How to conduct a clinical trial

Reviewed and revised 26 August 2015

OVERVIEW

• This page provides a broad overview of the complicated process of conducting a clinical trial
• Conducting a clinical trial can be conceptualised as having 14 key steps

STEPS

(1) Research question

• is it worth while?
• could it help improve patient outcome?

(2) Define hypothesis

• take an opinion on what the answer to the question is.

(3) Literature review

• make sure no one has already conclusively answered the question
• make sure it’ll benefit patients

(4) Involve statistician, and discuss:

• type of study
• calculate sample size (alpha and beta values, power, effect size , variance of population)
• how to reduce bias:
  • -> blinding (single, double or triple)
  • -> crossover
  • -> randomization
  • -> controls

(5) Develop inclusion & exclusion criteria

• stick to them when carrying out study!

(6) Calculate sample size needed

• beta value and power (small sample size is the commonest cause of a false negative result)
• alpha value (e.g. p < 0.05; limits false positive results)
• variance of population
• effect size
• statistical test

(7) Develop methodology

• treatment
• doses
• side effects
• placebo
• how to collect information

(8) Gain ethical approval

• consider professional virtue of medical loyalty
• autonomy, non-maleficence, beneficence & distributive justice.

Some issues include

• use of placebos when efficacious therapy available
• whether every research participant needs to sign a written consent
• whether informed consent is necessary (ICU patients)
• accepting drug company gifts
• misleading information from the drug reps.

(9) Perform pilot study

• modify protocol based on pilot study

(10) Carry out study

• patient information & informed consent important
• collect data
• measure outcomes
• monitor safety

(11) Analysis

• using appropriate statistical methods

(12) Compare to null hypothesis

• proven or disproven
• statistically significant difference or not

(13) Consider errors

• sources of bias
• patients lost to follow up
• adverse effects
• surprising benefits

(14) Submit for publication

• survive the peer review process and hope that the article is accepted!
• educate others of the findings so it can be implemented into clinical practice (knowledge translation)

References and Links

LITFL

• CCC — How to conduct a clinical trial
• CCC — Early stopping of clinical trial
• CCC — How to analyse a clinical trial