EXAM II Material Flashcards
What is a null hypothesis?
A research perspective which states there will be NO (true) difference b/w the groups being compared
Stat Perspectives:
- Superiority 2. Inferiority 3. Equivalency
Alternative Hypothesis
A research perspective which states that there will be a true difference b/w the groups being compared
What are the 2 factors in which study populations are based upon?
- Ethics
- Equipoise
Which type of study design involves no researcher-forced group allocation?
Observational
considered “natural/freely”
What type of study involves researcher-forced group allocation?
Interventional
considered “experimental”
Investigator-selectes interventions (exposure)
What are study population selection based upon?
- The research hypothesis/question
- Inclusion and Exclusion selection criteria (interventional) and Case & Control group or Exposed & Nonexposed group selection criteria (observational)
- Ethics (principles of bioethics must be met)
- Equipoise
Define Equipoise
Genuine confidence that an intervention may be worthwhile (risk vs benefit) in order to use it in humans (greater benefit than risk)
Must have genuine reservations about the treatment in order to withold information
What are the 4 key principles of bioethics?
- Autonomy - self-rule/self-determination
- Beneficence - to benefit the patient, not society
- Justice - equal & fair treatment
- Nonmaleficence - do no harm
What is the Belmont Report and what is it used for?
A document used to know whether a research is ethical
Issued by National Commission for Protection of Human Subjects of Biomedical and Behavioral Research
What are the 3 guiding principles for the ethical conduct of research methodology?
- Respect for persons (voluntary)
- Beneficence (research risks are justified by potential benefits)
- Justice (risk and benefits of research are equally distributed)
Define Assent
Agreement to participate, based on being fully and completely informed, given by mentally-capable individuals not able to give legal consent (i.e. children and adolescents)
Not mentally capable
What is the role of the IRB (Institutional Review Board)?
aka “Ethics committee”
Determines whether a study is ethical (proper/safe)
This MUST occur Before a study begins
To protect human subjects
All observational and interventional studies
What is the role of the Data Safety & Monitoring Board (DSMB)?
Protects the safety of the patients After a study starts
What is the key difference between Observational and Interventional study designs?
In Interventional study designs, the researcher allocates and forcefully allocates study subjects into forced-intervention groups
Which type of study design is most likely to prove causation?
Interventional
What are the advantages of Interventional studies? (2)
Most likely to prove causation
Only study used to become approved by FDA
What are the disadvantages of Interventional Studies? (4)
Costly
Time consuming/complex
Ethical considerations
Generalizability/External Validity - too restrictive
What is an explanatory interventional study?
They’re done when you’re trying to explain the impact of the intervention and prove causation
Super restrictive w/ inclusion criteria
What is a pragmatic interventional study?
When they mirror the real world w/ a clinical environment in which:
NO PLACEBO USED
“regular” people from community join (can have multiple diseases)
Allows physicians to use their own judgement
List the disadvantages of pragmatic interventional studies
Lost the advantages given by explanatory studies
Loss of control and rigidity
More confounding can occur
What are the elements of interventional clinical trial in which it may prove causation?
Has the strongest evidence = causation
Must go through FDA approval
Has randomization, exposure and intervention
How many randomizations occur in a simple study and when are they used?
Randomization occurs once, used usually to test a single hypothesis or begin with the effect of one drug
How many times does randomization occur in a factorial study and when it the study used?
Once initially, then later on.
Usually used to test multiple hypothesis at the same time
What’s an advantage of factorial design study?
Improves the efficiency for answering clinical Qs
What are the disadvantages for factorial study designs? (4)
Requires larger sample size
Higher risk for people dropping out
Increased complexity
May restrict generalizability of results
What is unique about a parallel study design?
No switching/crossover occurs after the initial randomization (can be a simple or factorial study)
What is unique about a cross-over study design?
Groups can crossover from one intervention to another during the study
This allows for a smaller sample size
Between and within group comparisons are possible
What are the disadvantages of crossover study designs?
