clinical trial methodology

Question 1

Advantages of cross over trials

Answer 1

Between patient variability is reduced as patients act as their own control. Reduces number of participants required.

Question 2

Disadvantages of cross over trials

Answer 2

Harder to manage if numerous treatments are given at multiple time points. May get carryover effects. May get rebound effects. May get period effects. Participants are more likely to drop out before the clinical trial is over.

Question 3

What is the required length of a washout period?

Answer 3

At least 3 half-lives.

Question 4

What is the rebound effect?

Answer 4

Administration of a drug may cause a spike once withdrawn, e.g. giving an antihypertensive drug may reduce blood pressure, but once the drug is stopped it may cause an excessive spike in blood pressure. Administering the second drug whilst this occurring has potential to make results falsely better.

Question 5

What is the period effect?

Answer 5

Drugs may have different effects depending on whether they’re administered first or second

Question 6

What is the run in period?

Answer 6

The period before drug administration in parallel design studies where ineligible patients can be screened out and eligible patients may be weaned off previous drugs, etc.

Question 7

Advantages of parallel design trials

Answer 7

Easy to manage and interpret. It is suitable for all diseases. Can compare numerous treatments without over complications.

Question 8

What is the inclusion criteria?

Answer 8

Patient characteristics which make them eligible for participating in a study

Question 9

What is the exclusion criteria?

Answer 9

Patient characteristics which prevent them from participating in a specific study, e.g. pregnancy, concomitant medications, smoker

Question 10

What is factorial design?

Answer 10

Two treatments are evaluated in a single trial, patients may be randomised to receive 1 of the treatments, or both, or placebo. Usually incorporates 4 arms.

Question 11

Disadvantages to parallel design trials

Answer 11

Increased requirement for participants. There is between patient variability between groups. Must ensure that characteristics which may affect outcome (e.g. disease stage, age) ae balanced in each group.

Question 12

What is a clustered randomised design?

Answer 12

Participants are formed into groups based on organisations, etc. (e.g. hospitals) and are randomised to a treatment, rather than individual patients.

Question 13

Disadvantages to clustered randomised design

Answer 13

May get confounding and selection bias - variation in characteristics may cause influence on the treatment response, which cannot be minimised, e.g. affluent areas may be more likely to take advantage of treatment, therefore gain more support than deprived areas. May also get the cluster effect - outcome could be influenced by the investigator.

Question 14

What is group sequential design?

Answer 14

Intervention is given to a group of patients. The study can then be modified for the next group of patients. This is a flexible trial design.

Question 15

Examples of modifications for group sequential trial design

Answer 15

May drop treatment arms which don’t show effective results. May enrol more patients into treatment arms which show promising results. May adjust the dose so that the effect is seen but toxicity is reduced.

Question 16

Downsides to group sequential trial design

Answer 16

Promotes bias due to changing the trial design to better results - not a true reflection of what the real world response will be.

Question 17

Why are control groups important in a clinical trial?

Answer 17

Allows you to determine that the efficacy seen in the outcome of your trial, is due to the intervention rather than baseline characteristics, or disease progression, etc. Demonstrates what would have happened if the patient had not received the intervention, therefore it provides clinically relevant data on whether the intervention is useful.

Question 18

What are internal controls?

Answer 18

This involves having an arm in your clinical trial during the same time which receives a control, rather than the intervention.

Question 19

What are external controls?

Answer 19

This can involve subjects in another study that has already been carried out, e.g. historical controls. Not often used, but may be used where it would not be considered ethical to give patients a placebo.

Question 20

What is a placebo control?

Answer 20

Placebo appears exactly the same as the intervention, but does not contain the test drug. Usually used in double blinded trials. Allows you to assess the placebo effect.

Question 21

What is a no-treatment control?

Answer 21

Participants in this arm will not receive any treatment. It is not possible to blind patients or investigators to this sort of control. Only used if it is not possible to use a double-blind study.

Question 22

What is a dose-response control

Answer 22

Participants may be randomised to a fixed dose group for a comparison of groups on their final dose.

Question 23

What is an active control?

