Evidence Based Research Exam 1
What is an example of a double blind RCT?
All COVID vaccine studies
In a cohort study, what must be true of participants at the time of exposure?
All participants must be free of disease/condition being studied at the time exposure defined
Is a case-control prospective or retrospective?
Always retrospective
What is an experimental study?
An investigation where the researcher directly intervenes to provide an intervention or treatment to one group and not to another group (control) with the analysis focusing on isolating and estimating the effects of the treatment
What is an observational study?
An observational study is an investigation in which neither the subjects nor any variables are manipulated in any way
Is an observational study an investigation or an experiment?
An observational study is an investigation, NOT an experiment
Can a meta-analysis stand on it's own or is it part of a systematic review?
Both - not all systematic reviews include meta-analysis but all meta-analyses are found in systematic reviews
Does a cross-sectional study use probability or non-probability sampling?
Both probability and non-probability sampling are used in cross-sectional studies
Types of observational studies
Case Reports Cohort Case-control Cross-sectional
Case control study
Case group (known to have the outcome) and a control group (known to be free of the outcome) The goal is to select individuals in whom the distribution of exposure status would be the same as that of the cases
What are advantages of meta analysis?
- Greater statistical power - Greater ability to extrapolate to general population effected - Considered a very high level of evidence
Cohort study limitations
- If prospective, time consuming and expensive - If retrospective, availability of records and changes in monitoring/coding of outcomes means not all risk factors for the group under study are available - If prospective, can be inefficient for rare diseases (due to small sample sizes)
What are advantages to cross-sectional studies?
- Provide important info for health care planning and generation of hypotheses - No loss to follow up (participants only contacted 1x) - Can be conducted in a short time - Good generalizability - Less costly than cohort/case-control
What are advantages of case-control studies?
- Quick and inexpensive - Particularly appropriate for investigating outbreaks and rare diseases - Good for studying multiple exposures - Good for preliminary investigation of a suspected risk factor for a common condition
What are advantages of systematic reviews?
- Summarizes findings from multiple studies and makes info easier to read and understand - Reduces selection bias and more reliable/accurate results
Define Internal Validity
Internal validity is the extent to which the results (outcomes) of the research reflect the associations inherent in the variables used in the research and do NOT reflect associations with any other, extraneous or alternative variable or set of variables
What is the typical format of a peer-reviewed medical journal?
Introduction Methods Results Discussion (Conclusion)
What is the essential property of conditional probability?
It "shrinks" the sample space by giving you additional information
What are advantages of Primary Data collection method?
Less measurement error, typically suits objectives of the study better
Conclusions produced by __________ are statistically stronger than the analysis of any single study due to increased number of subjects, greater diversity among subjects, or accumulated effects and results
Meta-analysis
In a peer reviewed medical journal, in what section is the study design presented and discussed?
Methods section
Are sensitivity and specificity affected by prevalence?
Nah
How does a systematic review differ from traditional narrative reviews?
Narrative reviews are more descriptive and do not involve a systematic search of the literature Narrative reviews can often include selection bias
What is incidence?
Number of new cases of disease over some specified time interval
Would the number of patients in a study be considered "variables" or "observations"?
Observations = # of patients or data points in a study
Conditional probability example: - I roll a fair die - Outcome A is an odd number - Event B is the outcome is less than or equal to 3 What is P(A|B)? (conditional)
P(A) = (1,3,5) = 3/6 = 1/2 (odd #) P(B) = (1,2,3) = 3/6 = 1/2 (less than or equal to 3) P(A∩B) = (1,3) = 2/3 =1/3 (A and B occurs) P(A|B)=P(A∩B)/P(B) P(A|B)= (1/3)/(1/2) = 2/3
What is the formula for conditional probability?
P(A|B)=P(A∩B)/P(B) AKA the probability that A or B occurs divided by the probability that B occurs
Unconditional probability: What is the probability of two dice totaling 5? Conditional probability: What is the probability of two dice totaling 5 if neither die has a 1 or a 6 on it?
