2017-11-27

Definition of clinical trial

  • Intervention trials ideally take the form of a randomized controlled trial (RCT) in which the intervention is compared with a control and the allocation to treatment or control is randomized.
  • The control may be a placebo or an alternative intervention.
  • RCTs are often used for therapeutic trials in which different treatments for a given disease are compared in a clinical setting.
  • Since RCTs offer a unique opportunity to eliminate the problems that beset observational studies, the results of such trials are generally considered to be a “gold standard”

The CONSORT Statement

  • CONSORT Statement is an evidence-based, minimum set of recommendations for reporting randomized trials.
  • It offers a standard way for authors to prepare reports of trial findings, facilitating their complete and transparent reporting, and aiding their critical appraisal and interpretation.
  • Comprises:
    • 25-item checklist
    • flow diagram
  • http://www.consort-statement.org/

The P.I.C.O. Model for Clinical Questions

P Patient, Population, or Problem How would I describe a group of patients similar to mine?
I Intervention, Prognostic Factor, or Exposure Which main intervention, prognostic factor, or exposure am I considering?
C Comparison or Intervention (if appropriate) What is the main alternative to compare with the intervention?
O Outcome you would like to measure or achieve What can I hope to accomplish, measure, improve, or affect?
  • What Type of question are you asking?
    • Diagnosis, Etiology/Harm, Therapy, Prognosis, Prevention
  • Type of Study you want to find
    • What would be the best study design/methodology?

The P.I.C.O. Model /Specific questions

P I C O
Patient or Problem Intervention (a cause, prognostic factor, treatment, etc.) Comparison Intervention(if necessary) Outcomes
Tips for Building Starting with your patient, ask “How would I describe a group of patients similar to mine?” Balance precision with brevity. Ask “Which main intervention am I considering?”Be specific. Ask “What is the main alternative to compare with the intervention?”Again, be specific. Ask “What can I hope to accomplish?” or “What could this exposure really affect?”Again, be specific.
Example “In patients with heart failure from dilated cardiomyopathy who are in sinus rhythm …” “… would adding anticoagulation with warfarin to standard heart failure therapy …” “… when compared with standard therapy alone …” “… lead to lower mortality or morbidity from thromboembolism. Is this enough to be worth the increased risk of bleeding?”

Clinical Trial Designs

May be classified :

  • according to the method used to allocate participants into treatment or control groups (non-randomised or randomised controlled trials)
  • according to the awareness of either participants or researchers or both of which group participants are allocated into (single, double-blind, or triple-blind studies)
  • according to the magnitude of difference between treatment and control groups that is expected (superiority or non-inferiority trials)

Comparison clinical trial designs

There are a number of different types of comparison trials possible:

  • Superiority to demonstrate that the investigational medicine is better than the control;
  • Equivalence to demonstrate that the endpoint measure is similar (no worse, no better) than the control;
  • Non-inferiority to demonstrate that the investigational medicine is not worse than the control;
  • Dose-response relationship trials to determine various dose parameters, including starting dose and maximum dose.

Parallel trial

  • After screening, patients are randomised into separate treatment groups. They remain in these treatment arms for the duration of the trial, analysis, and follow-up activities.

Crossover trial

  • Patient X and Y are randomised into two different treatment arms. Patient X receives Treatment A during the first period of the study; Patient Y receives Treatment B. After the first period is over, there is a washout period. Patient X then receives Treatment B for the second period of the study while Patient Y receives Treatment A

Matched-pair trial

  • After screening, participants are matched into pairs. Within each pair, one participant is randomised onto Treatment A while the other is randomised onto Treatment B.

Cluster sampling

  • Randomised trials can also use cluster sampling.
  • In cluster sampling, suitable geographical areas are found (for instance, city, region, etc.).
  • A number of these geographical areas are then randomly chosen.
  • For each of these chosen geographical areas, a proportionate subsample from the members of the study sample in that area are chosen, and these subsamples are then combined into a sample group.

Withdrawal trial

  • During a withdrawal trial, after the first specified period of time has elapsed, participants are randomised into two groups, one of which receives a placebo instead of continuing to receive the active treatment.

Factorial design

Factorial clinical trials test the effect of more than one treatment. This allows assessment of potential interactions among the treatments.

Effect Size, Power, Significance, and Sample Size

The relationship between the null hypothesis (the premise of the study that the treatments have equipoise) and the true effects of the treatments. Two statistical errors are identified: a type I (or \(\alpha\)-error) and a type II (or \(\beta\)-error).

Observation Reality
Treatment has no effect Treatment has an effect
Treatment is effective
H0 rejected 		Type I or \(\alpha\)-error Correct conclusion
Treatment has no effect
H0 accepted 		Correct conclusion Type II or \(\beta\)-error

Inorder to determine sample size (N) for study we have to decide:

  • Type I or \(\alpha\)-error (\(\alpha\)) or significan
  • Type II or \(\beta\)-error (\(\beta\)) eli power is (\(1-\beta\))
  • Effect size (\(\Delta\))
    • Clinically meaningful effect size is an expert decision, not statistical decision

Factors that affect sample size calculations

Factor Magnitude Impact on identification of effect Required sample size
P value Small Stringent criterion; difficult to achieve ‘significance’ Large ↑
Large Relaxed criterion; ‘significance’ easier to attain Small ↓
Power Low Identification unlikely Small ↓
High Identification more probable Large ↑
Effect Small Difficult to identify Large ↑
Large Easy to identify Small ↓

Sample size calculations in practice

  • There are plenty of web-based calculator such as:
  • In order to use them the following are needed to know
    • Type I or \(\alpha\)-error (\(\alpha\)) or significance
    • Type II or \(\beta\)-error (\(\beta\)) eli power is (\(1-\beta\))
    • Effect size (\(\Delta\))
  • Effect size (\(\Delta\)) is difference:
    • between means of two groups (continuous outcome, also standard deviation (SD) is needed )
    • between to proportions (dicotomic outcome, also base rate is needed)

Sample size calculations in practice/ Two means

Comparing two means

Comparing two means

Sample size calculations in practice/ Two proportions/ Formulas

\[ H_0:\mu_A-\mu_B=0~~~~~~~~~~~~H_1:\mu_A-\mu_B\neq0 \] \(\kappa=\frac{n_A}{n_B}\) \[ n_A=\kappa n_B \;\text{ and }\; n_B=\left(\frac{p_A(1-p_A)}{\kappa}+p_B(1-p_B)\right) \left(\frac{z_{1-\alpha/2}+z_{1-\beta}}{p_A-p_B}\right)^2 \] \[ 1-\beta= \Phi\left(z-z_{1-\alpha/2}\right)+\Phi\left(-z-z_{1-\alpha/2}\right) \quad ,\quad z=\frac{p_A-p_B}{\sqrt{\frac{p_A(1-p_A)}{n_A}+\frac{p_B(1-p_B)}{n_B}}} \] where,\(\kappa=\frac{n_A}{n_B}\) is the matching ratio; \(\sigma\) is standard deviation;\(\Phi\) is the standard Normal distribution function; \(\Phi^{-1}\)is the standard Normal quantile function;\(\alpha\) is Type I error;\(\beta\) is Type II error, meaning \(1-\beta\) is power

Sample size calculations in practice/ Two proportions

Comparing two means

Comparing two means

References

Licence