Research Flashcards
NICE 2013 Levels of Evidence
1++ High-quality meta-analyses, systematic reviews of RCTs or RCTs with a very low risk of bias
1+ Well-conducted meta-analyses, systematic reviews of RCTs or RCTs with a low risk of bias
1− Meta-analyses, systematic reviews of RCTs or RCTs with a high risk of bias
2++ High-quality systematic reviews of case–control or cohort studies; high-quality case–control or cohort studies with a very low risk of confounding, bias or chance and a high probability that the relationship is causal
2+ Well-conducted case–control or cohort studies with a low risk of confounding, bias or chance and a moderate probability that the relationship is causal
2− Case–control or cohort studies with a high risk of confounding, bias or chance and a significant risk that the relationship is not causal
3 Non-analytical studies (for example case reports, case series)
4 Expert opinion, formal consensus
NICE 2013 Grading of Recommendations
A. At least one meta-analysis, systematic review or randomised controlled trial (RCT) that is rated as 1++, and is directly applicable to the target population; or a systematic review of RCTs or a body of evidence that consists principally of studies rated as 1+, is directly applicable to the target population and demonstrates overall consistency of results; or evidence drawn from a NICE technology appraisal
B. A body of evidence that includes studies rated as 2++, is directly applicable to the target population and demonstrates overall consistency of results; or extrapolated evidence from studies rated as 1++ or 1+
C. A body of evidence that includes studies rated as 2+, is directly applicable to the target population and demonstrates overall consistency of results; or extrapolated evidence from studies rated as 2++
D. Evidence level 3 or 4; or extrapolated evidence from studies rated as 2+; or formal consensus
D (GPP). A GPP is a recommendation for best practice based on the experience of the guideline development group
FSRH GRADING OF RECOMMENDATIONS
A. Evidence based on randomised controlled trials
B. Evidence based on other robust experimental or observational studies
C. Evidence is limited but the advice relies on expert opinion and has the
endorsement of respected authorities
Tick: Good Practice Point where no evidence exists but where best practice is based
on the clinical experience of the guideline group
Sensitivity
TRUE POSITIVES
Correct identification of pts WITH the disease/ % chance that the test will ‘pick up’ someone with the disease
true positives / (true positives + false negatives)
High SnNout: A highly sensitive test with a negative result = reliably rules out the disorder
“If the test is negative we can be really sure that you don’t have the condition because the test picks up 99.9% of people with it”
Good sensitivity = good (high) negative predictive value.
Negative predictive value (NPV)
POST-TEST PROBABILITY that those with a negative test result are actually disease free
true negatives / (true negatives + false negatives)
Test of probability - As population prevalence increases, NPV decreases
Sensitivity and Specificity are constant properties of the test itself
Specificity
A measure of true negatives (correct identification of patients WITHOUT disease)
% chance that the test will correctly identify someone without the disease
(true negatives) / true negatives + false positives
If it’s high that means the tests value is in RULING PEOPLE IN to having the condition
SpPin: A highly specific test with a positive result = reliably rules in the disorder
“If the test is positive we can be really sure that you do have the condition because the test identifies 99.9% of people without the disease correctly”
Therefore, tests with good specificity also have a good (high) positive predictive value
Positive predictive value (PPV)
PPV calculated = true positive / true positives + false positives (all positive results)
POST TEST PROBABILITY that those with a positive test result are actually diseased
As population prevalence increases, PPV increases (and vice versa)
Relative risk and how it is derived
RR = the risk of the patient compared to a control group
RR >1 = you have higher risk than the control
RR <1 = you have a lower risk than the control
Calculated = experimental event rate (EER) / control event rate (CER)
Relative risk reduction tells us how much an intervention decreases the risk of an outcome, compared to those who did not have the intervention
Relative risk reduction = 1 – RR
Absolute risk and how it is derived
Absolute risk is the probability that a person will experience an event
Expressed as a percentage
AR = number of x event occurred in group / number of people in that group
Absolute risk reduction is the definitive amount that a patient’s risk is decreased with a given intervention (compared to their own pre-intervention risk)
Calculated = AR (control group) – AR (treatment group)
Number Needed to Treat
number of patients that need to take the new drug before we have one positive outcome/ prevent one event
Calculated = 1 / ARR
A lower NNT = better, more effective drug OR
The drug is being used in the best population, less unnecessary people being treated
Number Needed to Harm
how many people need to take a new drug before there is one adverse outcome
Calculated = 1 / absolute risk increase
A lower NNH = a poorly tolerated or more risky treatment, treatment targeted for wrong groups
P values and 95% confidence intervals
P value <0.05 = “tells us that the results are highly unlikely to have occurred by chance”
Confidence interval of 95% = a measure of how sure we are that the true value lies within a certain range
A smaller range of CI = more precise and certain results
CI not including 1/0 = that at no point in the range of possible results would there be no significant difference between the two groups
Together the p value and CI mean that if someone were to repeat the experiment, they should also get the same/ very similar results
