriages
- National death registers
Public Health Surveillance Systems
- i.e., Canadian Chronic Disease Surveillance system
- Stats Canada??
In most countries, it is a legal requirement that:
- vital events, such as births and deaths, are registered
- a medical practitioner completes a death certificate whenever anybody dies
Deaths are classified according to standardized rules:
- International Classification of Diseases (ICD) (up to version 11)
- coding rules on underlying cause(s) from causes listed on death certificates
a. ensure comparable data over time, between countries, and
b. aid storage, retrieval, statistical manipulation and analysis
In what ways we can use existing data?
Simple description – patterns by person, place and time
- e.g. prevalence of potential causes versus incidence of disease
Migrant studies
- Comparing people who move from a low-incidence area to a high-incidence area (or vice versa)
Ecological or correlation studies
- Comparing the prevalence of exposure and incidence of disease at a population level
- Note: ecological fallacy – patterns seen at a group level may not hold at the individual level
Big data
What is big data?
Volume
- Many variables per record and/or many records
Velocity
- Rapidly available (especially with smart technology)
Variety
- Many different types of data
e.g. sensors in smartphones can measure physical activity, population movements, social networks …
