Research & Epidemiology Flashcards
Elements of foreground questions for specific knowledge to inform clinical decisions or actions:
PICO letters stand for:
P: Patient or population. The first step in the PICO process is to identify the patients or population to be studied. More specifically, it describes patients‟ characteristics, such as age, gender, disease status, or any other patientrelated characteristic.
I: Intervention to be tested. Identifying the intervention is the second step in the PICO process. It is important to identify the exposure intended to be studied in the research project. This may include the use of a specific diagnostic test, treatment, adjunctive therapy, medication, etc.
C: Comparison used in the research project. It is the alternative exposure to which the intervention will be compared, which might be the standard of care or a placebo. The comparison component is the only optional one in the PICO question, since the researcher might study the intervention alone because either due to no interest in comparison or the lack of a comparable group.
O: Outcome to be measured as a result of the intervention. It is the evaluation of the intervention‟s effect. This may include cure or level of control of a disease, efficacy of a medication or a diagnostic test, etc.
Odd ratio used in which study ?
Odd ratio for
case-control but if difference is small, odds ratio will be useful for all (cohort,RCT,case-control)
Sensitivity equation:
True Positives/(true positives + false negatives)
False Positive Rate:
1 − specificity
False Negative Rate:
1 − sensitivity
Specificity equation:
True Negatives/(true negatives + false positives)
Positive Predictive Value:
True Positive/(true positive + false positive)
Negative Predictive Value:
True Negative/(true negative + false negative)
What is the best design you choose to study the
prevalence of a disease?
Cross sectional study
What is the best trial design to study the incidence of
a disease?
Cohort study
Which of the following studies is considered a gold
standard for analytical epidemiology?
Cohort study
A critical appraisal of a RCT takes into consideration the followings:
A- Randomization
B- Blinding
C- Precision of the estimate (CI)
D- Benefice versus harm
In External validity to check the Applicability of results to your patients several Issues needed to consider before deciding to incorporate research evidence into clinical practice which are …
- Similarity of study population to your population
- Benefit versus harm
- Patients preferences
- Availability
- Costs
The best type of study to evaluate a therapy or screening ?
Clinical trials
The best type of study for diagnostic accuracy?
Clinical trials and cross sectionals
The best type of study for prognosis?
Clinical trials (placebo arm of RCT) and cohort study
The best type of studies to evaluate the harm ?
Clinical trials and cohort study and case-control.
The best type of studies to study the etiology ?
Cohort studies and case control.
Likelihood ratio (LR+):
Sensitivity/1-specificity
(If the result is = or > 10 that means that test post test probability is higher and most likely they had the disease)
more likely a test result is to be found in the “diseased.”
Likelihood ratio (LR-):
1-Sensitivity/specificity
(if likelihood less than 0.2 that means the test role out the disease with high post-test propability)
less likely a test result is to be found in the “diseased.”
Each Outcome has five elements:
- Domain
- Specific measurement(s)
- Specific metric(s)
- Method(s) of aggregation
- Time point(s)
Pragmatic trials are
designed to evaluate the effectiveness of interventions in real-life routine practice conditions, whereas explanatory trials aim to test whether an intervention works under optimal situations. Pragmatic trials produce results that can be generalized and applied in routine practice settings.
Verification bias seen in
Cross sectional
verification bias is
type of measurement bias in which the results of a diagnostic test affect whether the gold standard procedure is used to verify the test result.
(sometimes referred to as “work-up bias”) occurs during investigations of diagnostic test accuracy when there is a difference in testing strategy between groups of individuals, leading to differing ways of verifying the disease of interest.
Publication bias
the failure to publish the results of a study “on the basis of the direction or strength of the study findings
To calculate NNT
NNT (number needed to treat)=1/ ARR
To calculate Absolute risk reduction (ARR) – also called risk difference (RD) –
= ARC - ART
ARC= the AR of events in the control group
ART= the AR of events in the treatment group
RR (relative risk)=
ART / ARC
RRR (relative risk reduction)=
(ARC – ART) / ARC
RRR= 1 – RR
AR (absolute risk)=
the number of events (good or bad) in treated or control groups, divided by the number of people in that group.
If the RR (the relative risk) or the OR (the odds ratio) = 1, or the CI (the confidence interval) cross the line of unity, what does that mean ?
then there is no significant difference between treatment and control groups.
If the RR >1, and the CI does not include 1, events are significantly more likely in the treatment than the control group.
If the RR <1, and the CI does not include 1, events are significantly less likely in the treatment than the control group.
How to check for publication bias
funnel plot can be used to examine whether ESs from smaller studies show more variability than those from larger studies.
(Can be checked only by logistic regression and eager ??)
If there is no bias, 95% of studies should lie within the triangle a. A larger sample size increases precision so small studies will be scattered widely at the base of the triangle whereas larger studies (more powerful) will be at the top and have a narrow spread.
Funnel plots should be symmetric in the absence of study heterogeneity and publication bias
Attritional bias can be avoided by
Intention to treat analysis
Referral bias or tertiary hospital bias
type of selection bias. People who are referred into studies are frequently different from those who are not, meaning that the results of a trial may not generalize well to the general population.
How to solve the drop-out (loss of FU) and demographic variations in cohort study and case-control?
