Randomized trials and drug testing Flashcards
Core intention of public health and clinical practice
to change natural history of disease so that death and disability can be delayed and the health of the patient or population can be improved
Randomized trial
used to evaluate effectiveness and side effects of a new treatment
Basic design of randomized trials
begin with defined population that is randomized to receive either new or current treatment
Follow the subjects to see how many are improved in the new treatment group compared to the current treatment
Issues to consider when designing randomized trials
- selection of subjects
- allocation of subjects to treatment groups
- data collection and analysis
- crossover
- factorial design
- non-compliance
Selection of subjects and allocation of groups
NB to clearly define in writing the criteria that will govern which subjects will be eligible for and selected in a study. Must be defined in a way that same subjects will be picked if chosen by someone else
Studies with comparison
Studies without comparison
Methods used for selection of subjects
Historical controls: comparing records of patients who were treated in the past before the new treatment became available to the outcomes of a disease that is being treated currently. Some problems include general changes in health over time
Simultaneous non-randomized controls: a simultaneous control group is selected but in a non-randomized manner e.g. people born on even days in group A and people born on odd days in group B. There must be no chance of selection bias.
Randomization: best way to reduce selection bias. assignment of next individual cannot be predicted and modified. Tables of random numbers can be used to assign people randomly into groups
Stratified randomization: if there is a known variable that we believe could influence prognosis
step 1- study population is separated into sections determined by each variable that may influence outcome of treatment
step 2- participants in each section are then randomly assigned to treatment groups
This assures that the groups are balanced and the data can be compares
Data collection of subjects
must be of the same quality
Treatment: essential for treatment group to which patient was assigned to be known and which therapy patient received and whether they completed treatment
Outcome: including improvements and changes and side effects need to be recorded. NB that a criteria be explicitly stated for all outcomes to be measured and that all outcomes measured comparably
Prognostic profile at entry: if there is a known risk factor for a bad outcome, we want to make sure that the randomization has resulted in groups that are comparable
Crossover
Planned crossover: subjects begin study on treatment A and later switch to treatment B and vice versa. Patients can then serve as their own control
Unplanned crossover: subject that was assigned by randomization to a particular group moves into another group
Certain factors to consider for planned crossover
Carry over: effects of the initial treatment affect the observations for the second treatment. There has to be a wash out period between treatments which ensures effects of second agent are only due to that agent
Psychological responses: individuals may react and feel differently when they start the first treatment versus how they behave when they start the second treatment. Researchers need to ensure that any differences observed are indeed due to the agents being evaluated and are not due to differences in psychological responses or adherence to treatment
Factorial design
same study population is used for testing two different drugs provided the drugs mode of action are distinct and the expected outcomes of the two drugs are different
Non- compliance
when patients do not act in accordance with the assignment treatment
Overt non-compliers refuse to participate and are referred to as dropouts
Covert non compliers do not take the agent as requested and do not inform the researchers of their refusal to comply thereby affecting the data
Ways of expressing the results of randomized trials
1 calculate efficacy of a drug or vaccine
-subtract rate of developing disease in the people who received the vaccine from the rate of developing disease in people who received the placebo
2- calculate the ratio of the risks or relative risk in two treatment groups
3- compare survival curves of the groups
3- estimate no. of patients who would need to be treated to prevent one unfavorable outcome
Generalize results by
Internal validity: degree to which the results of a study are true for the studied population. Depends on no. of systematic errors in measurements and how well the study was conducted
External validity: represents degree to which results of a study are relevant for populations other than the population that was studied
Phases in testing new drugs
Phase 1: small studies to asses the toxicity and pharmacology effectiveness of the drug (typically 20-80 patients)
Phase 2: testing efficacy and relative safety in people (100-200 participants)
Phase 3: large scale randomized trials to further test efficacy and relative safety
Phase 4: monitor that there are no adverse effects which may take years to develop or which occur seldom that they are only detectable in very large populations
Ethical issues involved in the use of clinical trials
- is it ethical to randomly withhold a drug from a patient -if it may cure a serious or life-threatening disease
- is it ethical to use a placebo or to withhold a treatment that is known to help?
- Can true informed consent be obtained
- In what situations should a trial be stopped earlier than originally planned