Answer 23

This is used to show efficacy of a known compound which will definitely cause an effect, to see how the intervention compares. Can be used to assess assay sensitivity.

Question 24

When is it ethical to give a placebo control?

Answer 24

It is only ethical to give a placebo control if there are no standard treatments available, e.g. in mitochondrial diseases. Also ethical if the condition is mild and not receiving a treatment will cause no harm to the patient.

Question 25

When might it be difficult to use a placebo control?

Answer 25

If the treatment in question is a topical cream, using a topical cream as a placebo will still cause some effect, e.g. moisturising the skin, may have knock on effect. In these cases the study may have to not be double-blind, e.g. might have to have different form of application.

Question 26

What are disadvantages of external controls?

Answer 26

There may be selection bias. As the data is already available there is a risk that external controls will be chosen to increase the apparent effect of the intervention. Can prevent this by choosing controls before study.

Question 27

When might external controls be required?

Answer 27

This would be suitable when the effect is dramatic, e.g. anaesthesia. Or when looking at a serious condition with high mortality rates, where there is no satisfactory treatments (e.g. no good comparator to compare to, use previous data as more ethical)

Question 28

What is an endpoint?

Answer 28

An outcome that is measured during a clinical trial to determine whether the intervention was a success or failure.

Question 29

What are the types of endpoints?

Answer 29

Safety endpoint and efficacy endpoints.

Question 30

What are the characteristics of a good endpoint?

Answer 30

Should be clinically relevant to the disease in question, should be validated from other trials (e.g. can’t just make up an endpoint, it needs to have been studied previously and proven to be effectively linked to disease), should have an objective measure/scale so it can be determine whether the endpoint is successful or not, must be achievable (e.g. for advanced cancer trial, shouldn’t aim to cure cancer, but reduce mortality, or reduce progression)

Question 31

Give examples of biomarkers.

Answer 31

Biomarkers are measured to indicate a normal biological process, or a pathogenic process, in response to an intervention. E.g. may assess blood pressure, heart rate, to determine whether the effect of an antihypertensive drug is clinically relevant. Also cholesterol - indicative of diabetes.

Question 32

What is a hard endpoint?

Answer 32

These are measurements which are well define and are definitive within the disease in question, e.g. death, time to disease progression.

Question 33

What is a primary endpoint?

Answer 33

An endpoint which provides evidence that the clinical effect of a treatment would support a regulatory claim, e.g. survival. It is enough by itself to show efficacy.

Question 34

What is a clinical endpoint?

Answer 34

These are variables which reflects how a patient might feel or function, e.g. stroke, survival. Focuses on the patient QOL.

Question 35

What is a secondary endpoint?

Answer 35

Any additional characterisation of a treatment, but ones which could not by themselves be convincing of a clinically significant treatment effect, e.g. increased nausea via patient questionnaire, or the pharmacokinetics of the drug. They contain valuable information but the trial was not specifically designed to evaluate them.

Question 36

What are surrogate endpoints?

Answer 36

These are biomarkers which are used instead of clinical endpoints, e.g. epidemiological evidence may be used to select a surrogate endpoint to predict the clinical benefit, or harm. For example, the clinical endpoint may be atrial fibrillation, but the surrogate endpoint would be the systolic blood pressure (before atrial fibrillation occurs).

Question 37

What is a soft endpoint?

Answer 37

Not an exact measurement, linked more to QOL, e.g. patient questionnaire results.

Question 38

What is a composite endpoint?

Answer 38

A combination of multiple clinical endpoints, used to reduce the required participant number when rare events are concerned in the primary outcome.

Question 39

Can secondary endpoints be correctly analysed if the primary endpoint is negative?

Answer 39

Yes. Although the aim of the trial follows the primary endpoint, there may still be relevant information about the intervention in the secondary endpoints. E.g. reduced nausea reports may be the secondary endpoint, however the primary endpoint may have been decreased disease progression. Disease progression may not have worsened but side effects have bettered.

Question 40

What are the advantages of using biomarkers and surrogate endpoints in clinical trials?

