Statistics

Overview The reporting of medical findings originally focussed on the description of individual patients with an unusual presentation. This would be a novel or rarely used method of treatment, or an unexpected outcome. Some reports described several such cases treated by the same practitioner. The 20th century saw the development of larger scale studies that involved (for example) the collection of the views of individual patients, the inspection of patient medical records, following patients over time using standardised record keeping, or performing comparisons of patients receiving different treatments. Recent years have seen the development of medical literature databases on the Internet, facilitating

Rationale for sampling A population is a complete group of individuals such as all the residents of a country. Involving the whole population is unrealistic in terms of cost and staff time. Sampling is preferable and often the only way forward. Individuals who are reasonably easy to contact are selected making the study viable. Remember that due to random fluctuation information gained from different samples will not be identical even with the same method of selection. Also, some of the individuals selected will not be willing and/ or able to be involved leading to non-response bias. Random sampling A sampling

Investigating a population Deductions about the characteristics of a population can be made using a representative sample drawn from the population. This procedure starts with a statement regarding the population and information from the sample is used to decide whether or not there is enough evidence to conclude that this statement might be false. The null hypothesis The initial statement about the population (the null hypothesis) is very specific. For the case of a single sample it needs to describe the population, introduce the characteristic under study, include a statistical measure of interest (e.g. the mean) and propose an assumed value for this measure.

Introduction Diagnostic testing is based on the phenomenon that components of blood and other body fluids are found at different concentrations in individuals with a particular disease relative to people without the disease. For instance, fasting blood glucose concentrations are raised in those who suffer from diabetes. Blood glucose concentration can therefore be used as a method for identifying undetected cases of diabetes in the population. In men, prostate specific antigen (PSA) levels in blood are higher on average in those who have prostate cancer and can be used to detect and manage progression of the disease. In these notes,

Summary Data can be binary, nominal, ordered, discrete quantitative, or continuous quantitative. Graphs include bar-charts, pie-charts, histograms, box-and-whisker plots and scatter diagrams. The type of graph used should be chosen according to the type of data being displayed. Be aware that graphs may be presented in misleading ways. Furthermore, graphs can be symmetrical, positively skewed or negatively skewed. Types of average include the mean, median, and mode. Measures of spread include the range, interquartile range, variance, and standard deviation. Means and measures of spread should be chosen according to the type of data being summarised and some measures are susceptible