Observational epidemiological studies Flashcards
Definition of incidence rate?
- Number of new cases of a specific disease arising within a popn over a specified time period divided by the person-time accumulated
= (no. new cases) / (no.s of person-years accumulated)
= #new cases/ #ppl at risk in a given time frame
Definition of risk?
- Number of new cases of a specifc disease arising within a popn over a specified time period divided by the number of persons at risk at the beginning of the time period
= (no new cases) / (no of persons at risk at beginning of obs period)
Definition/ formula for point prevalence
= (No of persons with disease at some time point) / (total popn at risk of disease at the same time point)
What are descriptive studies
- Descriptive studies characterise disease in terms of time, place and person with several aims.
When would a cross-sectional study be undertaken?
- If in a particular geographical region we were to determine the point prevalence of a particular disease -> then undertake cross sect. study (snapshot of region at one point in time)
- Describe what is happening at the present moment
- Used to determine the prevailing chara. in a popn at a certain point in time
- How many exposures and outcomes can we count???
Advantages of a cross-sectional study
- Inexpensive
- Quick
- Gather info about diff groups in short amount of time - study multiple exposures and outcomes
- Can identify cohort for follow up
- Few ethics
- Can use entire popn
Disadvantages of a cross-sectional study
- Time sequence and causation (disease and exposure measured simultaneously and so sometimes hard to interpret results in terms of cause and effect)
- Cohort effects when interpreting relationships with age
- Problems with interpretation of prevalence which is a mixture of incidence and survival relating to a disease
- Large sample sizes necessary to generate usable information (reqs. entire popn to be studied to create useful data)
- Selection bias/ information bias/ confounding/ length bias (long disease courses have higher prevalence so it is overestimated)
- Does not address temporality
- Not for rare/ latent diseases
What are ecological studies
Studies in which the units of obs are groups of individuals rather than indiv themselves, for example, popns of diff countries, towns, health regions/ districts
- Associations between exposures and disease at popn level
- Measure rate of death or disease in popns and the popn rate of a RF
- can be used for hypotheses generation
- provides snapshot of the popn vs the indiv
Advantages of ecological studies
- Quick
- Inexpensive (uses routine data)
- Hypothesis generation
- Useful for popn exposure
Disadvantages of ecological studies
- The group of people described by disease data are often not the same people as those described by the exposure data
- > Cannot assume from a relationship found amongst groups of indiv that the same relationship also holds at the indiv level - ECOLOGICAL FALLACY - Outcome measures likely to be biased and inadequate as data not usually collected for all members of a popn?
- Subject to confounding which cannot be controlled
- Prone to sampling bias - as we can only use the info available to us which may not be representative of the whole popn
- Prone to information bias as recording may differ between popns
- The time relationship between the measurements for exposure and disease is often unclear - when was the data collected for exposure/ disease?
Definition of cohort studies
Studies in which a group of indivs who are defined on the basis of their exposure to a RF and are then followed up over time to determine who develops the disease of interest
-> can be defined either prospectively or retrospectively
Advantages of cohort studies
- Provide estimates of absolute and cumulative (+relative?) risks of disease
- Particularly useful for following up rare exposures
- Eliminate some of the sources of bias which are unavoidable in case-control studies which rely on recall of past events, i.e.g if carefully conducted then recall bias is eliminated
- Little doubt about cause and effect because exposure is measured before disease occurs
- Unlike case control studies that address a single disease, there is no limit to the disease for which the risk of occurrence may be related to exposure - investigate multiple outcomes
Disadvantages of cohort studies
- Not useful for v rare diseases - as few of the indiv exposed will develop the disease (case control would be better ) - a lot of exposed indivs would need to be recruited to be sure of enough outcomes occurring
- Often involves following up large numbers of people over decades they are are EXPENSIVE and DIFF to maintain, particularly from a funding stream perspective (need to wait for outcome to occur after recruiting on the basis of exposure) - loss to follow up problem for diseases with long latency periods
- If disease is rare and we only need relative risks assoc with different levels of exposure to RF then unnecessarily expensive
- Can only investigate ONE exposure
Definition of case-control studies
Studies in which a group of indiv with a particular disease (the cases) are retrospectively compared with an app. selected group of controls in relation to their previous level of exposure to the RF of interest (controls are individually matched) - cases have disease and the controls do not
- Observational retrospective
Advantages of case-control studies
- Effective use of resources to study the aetiological factors for rare diseases
- Explore multiple exposures (RFs) for a disease
- Can estimate relative risk of disease associated with different levels of exposure to a RF through the odds-ratio approximation to the relative risk
- Recruitment is based on presence or absence of outcome, so do not have to wait for this to occur - good for diseases with long latency periods
- Groups for comparison are based on the presence of the outcome: active case finding overcomes the problem of rare outcomes
- Quick/ cheap
- Small sample sizes
Disadvantages of case-control studies
- Recall bias, e.g. AXR and childhood cancer
- Address single disease only (can investigate only one outcome)
- Can’t estimate absolute risk or incidence rate
- Can’t calculate incidence or prevalence?
- Temporal relations?
Sources of controls in case-control studies
- From underlying cohort from which the cases were ascertained
- Hospital controls
- GP controls
- Neighbourhood controls
- Random digit dialling by telephone
What are the different sources of bias in case control studies?
- Recall bias
- Recording bias
- Interviewer bias
- Response bias
- Sampling bias
What is relative risk?
(Incidence rate of disease among those exposed to the factor)/ (incidence rate of disease among those unexposed to the factor)
What is the criteria for assessing causality (BRADFORD HILL CRITERIA)
- Biological plausibility - does it make biological sense
- Time - Logically a cause must PRECEDE its potential effect
- Strength of association - Stronger ass of an exposure with disease occurrence then the harder to conceive of likely confounders which might explain the association
- Biological gradient or does-response relationship - causality as a plausible interpretation is strengthened if there is a strong dose-response
- Consistency - Consistent with other studies in diff popns, places and times add weight to a cause-effect interpretation
- Specificity - if the supposed cause if ass. with one disease only, or the disease is ass with one cause only can add weight to a causal interpretation
- Coherence - should not contradict what is already known about the natural history and biology of the disease
- Experiments - occasionally natural experiments offer themselves, such as tap water fluoride levels and specific disease outcomes
- Analogy - v rare and potentially dangerous
What are the aims of descriptive studies?
- To alert the medical community to what types of persons (Age/ sex/ social class/ geographical region/ calendar year) were most/ least affected by disease
- To assist in the EB planning of health and medical care facilities (e.g. how many oncology beds are needed according to health region)
- To provide suggestion concerning disease aetiology for further investigation using analytic studies (hypothesis generation)
What are the sources of bias in cohort studies?
- Healthy worker effect
2. Lost to follow up bias
How can adjustment for confounding be achieved in cohort studies?
- Achieved by use of standardisation, stratification or use of a suitable regression model
- Indirect/ direct standardisation
What are the different possible regression models for cohort study analysis?
- Linear regression model
- Logistic regression model
- Cox regression model
- Poisson regression model
What is RECALL bias?
Among study subjects the extent of recall differs systematically between cases and controls
What is RECORDING bias?
More illnesses and more info recorded in medical records
What is INTERVIEWER bias?
Non-neutral approach to interview cases and controls
What is RESPONSE bias?
If response rate is LOW or differs between cases and controls
What is SAMPLING bias?
Ig not clear what comprises the underlying cohort
