- 1 A
- 2 B
- 3 C
- 4 D
- 5 E
- 6 F
- 7 I
- 8 J
- 9 K
- 10 L
- 11 M
- 12 O
- 13 P
- 14 Q
- 15 R
- 16 S
- 17 T
- 18 U
- 19 W
- 20 References
agent based model
Agent-based modeling is a special individuals-based method of computer modeling and simulation, closely linked with complex systems, multi-agent systems, evolutionary programming, cellular automata. The modeling is characterized mainly by the possibility of connections between the micro and the macro-level to model explicitly.
This type of modeling is particularly apply if the focus of a question is not the stability of an equilibrium or the assumption that a process will returns in an equilibrium, but the question of how a system of changing conditions can adapt.
The main difference to the standard form of discrete event simulation is that in an agent-model entities (here: "agents") each other and perceive their environment. This technique is ideal for the modeling of infectious diseases such as where the transmission of infections is important.
Also called systematical error. Unlike uncertainty a bias effects the result by an error made during collection or interpretation of a data set. E.g. the sample doesn't represent the entire population.
Occurs when a study isn't double blinded and e.g. knowledge about a test, medication or else, influences a person in his approach to patients (e.g. a doctor treats a placebo-patient more thoroughly), medications or staff.
Work-up bias-Verfication bias
Occurs when patients are selected for a study by criteria, which were manifested in the results of the test/study itself.
A confidence interval gives an estimated range of values which is likely to include an unknown population parameter, the estimated range being calculated from a given set of sample data. Confidence intervals are usually calculated so that the percentage is 95%, but we can produce any confidence intervals for the unknown parameter. The width of the confidence interval gives us some idea about how uncertain we are about the unknown parameter. A very wide data interval may indicate that more data should be collected before anything very definite can be said about the parameter.
More about confidence intervals.
Cost-benefit analysis weighs the positive benefits against the treatment costs to evaluate if the treatment is feasible. The analyses gives information whether a treatment is profitable with it's costs or not.
Cost-utility analysis is an advancement of the cost-effectiveness analysis, however the outcome differs. The outcome is based on quality-adjusted lifeyears(QALYs). For evaluation, every year of life is allocated with life quality of corresponding value.
cost effectiveness analysis
Cost effectiveness analysis usses monetary units as input data and outcome is valued in clinical parameters, e.g. blood pressure or life expectancy. The analysis focusses on answering the question of how to spend a budget most efficiently. Costs and health improvement are taken into account during the analysis.
cost minimisation analysis
Cost minimisation analysis focusses on finding the cheapest way to a definite result using cost comparison. The analysis can only be utilized if treatment goals are met and comparable.
decision analytic techniques
Generally every technique used to reach a conclusion based upon data.
May include techniques like: decision trees, regression models, epidemiology models etc.
+ decision tree method
Decision tree analysis is very suitable for simple decision problems, where events with limited recursions are analyzed over a specific period of time.
It is a visual representation of all possible options and consequences which can occur to an investigated technology. Each of the studied interventions follow ramifications, the possible events with their associated probabilities of occurrence. Here can not only the probabilities of the different strategies, but also by patient characteristics depend. At the end of each path of the decision tree is an outcome. For each of the alternative interventions can be the expectation value of the clinical and economic consequences as a weighted average of all possible consequences are calculated, taking into account the chance of paths are used as weights.
Delphi panel or delphi method is an instrument to gain information about future trends, events, technical improvements, etc...
During the first phase of a delphi panel a group of experts is asked several questions. During the second phase the collected answers are once again presented to the experts to be commented and rated by them. This process is repeated several times during which a widely accepted opinion is obtained.
In this context, discounting values (monetary or non-monetary) describes a method of valuing a therapy using the concepts of the time value of money. Future cash flows are estimated and discounted to give their present values, as using future values for financing is more profitable than financing it immediately. Being able to invest the present funds, the profit made in the time between using and paying the therapy can be subtracted from the actual costs. Time can be translated into a monetary value.
In Germany the "Hannoveraner Konsens" specified the discount rate to 5%.
discrete event simulation
An event-oriented simulation model can clearly be described by a state model, an event calendar (or an event list), event routines, and time. The event calendar contains a list of future events with the name and nature of the event and the date of its occurrence. You can turn certain events happening in the future trigger. Thus, complex behavior can also be simulated. In the technical program implementation, the event list after time and always sorted the next occurring event processed.
The strength of the discrete simulation is that the chance or probability in the model with inclusive enough and frequent recalculation a statement about the expected probability of various system states returns.
The study should concern itself with a topic that is economically relevant or will be in the future.
Treatment is effective in routine under normal care conditions.
Treatment is effective under optimal conditions like clinical trials. Therefore results obtained under optimal conditions do usually not represent the every day routine in health care facilities. To assess the every day routine in treatment of a certain disease, the guidelines compiled by the Association of the Scientific Medical Societies (Arbeitsgemeinschaft der Wissenschaftlichen Medizinischen Fachgesellschaften - AWMF) should be used.
An epidemiological model is a mathematical model, which may be a computer simulation model of a disease for the purpose of studying the behavior of the disease in a variable animal population unter variable conditions of climate, density of population, mix of population, and so on. It may be an analytical model, an economic decision making model, an explanatory model or a predictive model. I It may also be a causal model, which allows the operator to vary the determinants of prevalence and observe the respective outcomes. It may permit only the use of fixed numbers so that it will always return the same answer to the same question, in which case it is a deterministic model, or it may introduce the element of chance into the selection of outcomes, in which case it is a stochastic model.
The follow-up study is settled a short time after the original study, judging long-term effects of therapies or gathering additional data.
