E-quiz 3b
Terms
undefined, object
copy deck
- Phases of clinical trials in humans
- Phase I sets tolerable doses, describes metabolism of drug Phase 2 determines therapeutic efficacy Phase 3 compares new therapy to placebo or existing option(s) Phase 4 monitoring for adverse effects in the general population
- Intervention studies
- The major distinction between cohort and intervention studies is: the treatment or exposure is randomly assigned to study subjects by the investigator Group assignment is unknown to researcher and subject whenever possible - “blindingâ€
- Pop hierarchy & subject selection
- reference population - those to whom you want to apply the results experimental population - those in which the study is conducted study population - those who are willing and eligible treatment / control arms - random assignment
- Quality of info
- accurate and complete confounder information consistency of data source(s) for both intervention and comparison maintaining high level of follow-up
- Actual # of events
- what is anticipated incidence in intervention and comparison arms sample size for study completeness and consistency of outcome ascertainment
- Types of intervention studies
- Preventative reduces the risk of developing a disease
- Example 1
- Researcher wishes to find out if a contraceptive film containing a specific drug is effective at preventing HIV infection. 250 commercial sex workers are enrolled and randomized to active or inactive film, all provided with condoms, and warned about the dangers of unprotected sexual activity. Followed for 6 months.
- Example 2
- Researchers wish to find out if a new drug is more effective at preventing death or severe disability in stroke victims than the current standard treatment. 120 incident stroke victims are randomized to standard or new therapy and followed for 1 year.
- Random allocation into study arms
- control/comparison arm may be a placebo or an alternate intervention allocation is random to enhance comparability of subjects in arms done after eligibility and willingness established use blocking to ensure balanced distribution of specific factors
- Ethical considerations
- Study participation must be voluntary Informed consent must be obtained Balance of doubt and belief regarding agent or procedure to be evaluated Cannot knowingly instigate harm, only modify its removal Anticipate unintended effects Stopping rules
- Stopping rules
- Stopping a study before its scheduled endpoint criteria established prior to initiation of the study interim measures of outcome monitored by an external body or group of researchers allows very good or very poor outcome to be shared or avoided
- Compliance/adherence
- more difficult to maintain compliance for long periods of time or for complex protocols sources of noncompliance - side effects, forgetting, dropping from study and cross-over measure and track non-complying subjects anyway Analyze on planned treatment status, not actual – “intent to treatâ€
- Ascertainment of outcome
- measures of outcome must be as objective as possible must be measured consistently & reliably across all subjects single or double blinding to avoid biases
- Factoral designs
- Increasing efficiency by testing multiple interventions, which have difference outcomes and do not use similar active pathways multiple (usually 2) stages of randomization, allowing for more than one intervention may look for synergy or examine interventions completely independently
- Analysis & interpretation
- compares rates of outcome in treatment and control adequate sample size and randomization decrease threats from chance and unknown confounders determine which subjects to include in analysis and how to deal with noncompliance - “intent to treatâ€
- Types of error
- Type I error --rejecting the null when it is true (alpha or p-value) Type II error -- failure to reject the null when it is false (beta value)
- Sample size
- Sample size calculations require estimates of: magnitude of effect proportion of population with either exposure or disease/ expected incidence; and designated levels for alpha and beta one- or two-sided test
- Power calculations
- power calculations use similar equations and parameters as sample size However in this situation you insert a sample size and solve for beta (power)