FinalExamReview-YamalTopics Flashcards
- Describe different types of studies and be able to rank in terms of strength of evidence
Rank in terms of strength of evidence (strong to weak)
1) Clinical Trial
2) Cohort Study
3) Case-control study
4) Cross-sectional study
5) case series
Descriptions:
Clinical Trial - A prospective study comparing the effects and value of intervention(s) against a control in human beings.
Cohort Study - Study where one or more samples (cohorts) are followed up prospectively and subsequent status evaluations with respect to a disease or outcome are conducted. In this type of study participants are identified based on their exposure characteristics (risk factors) associated with the disease or outcome of interest.
Case-Control Study - Study that compares patients who have a disease or outcome of interest (cases) with patients who do not( controls, and looks back retrospectively to compare how frequently the exposure to a risk factor is present in each group to determine the relationship between the risk factor and the disease.
Cross-sectional Study -
Case Series - Article that describes individual cases. (Advantages - can ID trends/disease; Disadvantage - lack of generalizability)
- Be able to identify different types of bias including their definitions
Information Bias - Misclassification of information/data collected
Confounding -the association between the predictor and outcome is distorted by extraneous factors (confounders)
Publication Bias - Studies with significant results are more likely to be published than studies without significant results
Selection Bias - A difference in characteristics of participants selected for a study and the population from which the participants were selected.
Recall Bias - Bias that occurs when participants are asked to recall events in the past that they may recall differently depending on current health status.
Ascertainment Bias - Loss to follow up, impacts perceived effect of treatment if participants are not loss to follow up at random.
- Define Phase I-IV trials
Phase I: Goal is to determine which dose of the drug is safe and most likely to show benefit.
Phase II: Goal is to identify agents with potential efficacy (NOT designed to definitively test efficacy). Discard agents without promise. Provide rationale for continuing the experiment.
Phase III: Goal is to define the “best” treatment which has implication of changing current practice
Phase IV: Goal is to monitor adverse effects, long-term morbidity and mortality after treatment is applied to large number of patients and follow up for a long period of time
Describe a Phase I (Clinical Pharmacology and Toxicity) clinical trial.
Primary Concern: Safety
Participants: Healthy volunteers or patients who have already tried and failed to improve on existing standard therapies.
Goal: Estimate tolerability and characterize pharmacokinetics and pharmacodynamics. (Max. Tolerable Dose etc)
n is typically 15-30 participants
Clinical Considerations: Route, schedule and starting dose, Plan for escalation, patient selection, extent of toxicity
Statistical Considerations: Number of patients per dose, Well-defined escalation theme, stopping rule, decision rule for recommended dose
- Be able to explain how the eligibility criteria can affect generalizability and differences in results between phases of studies.
Goals:
- setting certain eligibility criteria can increase the likelihood of detecting an effect
- minimizing the possibility of adverse effects (exclude pregnant women, exclude people without adequate liver/renal function)
Advantage:
Heterogenous groups let us study the mechanism in action
Impact:
- Many studies are performed in academic medical centers which may not be representative of general population
- underrepresentation of minorities
- underrepresentation of women
- participants may be different from non-participants in many significant ways (health, disease, racial or sex, etc)
- Identify differences in outcomes and how they would be analyzed
Ideal Outcome:
- valid and reliable
- easy to observe
- free of measurement error and other bias
- ideally complete observation in all subjects
- capable of being observed independent of treatment assignment
- clinically relevant
Types:
- continuous
- categorical (attractive for descriptive purposes, but we lose power and can lose information, difficult to generalize as we saw in the studies we read on COVID treatments)
- dichotomous
- count, ordinal
- nominal
- time to event
- composite or global
- Be able to explain properties of good surrogate outcomes and what their limitations are
Surrogate - A biomarker that is intended to substitute for a clinical endpoint; expected to PREDICT the effect of an intervention on the clinical outcome.
- strongly associated with definitive outcome
- part of the causal pathway
- yields the same inference as the definitive outcome
- responsive to the treatment
- short latency
Prentice Criteria:
- proposed outcome predictive of true (clinical) outcome
- proposed surrogate must fully capture effect of intervention on (clinical) outcome
Examples:
1) Treatment of hypertension:
- Clinical outcome - Myocardial Infarction
- Surrogate - Change in blood pressure
2) Treatment of MS
- clinical outcome - relapse lesions
- surrogate - MRI changes
- Identify issues with incorporating baseline measures of the outcome into change from baseline as the outcome and be able to discuss alternative approaches.
- if the baseline is related with the outcome can induce spurious correlation
- percent changes is dependent on what the baseline value is (we can have the same percent increase, but magnitude of change is very different)
- Identify the various types of control groups and their roles in different types of trials.
Placebo - same look and delivery method as treatment, but lacks an active ingredient.
No Treatment - control group has no treatment
Standard of care - control group receives the standard of care that is currently in practice.