Unconditional probability: You could roll (1,4), (2,3), (3,2), or (4,1) = 4/36 = 1/9 Conditional probability: You could roll (2,3) or (3,2) = 2/16 = 1/8 *Note the sample space is smaller b/c neither die has a 1 or 6 on it.
Meta-Analysis
Uses statistical techniques to summarize the results of other published peer-reviewed studies The aim is to provide reliable evidence
What are advantages of Secondary Data collection method?
Can be cheap or freely available
Equation for probability of a false negative
(1-sensitivity)
How are subjects assigned in an observational study?
The assignment of subjects into a treatment or control group is outside the control of the investigator
How are participants selected for a case control study?
The source of control is dependent on the source of cases To minimize bias, controls should be selected to be representative of the population that produced the cases (ex from the same hospital admissions as those with the outcome) controls are "matched" to cases over attributes such as age, gender, and race
Equation for probability of a false positive
(1-specificity)
How are subjects assigned in an experimental study?
(Like an RCT) each subject is randomly assigned to a treatment group or a control group before the start of the intervention
What are disadvantages to cross-sectional studies?
- Can measure prevalence but NOT incidence - Susceptible to bias due to low response - Only one data point (can NOT answer questions related to time)
What are disadvantages to systematic reviews?
- Can take months - Difficult to combine findings of different studies - Authors must decide on quality of sources
Cohort study strengths
- Clarity of temporal sequence between exposure and disease more clearly established - Allows calculation of incidence - Allows examination of multiple effects of a single exposure - If retrospective, can be inexpensive to obtain data
What are strengths of RCTs?
- Control of exposure - Randomization of subjects reduces risk of bias - Can provide direct evidence for causal effects
What are disadvantages of meta analysis?
- Difficult and time consuming - Not all studies provide adequate data - Requires advanced statistical techniques - Heterogeneity of study populations
What are limitations of case-control studies?
- Difficult to find sufficient controls - Prone to selection and recall bias
What are weaknesses of RCTs?
- Expensive, difficult to design, labor intensive, loss to follow up, poor compliance - Funding source and commercial interests of investigators can be sus - Highly controlled conditions may not reflect real world
What are the three properties of P(X)?
1. For any event X, P(X) ≥ 0 2. The probability of the sample space S is P(S) = 1 3. If X1, X2, X3.... are disjoint events, then = P(X1) + P(X2) + P(X3) etc
What is the probability of rolling a 1, 3, or 5? ex. P(1,3,5)
1/6 + 1/6 + 1/6 = 1/2
What is the probability of rolling a 1 or a 5 on a 6 sided die?
1/6 + 1/6 = 1/3
(D+ ∩ T-) is a.....
False Negative
(D- ∩ T+) is a......
False Positive
Cross-Sectional Studies
A "snapshot" - all measurements are taken at a single period point in time
What is research study design?
A framework, or the set of methods and procedures used to collect and analyze data on variables specified in a particular research problem
Define an outcome
A singular (atomistic) result of the random experiment
What is a prospective study?
A study in which 2 groups of subjects are selected/designed (typically one group consists of people with the risk factor and one without) and followed into the future (prospectively)
What is a retrospective study?
A study that collects data about past events *Note: in a retrospective study, the outcome of interest has already occurred
What is an event?
A subset of the sample space, which is a set of outcomes - which may be one of the outcomes or a collection of many outcomes
How does a meta-analysis compare to a systematic review?
A systematic review refers to the entire process of selecting, evaluating, and synthesizing all available evidence Meta-analysis refers to the statistical approach of combining the data derived from a systematic-review
Joint Probability/Marginal Probability =
Conditional Probability
If a random variable takes on uncountable many values it is _____ ex. Weight, temp, or blood pressure
Continuous
What are disadvantages of Primary Data collection method?
Costly, may not be feasible
Secondary Data collection method:
Data is collected by others, typically for purposes not related to those of the current study. Sources include: Administrative data (employment/hospital records, medicare), group records (census, insurance data)
Lung cancer/smoking/alcohol consumption/occupation Which of these is the dependent variable (outcome) and which is the independent variable (exposure)?
