Week 7 Flashcards
Observational studies
Defined as non-interventional, non experimental
Natural relationships between factors and outcomes
Cross sectional studies
Assessing a population, as represented by a study sample, at a single point in time
Reflect the situation of a disease or clinical outcome in a particular moment in a particular population
e.g. Enrolling participants who are either current smokers or non smokers and assessing whether or not they have decreased lung function
Longitudinal Studies
Employ continuous or repeated measures to follow particular individuals over prolonged periods of time - often years of decades
Generally observational in nature, with quantitative and/or qualitative data being collected on any combination of exposures and outcomes (without any external influences being applied)
Case control studies
Researchers identify study participants based on their case status - disease or no
Quantification of the number of individuals among the cases and the controls who are exposed allow for statistical associations between exposure and outcomes to be established
E.g. analysing the relationship between obesity and knee replacement surgery. Case = participants who have had knee surgery. Controls = those who have not. Comparison = odds of being obese if you have surgery compared to those that do not
Cohort studies
Identifying study participants based on their exposure status and following them through time to identify which participants develop the outcome of interest
E.g. cohort of 5766 men aged 35-64 recruited from workplaces followed for over 20 years to establish whether life circumstances during their childhood impacts health status
Survey studies
Survey respondents are asked a series of questions in a standard manner so that responses can be easily quantified and analysed statistically
Enables researcher to describe characteristics of sample being studied and make generalisations to the larger population of interest
Useful for collecting information about research phenomena that are not directly observable or measurable and for collecting data from people who are widely distributed geographically
Interventional studies
Defined as those where the researcher intervenes or manipulates the variables at some point throughout the study
Randomised control trials
A trial in which subjects are randomly assigned to one or two groups
One (the experimental group) receive intervention that is being tested and the other (comparison group or control) receive an alternative (conventional) treatment
Features of a well designed RCT
The sample studied is appropriate to the hypothesis being tested
Effective (concealed) randomisation of subjects (eliminate selection bias)
Both groups treated identically
Investigator assessing outcome is blinded to treatment allocation
Patients are analysed within the group to which they were allocated regardless of their experience of the intended intervention
Analysis focus on testing research questions the initially led to trial
Pre-post studies
Measures the occurrence of an outcome before and again after a particular intervention is implemented
Single arm = one group measured before and after
Multiple arms = where there is a comparison between groups
Non-randomised trials
Interventional study designs that compare in a group where an intervention was performed with a group where there was no intervention
Good choice for when RCT is ideal but practical considerations (costs, unacceptability to patients etc) make a high quality RCT infeasible
Visualising continuous data - main aspects of the distribution
The location or centre of the variability - atypical value taken by the variable
The speed of the variability - how far the values extend from the centre
Shape of the variability in whether or not values are spread symmetrically on either side of the centre
Histograms
Useful plot for exploring the distribution of a continuous variable when you have a lot of observations
Need to split the range of possible values into intervals, called bias
Count the number of observations falling in each bin
Two peaks - bimodal distribution
Symmetrical - similar tails on left and right
Skew to right - positively skewed - often when fixed lower limit
Density plots
Show a continuous estimate of density
Need a software to create whereas histogram can be calculated and drawn by hand
Useful for comparing distributions between groups due to being easily able to overly them
Median
Simple measure of the location of a distribution is to put all the values in order form smaller to largest and see what values if the middle
50th percentile