Only suitable for long-term conditions which are not curable or which treatment provides short-term reliet
Longer duration
Carry-over effects during cross-over (wash-out required; which prolongs study)
Complexity in data analysis
Treatment-by-Period interaction - difference in effects of treatments during diff time periods
What is the selection criteria for Observational studies?
Exposed vs. Non Exposed
Case vs. Control
What is the selection criteria for Interventional Studies?
Inclusion vs. Exclusion
What selection criteria impacts generalizability of interventional studies?
External Validity
What characteristics of a study must go through a full board review by the IRB?
All Interventional Studies
Those that have more than minimal/low risk to patients; all medically related studies
What characteristics of a study must go through an exempt board review by the IRB?
Those that have low to no risk to the patients
i.e. use of existing data and/or specimens
No patient identifiers
What characteristics of a study must go through an expedited board review by the IRB?
Those that have minimal risk and/or no patient identifiers
What are the 3 Guiding Principles of the Belmont Report?
Respect for persons - research should be voluntary
Beneficence - risks of the research are justified by potential benefits
Justice - risk and benefits are equally distributed
What are the types of outcomes that can occur in an intervention trial that are used in the conductance of a study?
Primary - initial hypothesis is used for the conduction of the study
Secondary/Tertiary/etc. - generation of future hypothesis; less important than primary; no expectations
Composite - combines multiple endpoints into a single outcomes
Patient-Oriented - most clinically relevant & important to the patient & family (death, hospital stay, quality of life)
Surrogate Markers - direct endpoints, i.e. hypertension for risk of stroke, high cholesterol for risk of heart attack, etc.
What are the 2 forms of group allocation in conductance of an interventional clinical trial?
Random and NonRandom
What occurs during nonrandom group allocation in interventional studies?
There is not an equal chance of subjects being place into each of the groups.
There is no equal probability of subjects being selected or assigned to each group
What occurs during random group allocation in interventional studies?
There is an equal chance of subjects being placed into each group; may use a program that utilizes random numbers
What are the two purposes of randomization? Which one is the primary purpose?
To eliminate bias <– primary
To try and make groups as equal as possible
Define randomization
The attempt to reduce systematic differences/bias between groups which can have an impact on the results/findings
What are the 3 types of randomization?
Simple
Blocked
Stratified
What does simple randomization ensure?
That you have an equal probability during study group allocation within one of the study groups
What does blocked randomization ensure?
That there is an equal number within each intervention group
Researcher does not know (if so = bias); usually an outsider is in control of the randomization process
What does stratified randomization ensure?
That there is a balance within the known confounding variables (characteristics)
i.e. want equalness in factors such as age, sex, etc.
Define post-hoc survey
Surveys done at the end of the study to find out if subject is able to determine which interventional group they belonged to.
Want a low predictability
Define an open-label study
An interventional study in which both researcher and subject know which intervention the subject is recieving
Usually investigator just wants to study the effects of something, and not necessarily other factors that are related to its effect
Define placebo/dummy treatment
Inert treatment that’s made to look identical to the active treatment
i.e. dosage frequency, dosage amount, etc.
What is a double-dummy treatment?
More than one placebo is used
Requires equal placebo treatment
What is the placebo effect?
When a condition improves while subject is given placebo, usually due to power of suggestion or because care/attention is being given
What is a run-in/lead-in phase?
A phase before a study begins where subjects are given a placebo to see if they are able to comply and to determine a new base-line of disease by “washing out” existing medication
When does wash-out occur during a study?
Usually during the middle
When is it acceptable to use post-hoc analysis?
When it is planned to be implemented before the study even begins
After the study you try to find a difference in the outcomes that may have occurred
What is an add-in number
The anticipated amount of drop-outs or lost to follow-ups
What are the 2 ways in which drop-outs/lost to follow-ups are dealth with?
Keep their data (intent to treat); either use the last data before they left, of convert all subsequent info to a null-effect/no benefit
Disregard the data and remove from study
What are the positive impacts of intent-to-treat studies? (3)
Preserves the randomization process
Preseves baseline characteristics and group balance
Maintains statistical power
Define per-protocol/efficacy analysis
Pre-define a compliance/a level of completeness that subjects must adhere to
Usually 80-90% completion
Otherwise, data not included
How does a “mixing of effects” come about in an interventional study?
