Ukázka elektronické knihy, UID: KOS261987
11
BIOSTATISTICS
11.1 DEFINITION OF STATISTICS, METHOD AND CONTENT Biostatistics is the science that studies random mass phenomena, their quantitative features with the connection of qualitative property. It studies how to collect, organise, analyse and interpret numerical data. Probabilism is one of the aspects of statistical thinking especially in medicine. Probability is the basic tool in study and application of statistical methods. Statistics is divided in descriptive (older) and inductive (younger) statistics. The task of descriptive statistics is to organise and summarise numerical information. Inductive statistics analyses data. The basic method of statistics is the probability theory. The principal content of statistical data is analysis (judgement of various factors), decision (to decide which method is the most favourable) and prediction (success of treatment probability).
11.2 PRINCIPLES OF STATISTICAL ANALYSIS In the case of any investigation or research it is necessary to make the design and prepare the plan at first. The hypothesis about study problem is formulated. Next steps are definition of study population, choosing good method of random sampling and the exact criteria of the collection of the data. Mean value, standard deviation and probability can be calculated after the control of study data. For hypothesis testing it is necessary to choose adequate statistical test. The conclusion of an investigation or research is better made in co-operation with a statistician and a medical professional.
11.3 BASIC TERMS Statistical unit should be defined objectively, locally and in time. It means who or what, when and where is a study made. The entire population means all measurements (e.g. all patients, all observations) of our interest according to the definition of statistical unit. The sample is 99
Ukázka elektronické knihy, UID: KOS261987
simply a representative part of the whole population. It serves for estimation of entire population data. Random sample is determined completely by chance. Variables are the properties of each unit. Common variables identify study unit objectively, locally and in time (e.g. people over 60 years old in the Czech Republic, in year 2018). Investigated (observed) variables are the ones in the focus of our interest (e.g. diagnosis, smoking, drug abuse). In general, variables can be divided in quantitative (measurable) and qualitative properties. Quantitative continuous variables are expressed with the decimal point, e.g. weight 70.5 kg, level of glycaemia 5.3 mmol/l. Discrete variables also called discontinuous are described by integers, e.g. 20 cigarettes per day, number of beds, sick people, patients and physicians. Qualitative data are verbally expressed in alternative ways (2 variants, e.g. sex, smoker or non-smoker). Multivalued qualitative data have more possible values (e.g. education, profession, diagnosis of disease, marital status). The sample is a representative copy of the entire population thanks to random sampling procedure. Each individual is chosen randomly (by chance). Each unit (individual) has an equal chance to be selected into the sample (to be a part of the sample). The result of this procedure is a representative sample. In the sample there is made the incomplete investigation. The basic types of random sampling are sample method (e.g. by a lot, table of random numbers) and systematic methods (e.g. day of birth, surname). Stratified random sampling is also called proportional random sampling or quota random sampling. Stratified random sampling involves dividing the entire population into homogeneous groups. They are called strata. Random samples are selected from each stratum, e.g. for a study of incidence of upper respiratory disease in the Czech Republic it is necessary to use the regions as the strata for the study of this disease. Cluster method is used in the case of a very large population for study of selected group, e.g. study of a family- in the Czech Republic is randomly select one region, one town, one street, one house and finally one sample family. Non random sampling is also called judgemental, paired sample or matching. There are for example two groups smoker and non smoker, which correspond with the common variables (sex, age, profession, job, education, living place, case history and other important factors) and they differ in the study of risk factors for the disease, e.g. smoking risk for lung cancer. Probability theory is a branch of mathematics concerned with the analysis of random phenomena. The outcome of a random event cannot be determined before it occurs, but it may be any one of several possible outcomes. The actual outcome is considered to be determined by chance (in Encyclopaedia Britannica). Probability theory is the mathematical study of phenomena characterized by randomness or uncertainty. More precisely, probability is used for modelling situations when the result of an experiment, realized under the same circumstances, produces different results (typically throwing a dice or a coin). Mathematicians and actuaries think of probabilities as numbers in the closed interval from 0 to 1 assigned to “events” whose occurrence or failure to occur is random (in Science Daily). The probability of an event A is PA = m/n, m is occurrence of an event A and n means all possibilities (total number of outcomes). PA is relative frequency of event A and is valued 100
Ukázka elektronické knihy