- Sub-group analysis (stratified analysis)
- multiple regression analysis
The only problem about this it is going to decrease the sample size
Unacceptable Loss of follow up percentage depends on
Observed event rate
Median survival curve (half-life)
How long it takes to 55% of population survive
Applicability (3 Ps)
Paper of best available evidence
Physician experience
Patient’s value and homogenous group
The World Health Organisation define the maternal mortality ratio as
Maternal deaths per 100,000 livebirths
if I square (inconsistency) less than 50 % that means (or p value is Insignificant)
Heterogeneity is Insignificant
P value of effectiveness with drug
Type 1 error
One of the disadvantages of the mean is
its sensitivity to extreme values.
The p-value is
a probability that the observed difference is due to chance or due to sampling error.
The α level (also referred to as type I error) is
it is the error taking place when making the inference, and is the probability of rejecting a true null hypothesis.
the β (also referred to as type II error) is
the error when making the inference, and is the probability of accepting a false null hypothesis.
There are 5 ways to control for confounding:
In design stage:
▪︎Randomization (in clinical trial)
▪︎Restriction (including subjects who are homogeneous in terms of the confounder)
▪︎Matching (in case control studies)
In analysis stage:
▪︎Stratification (assessing the association separately for the levels of the confounder)
▪︎Multivariate analyses (statistical technique that removes the effect of the confounder)
The most effective way to control confounding is the multivariate analyses which necessitates collecting relevant information on the potential confounder so that it can be adjusted in the analyses.
Reporting bias or Observer Bias (Ascertainment Bias, Detection Bias, Assessment Bias) can be avoided by
Triple blinding
Tertiary (referral or centripetal) bias can be avoided by
Multi- centres studies
Odd ratio in case control equation
ad/bc
Relative risk calculation
=a /a+b ÷ c/c+d
Performance bias can be avoided by
- Standardisation of care protocol
- blinding of care providers and patients.
Detection bias can be avoided by
Double blinded studies
When you Compare 3 normally Distributed group which test Preferred to use
ANOVA test (Analysis of variance)
if you Comparing Pre-/Post treatment result which test Prefer to us
T-test for paired sample
Chi-sequare used for
to test relationship btw two categorical variable
Absolute risk reduction
= control event rate - exposure event rate
= C/c+d - a/a +b
Attributable risk (or excess risk/ ARI) calculation
= a/a +b - c/c+d
independent (non paired or student) t-test used for
Continuous with categorical variables
The probability of rejecting a false null hypothesis (B) =
1-B
One sample t-test used for
It is a test used for assessing the difference between a continuous variable and a fixed reference value, such as the difference between the observed SBP and the normal value (of 120 as an example).
Paired t-test:
It is a test used to assess the association between two continuous measurements, which are related to each other, such as heart rate before and after taking a certain medication.
Analysis of variance (ANOVA):
It is a test to assess the difference between a categorical variable with more than two levels and a continuous one, such as the difference in age between patients of different severity of illness (mild, moderate, and severe).
Correlation:
It is a test used to assess the association between two continuous variables, such as SBP and age.
Correlation:
It is a test used to assess the association between two continuous variables, such as SBP and age.
The value of the power desired (1-?), and it is usually considered at ?
The value of the power desired (1- beta), and it is usually considered at 0.8 (80%).
Risk difference is
also called attributable risk and it measures the difference in the risks between the exposure groups.
The way to avoid selection bias is by
employing strict inclusion/exclusion criteria, as well as applying a random selection from the target population.
Selection bias is
a systematic error that occurs at the stage where improper procedures are followed when recruiting subjects into the study.
Information bias is
another systematic error that occurs at the stage of data collection, where the information collected are inaccurate, and this could be at the level of the exposure, the outcome, or other factors. This could take place when imprecise measurement instruments or invalidated questionnaires are used. Another source of information bias is introduced by subjects, such as recall bias, or reluctance to tell the truth. A third source of information bias is introduced by the interviewer when he/she probes for specific answers.
information bias is avoided by
using validated questionnaires as well as following a structured method to collect information.
The most effective way to control confounding is
the multivariate analyses which necessitates collecting relevant information on the potential confounder so that it can be adjusted in the analyses.
Maternal Mortality Rate :
maternal death/100,000 women of reproductive age (women)
Perinatal Mortality Rate :
IUFD+1 st week death/1000 Total births
It is usually expressed at the rate per thousand total births over one year
Neonatal Mortality Rate :
NND (28day) / 1000 Live births
Nuremberg Code of experiment:
o Has voluntary, well-informed consent
o experiment aim at positive results
o It should be based on previous knowledge that justifies the experiment.
o Avoids unnecessary harm. o Shouldn’t be done when there is risk of death or disabling injury.
o The risks should not exceed the expected benefits.
o Staff must be fully trained & qualified.
o Subjects are free to quit the experiment at any point
o Stop the experiment at any point when continuation dangerous.
Baseline cohort (adding baseline characteristics) is like
Cross sectional study
Minimal reviewers (to abstract study)/ studies to do systemic review …
2 reviewers
2 studies
Steps of an audit Cycle
- Initial needs assessment
- Identification of standards
- Data collection
- Recommendations
- Re-audit
Standard error of the mean (SEM) =
standard deviation divided by the square root of n.
In ‘negatively skewed’ distribution locate mean, median and mode
‘negatively skewed’ where the distribution of data about the mean tails off to the left with the majority of points being greater (the median and the mode are greater than the mean).
Tests used to compare two means
