What types of bias are there?
What is information bias?
Differences in measurement in the groups
Correct people wrong info
What is recall bias?
Differences in the groups due to forgetting info
What is selection bias?
Wrong people correct info
What is confounding?
A factor associated with the exposure and disease of interest but not itself on the causal pathway
What is Incidence rate?
Number of new cases of a disease in a defined population in a defined period of time
What is prevalence?
Proportion of people affected in a defined population at a particular point in time
What is a cohort study?
Identify based on exposure then monitor disease outcome
Suitable for rare exposures
Less prone to bias
Problems with follow up
Analyse with incidence rate ratio and smr for external comparison
Can be prospective or retrospective
What is a case control study?
Identify on disease outcome and compare exposure
Suitable for rare diseases
Problems with recall and selection bias
Analyse using odds ratio then estimate relative risk
What is a randomised control trial?
Comparison of two treatments interventions or a placebo chance determins who gets what and it is planned
Can be ethical problems
What is a systematic review?
Overview of primary studies that used explict and reproducible methods
Uses bradford hill criteria
What is a meta-analysis?
Mathematical synthesis of two or more primary studies that addressed the same hypothesis in the same way
What are the advantages of a systematic review?
Increase prescision of overall result
Large amounts of info can be understood quickly
Results of studies can be compared
Reasons for inconsistencies can be identified
Limitations of systematic reviews?
Dependant on quality of primary studies
Susceptible to publication bias
Reader still needs to consider generalisability of the result
What is the observed incidence rate ratio?
=observed incidence rate in pop 1/ observed incidence rate in pop 2
What is the odds ratio?
Measure of the risk of disease in one group compared to another
=exposed disease x unexposed no disease/unexposed disease x exposed no disease
What is the standard mortality ratio?
Smr >100 means more incidence than expected
What is the error factor?
Measure of random variation
Large error factor can be due to small sample size
= exp(2xsquare root[(1/d1) + (1/d2)])
What is the 95%confidence interval?
= from the obseved value / error factor to the observed value x by the error factor
What is the p-value?
A measure if how likely the results are due to chance
If 95% CI excludes the null hypothesis value of 1...
reject null hypothesis
There is evidence of a difference
Result is not due to chance
What is evidence base medicine?
Use of best current evidence in making descisions about the care of individual patients
Involves the integration of best clinical expertise best evidence from research and patients unique values and circumstances
What are the levels of evidence?
Systematic review of RCTs
Well designed trial
Well designed non experimental study