Funnel plots help to assess whether there is publication bias or not. It consists of a scatter plot of treatment effect on the x-axis against the size of the study on the y-axis. If the plot turns out to be of an asymmetric form this might be an indicator of publication bias. If there is no bias at all the graph should resemble a symmetrical inverted funnel (see picture below).
For more information see the Cochrane Handbook for Systematic Reviews 
intention to treat
During the execution of a clinical trial, patients might drop out of the study or switch therapy arms. There are several possibilities to cope with these violations against study protocol.
Intention to treat (ITT) analysis: A patient might switch his/her therapy but his/her outcomes are counted among the therapy arm s/he was randomized at.
As Treated (AT) analysis: Outcomes of a patient are counted among the therapy s/he was treated last.
Per protocol (PP) analysis: Only outcomes of patients who were treated after study protocol are assessed. Drop-outs and patients who switched therapies are excluded.
The Jadad score is a rather easy method to judge the quality of a publication.
The score is obtained from three questions:
1. Was the study described as randomized (this includes the use of words such as randomly, random, and randomization)?
2. Was the study described as double-blind?
3. Was there a description of withdrawals and drop-outs?
Add a score of ‘1’ for each positive answer. Add a score of ‘0’ for each negative answer.
Add a point of ‘1’ if the authors described how they generated the sequence of randomization and if this method was appropriate.
Add a point of ‘1’ the method of blinding was described and appropriate.
Deduct a point of ‘1’ if the method of generating the sequence of randomization was described but inappropriate.
Deduct a point of ‘1’ if the study was described as double-blind but the method of blinding was inappropriate.
The obtained score should be larger than 3 and containing the score points for randomization and blinding.
key potential decision makers
These are organisations or persons that have major influence over the health care system in Germany, e.g. health insurance companies or politicians.
A league table is some sort of chart or list which compares several treatments or effects to judge which one is ranked higher or lower.
Markov models are powerful and easy-to-use tools for modeling of prevention, diagnosis and treatment of chronic diseases in which
- the parameters are time-dependent
- the date of the event plays a role, or
- events can occur repeatedly
Markov models represent the reality in the form of a sequence of individual states from which the relevant health states of patients reflect.
Meta-analyses sum up several studies which are comparable in study question and setting. This kind of analysis should be executed with extreme care. Transparency and traceability are indispensable. Meta-analyses provide the highest level of evidence and they will prefer in any economic evaluation.
The result of a meta-analysis is often illustrated with a forest plot. Therefore the logo of the Cochrane Collaboration is a stylized forest plot.
The small diamond represents the result of all studies regarded in the given meta-analysis.
meaningful and robust
The statistical power represents how meaningful the performed test is.
A statistic is called robust when a biases is avoided by the used estimation. E.g. using median instead of arithmedic mean.
More about robust statistics.
Monte-Carlo-simulation is a method of the stochastic, in which very often carried out random experiments represent the base. It is based on the results of attempts, with the help of probability theory is not analytically solvable, or only complex problems in the mathematical context to be solved numerically.
More about Monte-Carlo-Simulation or on Wikipedia.
Opportunity cost represents the difference between the best treatment and the next best treatment. The difference can be measured in monetary terms, lost time, health status or other desirable units.
The power of a hypothesis test is the probability of not committing a type II error. It is calculated by substracting the probability of a type II error from 1, usually expressed as:
Power = 1 - P(type II error) = (1-beta) 
The maximum power a test can have is 1, the minimum is 0. Ideally we want a test to have high power, close to 1.
More about statistical power.
"The quality adjusted life years” are an instrument of the economic evaluation for comparing costs of procedures and technologies in the health service regarding their results. A QALY is equivalent to one year spent in perfect health. Goal is to measure an improvement on the quality of life or a lengthening of life.
see league table
The research question is the topic the study is about.
It is essential to present information which answers the research question, regardless of the outcome.
See also research question
There are different kinds of resources occur in the health care system. They are strongly related to a treatment of a certain disease. Depending on the point of view resources can be allocated to patients, treatments or even hospitals.
There are a many statistical tests, unfortunately most of them are very specific, so it is important to choose the right one applicable for the data set. Normally a test statistic is ascertained based on the given data. Dependent on the desired sensitivity the test statistic is in the confidence interval or not. More about statistical tests.
Sensitivity analysis is the study of how the variation (uncertainty) in the output of a mathematical model can be apportioned, qualitatively or quantitatively, to different sources of variation in the input of a model.
The most common methods used for a sensitivity analysis are:
- simple sensitivity analysis (one-way or multi-way)
- threshold analysis
- probalistic sensitivity analysis ( Monte Carlo Sensitivity analysis)
More about sensitivity analysis.
The affected person (for example a patient) has the choice to either remain in the current state with a certain security in the current state of the encroachment, or to undergo a measure (e.g., operation) by which he can reach perfect health with a chance not to survive.
Transmission models are a widespread and accepted technique in the modeling of infectious diseases in the areas of Epidemiology and Public Health. They can be constructing with different degrees of complexity and can be deterministic or stochastic. An example can be the effects of vaccination, which the susceptibility of the population and possibly the strength of the risk of infection as well as other factors can change, thereby both terms of direct effects of vaccination as well as its indirect effects.
time trade off
The affected person (for example a patient) has the choice to either live the rest of his life with the current illness or to deal with a shortened life expectancy with the possibility of perfect health.
Uncertainty is one of the two main reasons for variability, because used parameters are acertained out of a sample. The parameters are well known for the sample but not for the whole population. Simply said, uncertainty is the lack of that knowledge. By contrast bias.
willingness to pay
In the economy the willingness to pay, the maximum amount a person would be willing to pay in order to receive a better state, either for a certain period or for good.