Answer 40

Biomarkers are exceptionally useful for testing hypotheses and altering treatment response accordingly, e.g. identifying an exposure-effect relationship. Allows acceleration through clinical trials. Surrogate endpoints are a direct measure of how a patient feels, e.g. you may expect a 5ms increase in WT interval to cause a person to feel dizzy, but they might not. Contrarily, less than 5ms shouldn’t cause a person to feel dizzy, but in some cases it might. Are therefore more accurate representations of what is actually happening.

Question 41

What are the characteristics of a good surrogate endpoint?

Answer 41

They must be relevant to the mechanism of drug action, and to the progression of the disease. Must be validated in other trials/meta-analyses. Must be a sensitive biomarker, to detect true negatives and prevent false positives.

Question 42

What is the ‘No Observable Adverse Effect Level’ in pre-clinical studies

Answer 42

NOAEL involves the concentration at which there is not an increase in adverse effects in pre-clinical animal models, in comparison to the control group. This dose should be used to calculate the human equivalent dose for phase 1 clinical studies.

Question 43

When is it not appropriate to scale doses to body weight/surface area?

Answer 43

When drugs are administered topically, etc. therefore are limited by the local absorption, it is not appropriate to scale to body weight. Instead the HED should be normalised to the drug concentration or amount of drug at the application site.

Question 44

What are the steps for determining MRSD via the NOAEL route?

Answer 44

Determine NOAEL in animal model. Convert this to HED via multiplication of surface area and conversion factor.. Then divide by a safety factor (usually 10).

Question 45

What are the steps for determining MRSD via the MABEL (minimal anticipated biological effect level) approach

Answer 45

Utilise all relevant pharmacogenetics and concentration-effect data from in vivo and in vitro studies to predict human MABEL. Integrate this into PK/PD data to predict the pharmacological response in humans at multiple dose levels. Account for inter-species differences in affinity and exposure and duration of treatment, to produce MRSD. Additional safety factor may be used if there is uncertainty of MABEL

Question 46

What is the Maximal Tolerated Dose (MTD)?

Answer 46

The dose one level below which the dose-limiting toxicity was achieved.

Question 47

What is the Modified Fibonacci dose-escalation method?

Answer 47

As the dose increases, the ratio which they increase by gets smaller, e.g. 2.00, 1.67, 1.50 then 1.33 for all subsequent increases.

Question 48

What is the Continual Reassessment Method of dose escalation?

Answer 48

An estimate dose-toxicity curve is produced based on preclinical data. The estimated MTD is administered. Based on whether toxicity occurs, dose can be increased or decreased until dose is found.

Question 49

What is the pharmacologically guided dose escalation method?

Answer 49

AUC for the MTD is similar in mice and humans. Therefore the AUC is determine at each dose level and each subsequent dose is determine based on the AUC of the previous dose and the target AUC..

Question 50

What are the disadvantages to the pharmacologically guided dose escalation method?

Answer 50

This does not take into account between patient variability. Moreover, this is assuming linear kinetics of the drug, which may not be the case.

Question 51

What are accelerated phase 1 designs?

Answer 51

Accelerated phase 1 designs incorporate a speedier approach via optimisation of the design (e.g. incorporating bioavailability with dose-escalation, high dose increments) to ensure faster progression onto further trials. These designs have greater risk than others, particularly when there is a steep dose dependent toxicity.

Question 52

What is adaptive dose finding?

Answer 52

This method of dose-escalation involves treating the next subject based on the outcome of the previous subject. This prevents all patients receiving ‘waste doses’. Fewer subjects are assigned to doses which are too high or too low.

Question 53

What are the disadvantages of adaptive dose finding?

Answer 53

You must have infrastructure with rapid communication between lab findings at the trial site with the unblinded analysis centre. Also would need rapid randomisation software to allow fast computation of the next patient to be assigned the next dose. Drug-supply must also be flexible as it cannot be estimated how much will be needed.

Question 54

What aspects are important for designing a patient questionnaire?

Answer 54

Answering the clinically relevant questions, and having appropriate answer options.