Lecture 5: How Do We Know If a Drug Works Flashcards
What is pharmacology?
it often aims to be a very applied life science
many important checks are in play before and after drugs are introduced to the market to ensure they are safe, effective, and used appropriately
What are clinical trails?
controlled human studies (after considerable pre-clinical development, optimization, animal testing) to assess dosage, administration, safety, efficacy
What are phase 1 clinical trials?
small scale (dozens of subjects), testing for tolerable dosing ranges, bioavailability, excretion
What are phase 2 clinical trials?
intermediate scale (hundreds of subjects), testing for efficacy (sometimes several different dosages), monitoring for safety in greater numbers of patients
What are phase 3 clinical trials?
large scale, randomized, double-blinded trial (both the patients and doctors are blind to the treatment being given), compared against placebo or current accepted treatment
What are systematic reviews and meta-analysis?
this is an approach to combine data from multiple trials, often after a drug has been approved
this approach can increase confidence in our view of the effectiveness of a drug, and help guide future policy regarding drug use
What do forest plots provide data on?
number of trials, size of each trial (size of symbol), outcomes of trials (not always consistent), overall summary of all trials
the OR (odds ratio) is the ratio of the event rate in treatment vs. control (which one is favored in individual trials, vs. overall)
this particular meta-analysis describes the incidence of diabetes in healthy patients treated with a statin vs. a placebo
What are some characteristics of utility and tolerability?
- a large therapeutic index is good - it means the drug is tolerated with minimal toxicity and gives a lot of flexibility for dosing
- there may be high individual variability in drug response (what is toxic for one person, may not be toxic for somebody else)
- other conditions (diet, liver function, kidney function) might need to be considered as they can affect sensitivity (both for effect and toxicity)
- therapeutic index is useful in drug development, but should not be used as a rigid criteria for administration in a clinical setting
What are the concepts of utility and tolerability?
for a drug to be useful, it has to have a beneficial effect, but also be tolerable (in terms of toxicity)
this can be quantified as the “therapeutic index” (ratio of the median toxic dose and effective dose)
in patient studies, effect or toxicity is often described using a quantal dose-response curve
this is different from the quantitative effect curves; instead, it shows the cumulative number of patients who have a pre-defined response to the drug
What is relative risk reduction?
relative risk reduction is defined as: 1 - (event rate in treatment group)/(event rate in control group)
what is an “event” - all cause mortality (DEATH BY ANY MEANS), cardiovascular event, rash, seizure
reporting the relative risk reduction for a drug (while de-emphasizing other important information) is sometimes seen as a misleading practice
Why might reporting the relative risk reduction considered misleading?
the main reason is that it does not convey the magnitude of the baseline risk… it does not capture the difference between a “large” reduction in something that is very infrequent, versus something very frequent
in the case there is indeed a big effect, meaningful and understandable by relative risk
What is absolute risk reduction?
a more descriptive way to report the benefit of taking a drug
the absolute risk reduction describes the absolute number of cases that are prevented by taking a drug (rather than a percentage relative to baseline)
absolute risk reduction = event rate in control - event rate in treatment group
What is the number needed to treat (NNT)?
another useful way to think about absolute risk, or population-level benefit of a drug, is the NNT
X individuals need to take drug in order for 1 to benefit from it
What are the interpretations of the NNT?
a low NNT is good: NNT of 1 means that just about everybody taking the drug will receive the desired benefit
high NNT is not good: means most people will not receive a benefit by taking a drug, means a lot of people taking the drug would be exposed to possible harm (without receiving benefit)
we can also consider harms, with a NNH (Number needed to harm), low NNH is BAD, high NNH is good
if there is a negative absolute risk, it means that an event has a higher rate in the experimental group, so you would describe 1/ARR as the number needed to harm
