L4. Clinical Trials in Respiratory Disease (Critical Appraisal) Flashcards
What are the levels of evidence and what do they represent?
- Systematic Reveiws of randomised controlled trials 2. Randomised controlled trials 3. Pseudo or non-randomised trials 4. Case series (observational) The RANDOMISATION is important to increasing the evidence and thus relevance to use of results in practice.
Define Internal Validity
Internal validity is the extent to which the results of a study are VALID in terms of accurateness and robustness - is it scientific? - did it answer the question? - was it appropriately designed, data collected and analysed?
What are the three factors that determine internal validity?
- Randomisation 2. Blinding 3. Intention-to-treat analysis
What is randomisation and what does it aim to achieve?
Randomisation is the random allocation of subjects into each group so that each group is identical in all aspects AIM: reduce confounders (less systematic differences) Note: if works based on having a large sample size
What is stratification? And what is the purpose?
This is a type of randomisation where there is randomisation within the subgroups of confounders. Example a confounder being age, that all people with this confounder are randomly allocated to different treatment groups.
What is the effect of randomisation?
It reduces selection bias
Define Binding and what does it aim to achieve?
This is where the subjects and the examiners (sometimes the analysis team) do NOT know what group they have been assigned to. AIM: to reduce information bias
What are some strategies of blinding?
Have a centralised process for multi-centre studies (endpoints): independent clinical endpoint committees. Have the outcomes determined by a strict, STANDARDISED, and objective criteria
What is Intention-to-treat analysis and what is the aim?
It is where the subjects of the trail are analysed according to the ORIGINAL group that they were assigned regardless of the treatment they ended up taking (switching to placebo due to side effects or dropping out). AIM: To reduce selection bias and to make a CONSERVATIVE estimate of the result which is better than making the treatment look better than it is.
What is the purpose of statistical analysis?
Because, aside from census data, the data collected is only a representation of the whole population and the statistical analysis extrapolates how well the data found reflects the TOTAL population.
What is the P value and what does it represent?
The P value (significance) is the probability that the result achieved was purely due to CHANCE and not due to the intervention. The cut off for statistical significance is usually a P<0.05. Eg. P=0.05 suggests that there is a 5% chance that there is no difference between the control group with the treatment group.
What is the confidence interval and what does it represent?
The confidence interval is the range in which there is a 95% confidence that the true population value lies within that range. If the NULL VALUE (which is 1 for ratio values or 0 for absolute values) is achieved then there is NO difference between the groups being compared. If the range was narrow: more precise result and wide: less precise. This is dependent on the sample size or POWER of the study (large n = small CI).
How do you examine the true effect of the study?
If the study showed an effect where in the population there was no effect then a type 1 error occured (false positive).
If the study showed no effect when in actual fact there is an effect in the population then this is a false negative or a beta error where the sample size was not large enough.
What is the Number to Treat analysis? What does it mean?
The number of people that is needed to undergo the treatment in order to see the effect in only 1 person.
It is a marker for the EFFICIENCY of the intervention
How do you calculate the number needed to treat?
NNT = 1 / absolute risk or rate of reduction