Another treatment - not necessarily considered “standard of care”
Historical Control (new treatment is used and compared to historical outcomes from database)
Prospective registries as controls
- Describe the randomization process and different types of randomization procedures
- independent central unit should be responsible for developing randomization process (biostatistician for single center or data coordinating center for multicenter)
- Best if investigators are blind to randomization
- Sequenced and sealed envelopes that have assignment
- randomization list kept in pharmacy for double-blind drug studies
- web-based randomization - capture basic eligibility data, gives randomization assignment
- Describe the role of blinding and types of blinding
Single Blinded:
- A trial in which the participant is unaware of which treatment group they have been assigned, but the investigator is aware of the assignment.
Advantage: design is similar to unblinded, simple to carry out
Disadvantage - investigators van add bias b/c they are aware of treatment assignment. Investigator may intentionally or unintentionally un-blind the participants.
Double Blinded:
- A trial in which neither the participant nor the investigators following the participants know the identity of intervention assignment.
Advantages:
- reduce risk of bias
- account for placebo effect
Disadvantage:
- must manufacture placebo to match treatment
- periodic sampling drug content
- lab test such as checking serum level may be helpful but must be confidential
Triple Blinded:
- patients, investigators and data monitoring committee are all blinded
- disadvantage: DMC’s ability to monitor safety and efficacy can be hampered and this can be counterproductive
- improvement: DMC is blinded at first but code can be broken upon request.
- Describe why it’s reasonable to plan for testing non-inferiority and then superiority, but not the other way around.
Non-inferiority is the gate keeper. Similar to the ANOVA, if ANOVA is significant then we go in and look in more detail.
Non-inferiority first because if you get significant result it is ok to go ahead and look at the superiority hypothesis.
Power - Superiority requires a very large power in order to detect significance. Sample size for non-inferiority trial is not required to be as large as for superiority.
- Show how an estimated sample size can incorporate an expected loss to follow-up rate.
Example: Assume a 15% loss to follow up that is assumed to be at random-
N*=N/(1-.15)
Example: Assume 15% loss to follow up and assume not at random thus additional sensitivity analysis may be required -
N*=N/((1-.15)^2)
- Show what the risk is of not considering the clustering effect in sample size estimation for cluster randomized trials
If cluster effect is not accounted for, then variance is underestimated, standard error looks smaller than it actually is and the likelihood of committing a type 1 error increases.
- Explain and show what is the most efficient design in terms of allocation (balanced vs unbalanced) between two study groups. What is the effect of unbalanced designs?
Most efficient is balanced allocation.
Increased variance, increased required sample size to achieve the same power
(Or if sample size is constant-> increased variance and decreased power)
- Describe what Cohen’s effect size is and general guidelines for magnitude of effect sizes.
Cohen's Effect Size: - based on behavioral science data - essentially the same as a z-score for two sample test with s-pooled as SE - D=(mu1-mu2)/s-pooled General Rules: 0.2<=D<0.5 (Small Effect) 0.5<=D<0.8 (Medium Effect) D>=0.8(Large Effect)
- Describe what the differences are between the (1) Haybittle & Peto, (2) Pocock, and (3) O’Brien and Fleming sequential boundaries are with respect to the likelihood of stopping early and for the final look.
(1) Haybittle & Peto - constant until last time point and then increased change to reject null
(2) Pocock - constant throughout
(3) O’Brien and Fleming Increasing probability of rejecting null at every look
- What is the advantage of alpha spending designs over group sequential designs?
Group sequential designs are cumbersome and alpha spending designs are less cumbersome.
- Group sequential designs can have restrictive monitoring times that require equal increments of information (# patients) and sometimes causes administrative difficulties
- Alpha spending are more flexible than group sequential designs so the idea is to spend(distribute) the total probability of false positive risk (Type I error) as a continuous function of the information time (alpha spending function)
- In a 3+3 design, how is the maximum tolerated dose determined?
3+3 Design:
- Treat 3 participants at dose level k
- If no DLT, escalate to dose level K+1
- If 2+ DLTs, de-escalate to dose level K-1
- If 1 DLT, treat 3 additional participants at dose level K
- (Now 6 total participants) If 1 DLT, escalate to dose level K+1
- (Now 6 total participants) If 2 DLT, de-escalate to dose level K-1
MTD is the highest dose where 0 or 1 DLT observed
NOTE: DOSE LEVELS MUST BE PRE-SPECIFIED
- What are some strengths and limitations of the 3+3 design?
Limitations: - Ignores dosage history other than the previous 3 patient cohort -Imprecise and inaccurate MTD estimation -low probability of selecting true MTD -high variability of MTD estimates -Dangerous outcomes Can improve with 5+5 or CRM
Strengths?
- simple design, small, doesn’t require statistician.
Define Maximum Tolerated Dose.
- Every person has a unique Maximum Tolerated Dose before reaching a Dose Limiting Toxicity
- Ratios/Proportions of people at each dosage level for which that dose is their MTD
- We look for the MTD that has a ratio of some target percentage (say 25%)
- Describe the intention-to-treat principle.
all participants randomized and all events, as defined in the protocol, should be accounted for in the primary analysis
