Medicine SC080: Is This The Best Drug For Me?

Question 1

Most important properties of a drug

Answer 1

1. Efficacy
2. Quality and Safety
3. Tolerability
4. Cost-effectiveness

Question 2

Why do we need new drugs?

Answer 2

1. New conditions
2. New pathogens
3. New understanding of disease mechanisms (e.g. new receptor, enzyme / pathway identified)
4. New mutations
5. More choices
6. To differ from generics (more competitive)
7. Resistance
8. More specificity
9. Better efficacy
10. Better safety
11. Better tolerability
12. Better compliance

Question 3

Clinical trial

Answer 3

• Assessment of a **method of treatment in a **planned experiment involving ***patients

Phase 1:
- Test an experimental drug or treatment for the first time
- Small group of people (20-80)
- Evaluate **safety, safe dosage range, identify **SE

Phase 2:
- Larger group of people (100-300)
- Evaluate ***effectiveness and to further evaluate its safety

Phase 3:
- Large groups of people (1000-3000) (usually randomised + blinded)
- Confirm its **definitive effectiveness, monitor SE, **compare it to commonly used treatments, and collect information that will allow the experimental drug or treatment to be used safely

Phase 4:
- ***Post marketing studies
- Delineate additional information including drug’s risks, benefits and optimal use
- Rare side effects may not be revealed in clinical trials (if the incidence is 1 in 10,000, then the chance of missing this is 37% when 10,000 patients have been exposed to the treatment)
- Clinical trials may not detect long term adverse effects like cancer
- SE if unexpected, may be missed (e.g. cough and ACEIs)
- Standard assessment of adverse effects may miss side effect (valvulopathy after dexfenfluramine)

Reporting adverse drugs reactions:
- Voluntary in HK to Drug Office, Dept of Health
- Voluntary in Mainland China, but there are targets
- Yellow card system in the UK
- In USA, reports from professionals, consumers and manufacturers to the FDA Adverse Event Reporting System

Question 4

Trial designs

Answer 4

Best trials are:
1. Prospective

2. Controlled
   - need to know how good new treatment compared to no treatment (placebo) and current standard / best treatment
3. Randomised (***most important)
   - even out biases (a process at any stage of inference tending to produce results that depart systemically from true values)
   - facilitates blinding
   - allows null hypothesis that any difference between treatment groups are due to chance
4. Blinded
   - knowledge of assignment influences the subjects’ response, investigators’ conduct + evaluation
   - sometimes, blinding cannot be achieved easily, e.g. warfarin needs dose adjustments
   - some trials use blinded evaluation of endpoints, especially if the endpoint is objective, such as a biochemical test or an unequivocal event such as death

Question 5

Study subjects

Answer 5

• Subjects characteristics are important
• Defined by:
  —> Inclusion criteria
  —> Exclusion criteria
• Disadvantage: Affect ***generalisability of the conclusions

Question 6

Power of study

Answer 6

• Usually calculated by a statistician, who needs to know
  —> P value required (usually <0.05)
  —> Effect size: Anticipated difference between treatments relative to SD
  —> Drop out rate
• Usually at least a power of 80% to detect a difference between treatments at the 5% level of significance is expected in a clinical trial

Question 7

Adverse events

Answer 7

Definition:
- Any untoward event (does NOT need to be related to treatment) (Might not have causal relationship (i.e. not an ADR because of the drug))

Documentation:
- ALL adverse events must be documented
- Management: No action / Observation / Treatment / Hospitalisation / Surgery etc.
- Outcome: Fully / Partially recovered, Persistent, Death

Question 8

Efficacy and Effectiveness

Answer 8

Efficacy: How good the drug works under ideal conditions
- in those taking drug as prescribed (“Per protocol”)

Effectiveness: Efficacy + Take in account tolerability
- in those offered drug (“Intention to treat”)

Question 9

***Important statistical terms

Answer 9

1. Range, Median
2. Mean, SD, Standard error of the mean, 95% CI
   - 95% CI = Mean +/- 1.96 (or 2)xSE
   - SE = SD / √N
3. Null hypothesis, P value
   - Null hypothesis: any difference observed is due to chance
   - P value: probability of obtaining a test statistic at least as extreme as the one observed
   - Type 1 error: null hypothesis wrongly rejected when it is true i.e. false positive
   - Type 2 error: null hypothesis not rejected despite being false i.e. false negative
4. T-test, Chi-square test
   - T-test: check if 2 means are reliably different from each other
   - Chi-square test: check for patterns / relationships in categorical variables
5. Correlation coefficient
6. Odds ratio
   - (Odds that a case was exposed) / (Odds that a control was exposed)
   = (Case exposed/Case unexposed) / (Control exposed/Control unexposed)
7. Relative risk, Relative risk reduction
   - RR = EER (experimental event rate) / CER (control event rate)
   - RR reduction = 1-RR (or ARR / CER)
8. Absolute risk reduction, NNT
   - ARR = CER - ERR
   - NNT = 1/ARR

Question 10

Statistical tests used in clinical trials

Answer 10

Continuous data:
- Summary statistic: Mean, SD, SE, Median
- Statistic test: ***T-test, Mann-Whitney
- Difference tested: Mean, Rank order
- If outcome depends on several factors —> Multivariate method: Multiple regression, None as sample size usually small

Categorical data:
- Summary statistic: Frequencies
- Statistic test: ***Chi-square test
- Difference tested: Observed vs Expected frequencies
- If outcome depends on several factors —> Multivariate method: Logistic regression

Time to event (“Survival data”):
- Summary statistic: Median
- Statistic test: Log-rank
- Difference tested: Observed vs Expected events
- If outcome depends on several factors —> Multivariate method: Cox proportional hazard

Question 11

Limitations of clinical trials

Answer 11

1. Patient selection
   - Inclusion / Exclusion criteria
   - Other known selection bias
2. Protocol restrictions (e.g. fewer drug interactions)
3. Good medical care, Frequent follow up (in real life may not be the same)
4. Good compliance (in real life may not be the same)
5. Limited duration of trials, some effects (e.g. carcinogenesis) may be long-term
6. Very expensive
7. Address a limited number of hypotheses

Question 12

Reliability of different study designs

Answer 12

Ascending reliability:
1. Case report (Retrospective)
2. Case control study (Retrospective usually)
3. Cohort (Prospective usually)
4. RCT (Prospective usually)
5. Meta-analysis (Retrospective usually)
6. N=1 trial (trial directly on your patient, prospective usually)

Question 13

Systematic review vs Meta-analysis

Answer 13

Systematic review:
- There may be a number of trials addressing the same research question
- Results may vary between them
- Need to summarise the results + arrive at conclusions that take into account of all the trials

Meta-analysis:
- A statistical way of combining the results of different studies
- Usually the results of a trial is given a certain statistical weight (a large trial is given more weight)

Question 14

Pharmacoeconomics (SpC Revision)

Answer 14

Methodology:
1. Cost benefit
2. Cost effectiveness
3. Cost minimisation
4. Cost utility

Cost measure:
- ALL measure in terms of dollars

Outcome measures:
1. Cost benefit: Dollars (Net gain / loss of money)
2. Cost effectiveness: Natural units (e.g. Cost per life, Cost per stroke prevention)
3. Cost minimisation: Equivalent (Which drug is cheaper when equivalent)
4. Cost utility: QALY