Basic Statistics in Clinical Research Slides created from article by Augustine Onyeaghala (MSc, PhD, PGDQA, PGDCR, MSQA, FMLSCN) and available on Global Health Trials (www.globalhealthtrials.org) Dr. Augustine Onyeaghala Biomedical Scientist/ Clinical Research Consultant 15/7/2015, Ibadan, Nigeria www.theglobalhealthnetwork.org What is statistics? the collection, arrangement, analysis, interpretation and reporting of data So….. Statistics is just something that a statistician does, in a trial, and it happens just after we’ve finished collecting the data, right? Wrong! Knowledge of statistics is important right through the project: – – – – – – – – – – – – Sample size Blinding Randomisation procedure Inclusion/exclusion criteria Outcomes Type of statistical test used Interpretation of data Type of control group Data management Missing data Confounding factors Data safety monitoring Problems seen in trials, which could result in defective statistical analysis: Poor study design Inadequate recruitment Missing data Adherence to inclusion/ex clusion Confounding factors Bias Wrong application of statistics Case study: Remedy X Hypothesis • • Hypothesis: Local Remedy X has an effect on plasma glucose levels Null hypothesis: Local Remedy X has no effect on plasma glucose levels Null Hypothesis Alternative Trial designs: • The investigational drug (new drug) is betterhypothesis than or superior to the standard, control drug or placebo (superiority trial design) • New drugs perform as good as the standard treatment (equivalence trial design ) • New drug is less effective than the standard treatment (inferiority trial design) P values • P = probability value shows us whether the difference observed is just due to chance, or if it’s statistically significant. • If P>0.05, accept the null hypothesis (i.e. there IS NOT a statistically significant difference) • If P<0.05, reject the null hypothesis (i.e. there IS a statistically significant difference) BUT…… • Results are always probabilistic – you have never proved either hypothesis, simply indicated that the probability that the null in true is lower than your critical value so that you can reject the null, and accept the alternative as the most probable explanation. Bias • This is an error associated with the study design, conduct, analysis and publication that exaggerates or underestimates the effectiveness of the investigational product. Randomisation and blinding Randomisation and blinding are important ways of minimising types of bias. Variables: Confounding Variables These are factors that are not normally measured during the study, but may be accountable for the effects observed in research. Distribution of data Bell curve = normal distribution Non-parametric • Questionnaire answers are a good example of data that is not normally distributed Why does distribution matter? Expectation of your data distribution impacts the statistical tests you could use to analyse the data. Parametric tests include: t-test What if my data is not normally distributed? • Some data might not be normally distributed: Takeaway message: Researchers should decide and choose at the planning stage the type of statistical technique which should be applied to enable them arrive at a good conclusion. Types of Error the incorrect rejection of a null hypothesis which is actually true (a "false positive") the failure to reject a false null hypothesis (a "false negative"). Study Design is important for statistics! • Design: visit Equator-Network.org for free, useful advice on study planning and analysis • Patient selection (appropriate sample size, sampling, inclusion/exclusion criteria etc) • Define the measurable outcome and look out for it. • State the null hypothesis clearly, and avoid type1 and type 2 errors. • Ensure randomization and blinding. • Use the right and adequate control group. • Avoid confounding factors • Apply the right statistical calculations Some free eLearning courses in statistical analysis: • Online Statistics Education: an Interactive Multimedia Course (onlinestatbook.com) • Biostatistics Lecture Series from John Hopkins (ocw.jhsph.edu) • Essentials of Probability and Statistical Inference from John Hopkins (ocw.jhsph.edu) • Introduction to Biostatistics from John Hopkins (ocw.jhsph.edu) • Basic Biostatistics Concepts and Tools (Robert Stempel College of public health and social work) Acknowledgements: Original article written by Augustine Onyeaghala (MSc, PhD, PGDQA, PGDCR, MSQA, FMLSCN) and available on Global Health Trials (www.globalhealthtrials.org) With thanks for slide preparation to Brigid Davidson Funding The Bill and Melinda Gates Foundation