When the investigator allows subjects to switch groups and where ever they end up is where they are evaluated
“As Treated”
A way to handle drop-outs of lost to follow-ups
What are the ways that an investigator can assess the compliance/adherence of a subject?
Bottle counter-tops
Measuring drug levels
Pill counts at each visit
What are methods done to improve adherence/compliance?
More frequent follow-ups and communication
Treatment alarms/notifications
Dosage containers
What is one aspect in which observation and interventional studies have in common?
Both randomly select their groups and can use drugs in the study
Case-control studies are dependent upon the person’s
Disease status
Common uses for a case-control study
Testing for multiple exposures compared to 1 outcome
Disease is rare
If there is an outbreak
Time/cost effective
To determine odds and OR
Dynamic populations
If disease has long induction/latent period
When there are ethical concerns in interventional study
What type of data is useful when selecting cases?
Accurate, medically reliable and efficient
Best when clinically supportive and definable by using published professionally recognized and accepted diagnostic criteria or from multiple sources of data
What do you want to avoid during selection of cases for a study?
Misclassification error
A control group gives investigator information in regards to the:
Baseline risk factor of where the diease may have started from/where the cases are drawn from
Selection Bias
Bias in which the proper randomization of groups is not achieved and the groups are not a representation of the normal/standard population
Leads to a non-valid association of two variables
Most common bias
How can selection bias negatively impact a case-control study?
The odds ratio can end up being the same among the categories (such as age groups); null
Cohort studies are based on _____ while case-control studies are based on _____
Outcome
Exposures
Nested Case-Control Study; 3 ways to select controls
A case-control study that is done prospective to a cohort study where the those that end up developing the disease/outcome in the cohort study end up being the “case”
Used to evaluate other exposures
Can located by going back to baseline, looking at risk set or survivor sampling
3 ways in which control sampling can occur during a nested case-control study
Survivor Sampling
Base Sampling
Risk-set Sampling
Survivor Sampling
When an investigator uses those who did not obtain the disease/outcome during a cohort study, used for the nest case-control study as controls
Base Sampling
During a nested case-control study when the investigator finds the controls from persons who were non-diseased from the start of a cohort study
Risk-Set Sampling
During a nested case-control study when investigators obtain their controls from non-diseased persons during study period at the same time the case was diagnosed
Define Case-Crossover Study
When a subject serves as both a case and a control in a case-control study
Selection bias is not an issue
Addresses the issue of “temporality” (whether cause precedes effects)
Can look at diff amounts of risk levels thruout study
What are the 2 types of matching when conducting a case-control study
Individual Matching
Group Matching
Individual Matching in Case-Control studies
When you select cases based on unique/individual characteristics
Group Matching in a Case-Control Study
Performed when you want the same characteristics within case and control groups
Must select cases first
Do not want to select based on a risk factor otherwise won’t be able to differentiate
i.e. 30% beautiful black women cases, so want 30% beautiful black women controls
What information do you know and must determine in a 2x2 table when doing case-control studies?
You know the outcome, but not the exposure/risk factor, therefore you know the total number (A+C), but you do not know the value of A or C independently
What is the selection criteria of subjects based upon in observational study designs?
- Case and control
- Exposed and non-exposed
What is the selection criteria of subjects in interventional study designs?
Must have Inclusion and Exclusion
Inclusion-characteristics to be in the study
Exclusion-characteristics where they cannot be in the study
Benefits of having exclusion and inclusion criteria
Increases the likelihood of producing reliable and reproducable results and decreases the likelihood of causing harm to the patients
List all observational studies in order of increasing evidence
Cases - reports & series
Ecological
Cross-sectional
Case-Control
Cohort
Which studies of observational are analytical?