Dependent (outcome): Lung cancer Independent (exposure): Smoking, alc consumption, occupation
Case Report
Describes experience of a single pateint or a group of patients with similar diagnosis - typically something unusual
If a random variable X is finite or countable, it is called _____ ex. # of patients in a waiting room or # of operations done today
Discrete
What are the (2) classes of random variables?
Discrete and Continuous
What is the difference between efficacy and effectiveness?
Efficacy is the performance of an intervention under ideal and controlled circumstances Effectiveness is it's performance under "real world" conditions
Probabilities are typically assigned to _____
Events
What type of variable would age, gender, weight, or rurality be considered?
Exposure Variables
Define External Validity
External validity is the ability to generalize the results of the population under study
Can an experimental study establish causal effects?
Hell yeah brother
What is useful about a case report?
Identifies new and unusual conditions and good for formulating hypotheses
Double blind RCT
RCT where neither the patient nor physician knows who is getting treatment and who is getting control
As prevalence increases, what happens to PPV?
PPV increases with prevalence
Cohort Studies
Participants are defined on the basis of presence or absence of exposure
What are disadvantages of Secondary Data collection method?
Potential for large measurement error, not compiled for any particular study's objectives
What is a systematic review?
Provides an exhaustive summary of scholarly literature related to a particular research topic or question
What is useful about a cross-sectional study?
Provides great information about prevalence of disease/behavior in a population
What is specificity?
Specificity is the proportion of individuals without the disease that have a negative test result The proportion of true negatives that are correctly identified by the test The conditional probability of a negative test result given the patient does not have the disease P(T-|D-)
Negative Predictive Value equation
TN/(FN+TN) d/(c+d) The probability that a negative test is a true negative
Specificity equation
TN/(FP+TN) d/(b+d) The probability of a negative test among patients without disease
Sensitivity equation
TP/(TP+FN) a/(a+c) The probability of a positive test among patients with disease
Positive Predictive Value equation
TP/(TP+FP) a/(a+b) Probability that a positive test is a true positive
The Dependent variable is the _______ and the Independent variables/predictors are the ________
The Dependent variable is the outcome variable and the Independent variables/predictors are the exposure variables
What is Negative Predictive Value
The conditional probability a patient does not have the disease given a negative test result P(D-|T-) AKA the proportion of patients with negative test results who are correctly diagnosed
What is Positive Predictive Value?
The conditional probability that a patient has the disease given a positive test result P(D+|T+) AKA the proportion of patients with positive test results who are correctly diagnosed
Cross-sectional studies are done at the level of...
The individual (often with surveys)
Primary Data collection method:
The investigator is the first to collect the data. Sources include: Medical examinations, interviews, randomized control trials
What determines the type of study design used?
The nature of question The goal of research The availability of resources and/or data
What is the value of the variable of interest also called?
The outcome variable, or "outcome"
What is a false positive?
The percent of people who do not have the disease but test positive The conditional probability that a patient tests positive given they do not have the disease P(T+|D-) AKA healthy people wrongly identified as sick
What is a false negative?
The percent of people who have the disease but test negative The conditional probability that a person will test negative given that they do have the disease P(T-|D+)
P(X) reads as....
The probability of X
What is prevalence?
The proportion of all persons who have a condition at or during a particular time period (a+c)/(a+b+c+d)
What is sensitivity?
The proportion of individuals with the disease that have a positive test result The proportion of true positives that are correctly identified by the test The conditional probability of a positive test result given the patient has the disease P(T+|D+)
P(A∩B) = P(A)P(B) .... What does this mean?
The random variable A and B are Independent
What is the sample space (S)?
The set of all possible outcomes
Are PPV and NPV affected by prevalence?
They sure are!
Randomized Controlled Trial (RCT)
Two groups (one treatment and one control) where patients are randomly assigned to all groups
Virtually all medical research explains the relationship between...
Two or more variables
What is the hypothesis?
Whether one or more variables are associated with change in the value of the variable of interest
Should unpublished research be used in a systematic review?
Yes - using only published threatens validity as this may lead to a misrepresentation as it does not include findings from ALL existing research