Sample size calculations Sample size calculations need to be done before data collectionSample size calculations are done in a statistics packageFactors that influence sample size calculation:Expected effect size - if expected effect size is ↑, can have ↓ sample size to detect differenceExpected variability (SD) of data - look at previous trials for approximationSignificance level … Continue reading Statistics in Clinical Trials/Metanalysis

Observational studies include: cohort; case-control; and cross sectional studiesThey are non-randomised & non-experimental Cohort studyProspective Used to determine aetiology & natural history of diseaseDefine cohort --> assess risk factors to follow --> follow forward in timeUsed to calculate risk ratioRisk ratio indicates increased/decreased risk of disease associated with factor of interest AdvantagesDisadvantagesTime sequence can be reviewedCostly Provides … Continue reading Observational Studies/Epidemiology

Clinical Trials Interventional/experimental trialsUsually include randomisationThey are prospective (individuals are followed forward from some point in time)Causal analysis – test causal relationship between exposure of interest & outcome Phases of clinical trials Translational pathway of clinical trials:  Pre-clinical trials:Before testing in humansHelps decide whether the drug is ready for clinical trials (from ‘bench to bedside’)Look … Continue reading Clinical Trials/Ethics

Time to event (TTE) dataAnalysis of data from a point in time until a particular event. In many oncological studies event is death or event is disease progression.It is not normally distributedIt is continuous data Survival data/analysis is a type of TTE data/analysis Data may often be censoredRight censored – patients who have not reached … Continue reading Survival analysis

Statistical inference is the process of hypothesis testing and using data to make conclusions about characteristics of populations.  Hypothesis testing Null hypothesis (H0) – assumes no effect in the populationAlternative hypothesis (H1) – if the null hypothesis is not true Alternative hypotheses can be:One sided – state the direction of effectTwo sided – do not … Continue reading Statistical Inference

Bias Occurs when there is a systematic difference between the results obtained from a study and the true population parameter – it can over- or under- estimate the true effect and thus create an incorrect association or conceal a real oneBias can occur at any stage of the research process from planning to publicationTypes of … Continue reading Bias and Confounding

Types of Data Types of data include:Non-numerical/qualitative/categoricalWhen a variable/observation can only belong to a distinct categoryIncludes nominal & ordinal NominalUnordered & mutually exclusive categories. Not possible to rank Examples – alive/dead, blood groupOrdinalOrdered  & mutually exclusive categories. Can be ranked. Difference or ‘gap’ between values can be ill-defined, Examples – mild/moderate/severe Numerical/quantitativeWhen a variable/observation takes … Continue reading Types of Data and summarising Data

Normal/gaussian distributionsThe mean is the peak of curve – it is symmetrical around its meanThe standard deviation determines width of curve (the larger the SD, the wider the curve)Normal distribution tends towards infinity (i.e. the line never reaches the axis)Reference range for a sample = mean +/- 2 standard deviation Non-normal distribution/skewness: NEGATIVE SKEW or … Continue reading Normal and non-normal distributions

Simple calculations based on SSD, PDD and TMR SSD - the distance between the source and the surface of the patient PDD - percentage depth dose. The percentage of dose at a given point compared to a reference of 100% dose at Dmax TMR - Tissue maximum ratio. The ratio of dose at 2 different … Continue reading Simple calculations

IMAGING DURING TREATMENT CT - linear attenuation coefficient for each pixel is compared to water and gives us Hounsefield number  CT numbers represent Electron density of material  Compton attenuation is proportional to Electron density  MRI better for soft tissue contrast. But no electron density information. Prone to geometric uncertainty.   PET : F18 is a positron emitter  … Continue reading Imaging during Radiotherapy and Quality control