Study Design Flashcards
2 broad types of quantitative study design
Interventional = forced allocation into study groups (RANDOMIZED)
Observational = no forced allocation to study groups (RANDOMLY SELECTED)
5 types of interventional studies in order of increasing strength of evidence
Phase 0 Phase 1 Phase 2 Phase 3 Phase 4
5 types of observational studies in order of increasing strength of evidence
Case reports/series Ecological Cross-sectional Case-control Cohort
What is a null hypothesis (H0)?
Research perspective which states there will be no (true) difference between the groups being compared
[contrast with alternative hypothesis (H1) — states there will be a true difference between the groups being compared]
What are the 3 statistical perspectives taken in interventional studies regarding the null hypothesis?
Superiority (ex: trying to prove one drug is better than another, or better than placebo)
Noninferiority (ex: trying to prove one drug is not worse than another)
Equivalency (ex: trying to prove one drug is equivalent to another)
Prospective vs. retrospective vs. ambidirectional
Prospective — outcome not yet known
Retrospective — outcome is already known
Ambidirectional — first looking retrospectively, then looking prospectively for additional outcome occurrences
What interventional and observational studies can be done in a prospective fashion?
Interventional — phases 0-4
Observational — only cohort
________ validity = generalizability/applicability of acquired information to non-study patients
_______ validity = validity of acquired information based on study methods
External
Internal
2 key questions to selecting the correct study design
- Is researcher allocating to forced intervention? (Yes=interventional, no=observational)
- For observational studies, how were groups organized? (Disease status, exposure status, common factor, large pop.)
Observational study design in which groups are organized by disease status
Case-control/nested case-control
Observational study design in which groups are organized by exposure status
Cohort
Observational study design in which groups are organized together due to a “common factor”
Cohort
Observational study design in which groups are organized based on data collected across large population
Cross-sectional
Useful study design when studying a rare disease; commonly generates an odds of exposure for each, then an odds ratio as the measure of association
Case-control
Case-control studies are customarily conducted in a _______ fashion
Retrospective
Strengths of case-control studies
Useful when diseases have low occurrence
Useful in determining associations (NOT causation)
Good for assessing multiple exposures of one outcome
Useful when disease has a long induction/latent period
Good when ethical issues limit use of interventional studies
Type of case-control studies derived out of, or conducted after, a prospective previous (or ongoing) study-type (often cohort or interventional study)
Nested case control study
Study design useful when studying a rare exposure; commonly generates the risk of disease/outcome for each, then a risk ratio/relative risk as a measure of association
Cohort study
T/F: a cohort study may not always describe how groups are allocated
True
[once cohort is selected, can divide based on exposure OR disease/outcome]
Strengths of cohort studies
Useful when exposures have low occurrence
Useful when you have a group of study subjects that can be studied to answer numerous research questions
Useful in determining associations (NOT causation)
Good for assessing multiple outcomes of one exposure
Can be conducted ambidirectional
Good for long induction/latent periods (retrospective)
Able to represent temporality (prospective)
Good when ethical issues limit use of interventional studies
Observational studies that capture health/disease and exposure statuses in a large population at the same time; aka PREVALENCE study
Cross-sectional
[information gathered represents what is occurring at a point in time/time-frame across a large population, acquired without regard to exposure or disease/outcome status]
Focuses simultaneously on disease and population characteristics, including exposures, health status, health-care utilization, etc; seeks associations (NOT causation), generatese and tests hypotheses (possible associations), and by repetition in different time periods, can be used to measure change/trends (not in same pts)
Cross-sectional studies
Many _______ studies are large-scale, national surveys or databases capturing different aspects of the ‘population’
Cross sectional
The Framingham Heart Study and the Nurses Health Study are examples of …
Cohort studies
NHANES and NCHS are examples of _____ studies
Cross sectional
[NHANES is National Health and Nutrition Examination Survey, NCHS is National Center for Health Statistics…if it has the word NATIONAL in it — likely a cross sectional]
Purpose/focus, population studied, sample size, and duration of a phase 0 interventional study
Assess drug-target actions and pharmacokinetics in non-therapeutic/non-diagnostic doses (first-in-human use)
Healthy (or diseased patients (oncology)) volunteers
Very small N (e.g., <20)
Very short duration (e.g., a single dose to just a few days)
Purpose/focus, population studied, sample size, and duration of a phase 1 interventional study
Assess safety/tolerance and pharmacokinetics of one or more dosages (first-/early-in-human use)
Healthy or disease volunteers (depends on disease)
Small N (e.g., 20-80)
Short duration (e.g., several days to a few weeks)
Purpose/focus, population studied, sample size, and duration of a phase 2 interventional study
Assess effectiveness (continues to assess safety/tolerability; expands on Phase 1 purpose)
Diseased volunteers (may have narrow inclusion criteria for isolation of effects)
Larger N (e.g., a few to several hundred)
Short-to-medium duration (e.g., a few weeks to a few months)