Cohort
Case-Control
Cross-Sectional
Define external validity
How well the data of the study can be externalized to the general population, beyond the study population = generalizability
Internal Validity
Whether or not you can make inferences based on the findings within the study; whether a conclusion is valid or not
Coffee drinking and lung cancer example; coffee drinkers are more likely to smoke so coffee drinking is not the cause
Descriptive Epidemiology/studies
An attempt to understand a population’s health status by using descriptive elements; age, sex, race, etc.
Surrogate Markers/End points
Usually physiological effects that can cause disease
i.e. cholesterol levels for heart disease, blood pressure for hypertension
Analytical studies
Attempt to analyze two comparisons groups
When is it necessary to use a sample as oppossed to population in a study?
When studying a complete population is not feasible
What is one of the most common bias?
Selection bias
Who enforces the rules and regulations of the IRB?
Office of Human Research Protections
Examples of patient-oriented outcomes
Things that the patient cares about; best methodological study
Morbidity, symptom improvement, cost reduction, quality of life
“Run-In”/”Lead-in” Phase
Occurs before a study begins when subjects are blindly given one or more placebos for initial therapy during a defined time period to determine a “new” base-line of disease (standartization)
Able to assess study protocol compliance
Able to wash-out any exisiting medication
4 Proper Steps in Selecting Cases for a case-control study
- Be consistent
- Select Objectively
- Be Accurate
- Validity
Define efficacy and state which phase a drug’s efficacy is usually assessed
Obtaining a desired/intended result; usually during Phase 2
When would single-blind masking be used
When the study is objective as opposed to subjective; therefore investigator isn’t interacting with the subjects; so it’s ok for them to know which groups subjects are being allocated
Subjects do not know which group they’re in
When is it necessary to use double-blind masking?
When investigator is interacting with subjects; therefore someone outside of the study must know which intervention groups were given what
Baseline Risk Factor
Where the disease started from; what the initial risk factor was
What are the advantages of using intent-to-treat studies?
- The randomization process is preserved
- Preserves baseline characteristics and group balance at baseline which controls for known and unknown confounders
- Maintains stat power (original sample size)
List the 3 ways used to handle drop-outs/lost-to-follow ups
- Intent to treat
- As Treated
- Ignore them (include only those that finished; compliance is pre-defined 80-90%)
What is a measure of association that can be assessed in case control studies?
Odds Ratio
Can case-control studies be prospective?
NO - Always retrospective
What is a descriptive study
Study where data is collected without any changes to the environment; No manipulation of variables
Case-Control,
In case-control studies, Internal validity is dependent upon the selection of ____
Controls
Must have a good control group in determining whether a conclusion is valid or not
Sources in which you can select controls for a case-control study (3)
- Industry/Organization
- Family/Spouses/Friends
Ensures similarity in genetics, environmental, etc.
- Outbreak-sources of Control
Were involved in the same event
What is group allocation based upon in Cohort studies?
- Exposure status
- Commonalities; subjects have something in common
What is a measure of association that can be assessed in cohort studies?
Risk Ratio
Which type of study is good for measuring incidence rates?
Prospective Cohorts….why?
What are the 3 different types of cohort studies?
1. Birth Cohorts
2. Inception Cohort - people that are assembled at a given point based on some common factor
i.e. Nurse’s Health Study
3. Exposure Cohort i.e. 9/11 attack
What are the 3 sources where unexposed persons can be located in cohort studies?
