What are the two main methods used to conduct research?
- Qualitative methods
- complex psychological/social/cultural issues
- e.g. Interviews/discussion groups - Quantitative methods (research methods tend to use this type)
- highly standardised data collection
- structured
What does studying the epidemiology of disease involve?
- Quantitative description of the DISTRIBUTION of the disease
- how much disease is there in any population
- what is the distribution of this disease in time/place/person - Identification of the DETERMINANTS of variation in disease distribution between and within populations
- what factors are associated with an increase in risk of disease
- what factors seem to protect against disease - Testing of the hypothesis
What approach is used to help design your research question?
PICO approach:
- Patients/people/population
- Intervention
- Control/comparison
- Outcome (what is it? How to measure it?)
Following his approach can help refine your general research q into something that can be tested experimentally
Whilst it may only seem relevant to clinical trials it can also be adjusted to apply to studies of causation, diagnostic tests or prediction
Why should you always question the research?
Because associations can exist but may involve a third variable
E.g. Increased ice cream consumption is associated with higher temp which is associated with increased shark attacks. So an association exists between increased ice cream consumption and shark attacks but there is no real link
What are the five main types of quantitative research methods?
- Case-study
- Cross-sectional
- Case-control
- Cohort
- Clinical trial
The strength of evidence they obtain increases as you move down the list
What are the three types of population involved in study design and how are they defined?
Why is this stage important?
- Target population: group that we want to know about
- Study population: group that we choose to sample from
- Participant population: those that actually contribute to the study
Results from (3) will be used to infer things about (1) so it is important that (2) is representative of (1)
This is important because the further removed from (1) your (3) is the less likely you can generalise your results to (1)
What are two examples of sampling approaches?
- Random sampling: every member of population must have equal chance of being selected
E.g. Every 5th person on electoral list - Systematic sampling
E.g. People coming into a clinic ward/GP
What are the four main definition groups for measuring disease?
- Clinical (signs & symptoms - can be difficult to standardise)
- Statistical (test results outside of normal range)
- Prognostic (measurement/biomarker/lab result)
- Operational (different management/treatment plan is needed)
What are the two widely used basic measures of disease in populations?
- Prevalence: the no. of ppl in a population with disease at any given time, divided by total no. of ppl in the population. It is s proportion, not a rate.
- Incidence: the no. of ppl who develop disease over a period of time. It is a rate, expressing the development of disease per unit of time at risk
What is disease prevalence dependent on?
- The number of new cases (i.e. incidence): more new cases = higher prevalence
- Duration of disease: longer duration = higher prevalence.
Explain disease definitions and why adapting them can affect research vs. clinical setting
Disease can be thought of as a binary variable (e.g. Affected/unaffected). It would be nice to have a cut-off so we have two distinct populations but not always this simple and is usually continuous scale and may get some overlap between populations
Disease definitions often need to be adapted to make them suitable for research studies. Therefore definitions differ from research to those used clinically
This can affect how generalisable/translational our results are in a clinical setting
How is incidence measured?
Measuring it involves taking a population of ppl, following them for a period of time and identifying all episodes of disease that occurs
You then calculate the number of ppl who died and adjust the person-time accordingly. You then use the following formula:
Number of persons who died/Total person time at risk