Purpose/focus, population studied, sample size, and duration of a phase 3 interventional study
Assess effectiveness (continues to assess safety/tolerability)
Diseased volunteers (may expand inclusion criteria and comparison group(s) for delineation of effects; may take superiority, noninferiority, equivalency perspective)
Larger N (e.g., several hundred to a few thousand)
Longer duration (e.g., a few months to a year+)
Purpose/focus, population studied, sample size, and duration of a phase 4 interventional study
Assess long-term safety, effectiveness, optimal use (risk/benefits)
Diseased volunteers (expand use criteria for delineation of long-term safety/effects)
Larger (population) N (e.g., a few hundred to tens-of-thousands)
Longer, wide-range durations (e.g., many months to several years; ongoing)
Last phase of interventional studies before drug goes on the market
Phase 3
Subjects randomly allocated once into a single treatment group; no further randomizations into sub-treatment groups
A. Simple B. Factorial C. Parallel D. Cross-over E. Pragmatic
A. Simple
Subjects randomly allocated into an initial group, then further randomly-divided into sub-group (i.e., multiple randomizations
A. Simple B. Factorial C. Parallel D. Cross-over E. Pragmatic
B. Factorial
Useful in testing multiple hypotheses at the same time because increases sample size requirement
A. Simple B. Factorial C. Parallel D. Cross-over E. Pragmatic
B. Factorial
Subjects simultaneously yet exclusively managed in a single treatment group, with NO switching of groups after initial randomization
A. Simple B. Factorial C. Parallel D. Cross-over E. Pragmatic
C. Parallel
Subjects forcibly switched to other treatment group after initial treatment assignment; between- and within-group comparisons possible; must differentiate wash-out vs. lead-in
A. Simple B. Factorial C. Parallel D. Cross-over E. Pragmatic
D. Cross-over
Wash-out vs. lead-in
Wash-out = clear effects of previous group; seen in cross-over studies
Lead-in = a wash-out period prior to starting a study (e.g., stop current med before starting study on a new med)
Traditional interventional studies are limited/restrictive in terms of group allocation and criteria. What type of INTERVENTIONAL study allows more flexibility so that physicians can alter dosages, add treatments, etc. to best manage the patient/disease?
Pragmatic interventional studies
______ group allocation procedure most commonly utilized; subjects have equal probability of being assigned to each of the pre-defined intervention groups
Random
[contrast with non-random = some investigator-developed way to select patients]
______ randomization = equal probability for allocation into one of the study groups
Simple
_____ randomization = ensures balance within each intervention group [when researchers want to assure that all groups are equal in size]
Blocked
_____ randomization = ensures balance within known confounding variables (i.e., gender, age, disease severity); can also pre-select levels to be balanced within each interfering factor (confounder)
Stratified
Inert tx made to look identical in all aspects to active tx
Placebo (dummy therapy)
[double-dummy = more than 1 placebo used]
Improvement in condition by power of suggestion of being treated
Placebo-effect
Study subjects change their behavior solely due to awareness of being studied/observed
Hawthorne-effect
Ethical conduct of research guided by _____ ______; contains key guiding principles — respect for persons (research voluntary, subjects autonomous), beneficence (risks are justified by potential benefits), justice (risks and benefits are equally distributed)
Belmont report
Consent vs. assent
Consent = agreement to participate, based on being fully and completely informed, given by mentally-capable individuals of legal consenting age (i.e., adults, age 18 in most states)
Assent = agreement to participate, based on being fully and completely informed, given by mentally-capable individuals NOT able to give legal consent
3 types of masking/blinding in interventional studies
Single-blind — subjects are not informed which group they’re in, but investigators know
Double-blind — neither investigators nor study subjects know which intervention/treatment group they’re in
Open-label — both investigators and study subjects know
Procedure for assessing placebo-effect and hawthorne-effect, as well as compliance, before a study begins
Run-in or lead-in phase
Subjects can be give placebo for initial pre-study time frame to determine new baseline of disease; can assess study protocol compliance before actual protocol begins
Group whose role it is to protect human subjects from undue risk; all human studies must be reviewed by them PRIOR to study starting
IRB
______ = genuine confidence that an intervention/tx may be worthwhile (risk vs. benefit) in order to use it in humans
Equipoise
Level of IRB review used for all interventional (clinical) trials; more than low-to-minimal risk
Full board
Level of IRB review used for low-to-minimal risk studies; no patient identifiers
Expedited
Level of IRB review used for chart review or existing samples/data with no patient identifiers, no risk, environmental (community-site) studies
Exempt
Semi-independent committee not involved with conduct of study but charged with reviewing data as the study progresses to assess for undue risk or benefit; can stop the study early
Data Safety and Monitoring Board (DSMB)
2 ways of managing drop-outs/lost-to-follow-up
Include them — intent to treat [most conservative]; use last known assessment, or use baseline measurements. Preserves randomization process, maintains statistical power
Exclude them — per protocol [aka efficacy analysis]; must meet pre-established level of compliance/participation to be included. Biases estimates of effect and can limit generalizability
2 major advantages to interventional study designs
Cause precedes effect (can demonstrate causation)
These are the only designs used for “approval” process (on-label)