- Internal - same cohort but were unexposed (best method); if there are diff levels of exposure, pick lowest exposed
- General Population
- Comparison Cohort - match according the common characteristics
General advantages of Cohort studies
Good for multiple outcomes of one exposure
Rare exposures
Calculating risk and RR’s
Less expensive than interventional
When ethical issues cause limitations
Represents Temporality (Prospective)
When you have a long induction/letent period (Retrospective)
Advantages of Prospective Cohort Studies
More control over certain data collection processes
Temporality
Can be easier to track and follow-up on patients
Disadvantages of Prospective Cohort Studies
Not good for long induction/latency
Time, Expense, lost to follow-ups
Not efficient for rare diseases (good in case-control)
The exposure or amount of can change over time
Induction vs. Latency
Induction you display symptoms
Latency you get clinically diagnosed
Define cross-sectional studies
Observational, analytical and descriptive studies which examine relationships of health/disease to other variables of interest at the same time by studying large populations
State the other term that cross-sectional studies can be labeled as
Prevelance study; due to the study being at a snapshot in time
3 Variables that cross-sectional studies determine their study upon
- Person - that have the variable of interest i.e. exposure/disease
- Place - choose a geographical region
- Time - pick a specified time period
List the advantages of cross-sectional studies
Quick and easy
Useful for determining prevelance of disease/RFs, measuring current health status and planning for health service across populations, & evaluating differences in sub-groups within a pop at a given time
IRB approval is short and quick
Obtain data across large populations
Disadvantages of Cross-sectional studies
Temporal relationships is hard to determine since study is short/snapshot
Difficult to study diseases w/ low frequency
Survival association is difficult to determine if prevelance personnel are survivors
What are the 2 cross-sectional approaches in obtaining data?
-
Collect data on each member of the population
i. e. pregnancy-smoking data from KC Health Dept. (frequent in city/state evaluations) - Take a sample of the populations and draw inferences of the remainder
What is the sampling scheme most commonly used in cross-sectional studies? What are the subtypes?
Probability Samples
- Simple random samples; random # generator to select samples
- Stratified random samples; mutually exclusive strata (subgroups; age, sex, etc.) = oversampling is done on minorities
What are the common approaches to the collection of new information in cross-sectional data?
- Questionnaires/Surveys
- Physical assessments (lab, clinical, medical records, etc.)
List the 5 large scale cross-sectional studies
- NHANES - National Health and Nutrition Examination Survey
- NHIS - National Health Institute Survey
- NAMCS - National Ambulatory Care Survey
- NHCS - National Hospital Care Survey
- BRFSS - Behavioral Risk Factor Surveillance System
Describe methods, purpose, etc. of NHANES
Surveys health and nutritional status of adults and children
Source pop. is selected to represent US population of all ages
Interview Qs related to demographic, socioeconomic, dietary, health-related, etc.
Exams = medical, dental, physiological measurements, lab tests
Oversamples minorities
Describe methods, purpose, etc. of NHIS
Info. via non-institutionalized/non-hospitalized subjects
Source pop. to represent US pop of all ages
Household Interviews
Not as involved and in depth like NHANES
Describe methods, purpose, etc. of NAMCS
Surveys those who are outpatient/ambulatory = no overnight stays in hospital
Surveys obtain info. about provision and use of ambulatory medical care services
Non-federal, non-office-based
Involves patients who are seeking health care
Describe methods, purpose, etc. of NHCS
Combined national survey describes nation patterns of healthcare delivery in non-federal hospital based settings such as:
Inpatient departments and institutions
ER visits, outpatient depts., ambulatory surgery centers
Describe methods, purpose, etc. of BRFSS
Telephone surveys assessing health risk behaviors, preventative health practices, & health care access primarily related to chronic disease & injury
Adults and children
Monthly data collected
Youth BRFSS done in schools
Define Sensitivity
How well a test can detect presence of disease when the disease IS actually present
TP/TP+FN
Define Specificity
How well a test can detect absence of disease when the disease IS actually absent
Not predective
TN/TN+FP
Define Positive Predictive Value
How accurately a positive test predicts the presence of disease
PPV = TP/TP+FP
Define Negative Predictive Value
How accurately a negative test predicts the absence of disease
NPV = TN/TN+FN
Diagnostic accuracy/Diagnostic precision
Proportion of time that a patient is correctly identified as either having a disease or not having a disease with a (+) or (-) test
TP + TN / TP+FP+TN+FN
An LR+ >10 and a LR- <10 demonstrates that:
A test is most beneficial
If equal to 1 = NOT beneficial
Define Receiver Operator Curves
Graph used for screening tests to show a relationship b/w sensitivity and specificity for tests with numerical (continuous) outcomes
Define validity
Test’s ability to discern between those that DO and DO NOT have the disease
“Telling the Truth”
Reliability
Test’s ability to give the same result on repeated uses
A valid test is always reliable, a reliable test is not always valid
