1&2: Intro to biomech and modeling Flashcards
(14 cards)
What is biomechanics and what challenges does it face?
Biomechanics is the effect of forces in a licing body. It can be biomechanical-biomechanical interference or it can be impact of non-physiological loading, for example: hip implant against the bones and joints. Biomechanics are systems with different length scales. Molecular-collagen, cellular-endothelial cells, tissue-vascular tissue, organ-aorta, organimshuman.
Challenges
• In biomechanics, engineering methods need further development. Adapt properties of vascular tissue to mechanical and biomechanical environment. Need of couple models among structural, fluid, chemical and electric fields.
• It is often hard to validate models because of ethical and other constraints.
• For clinicians it is hard to integrate engineering solutions to clinical work flow.
• Limited market for the engineering solution. Threat to present medical companies
and work of clinicians.
• Engineering dominated decision makers are not aware of the clinical problems and clinical dominated decision makers are not familiar with engineering dimension
of the development.
• Legal problems, might fail to receive medical device approvement.
How do engineers and clinicians solve problems?
Engineers use powerful methods and they are principle dominated. They use a prospective and retrospective, meaning they try both untested things and the old and tried. It is a detailed and deterministic approach. Clinicians use retrospective, statistical approach.
What is a random variable?
A variable whose possible values are outcomes of a random phenomenon.
What is correlation amongst variables?
Correlation amongst variables refers to how close two variables are to having a linear relationship with each other. A correlation test explores whether or not a correlation between variables exist and provide a correlation coefficient -1 = r_1. A positive coefficient denotes a positive correlation, and a negative one a negative correlation. A positive correlation means that the dependent variable increases at increased independent variable. Negative the opposite. The large the absolute value of r is the stronger the correlation. Testing dependence of a single variable with respect to a single independent variable is called simple correlation testing.
What is regression modeling?
Regression modeling is used to predict dependent variables based on the values of independent variables, or a cause and effect relation between them. Predicting one dependent variable from one independent variable is called simple regression modeling.
What is a mean difference test?
A mean difference test can be used in order to explore whether or not two samples stem from different populations. Appropriate when the population is normal distributed and the sample size is 15 or more, and contains no outliers.
What is double blinding?
Neither study objects or analyst receive information about the treatment. For example the study is done by using two groups, one which get the real treatment and one which
get a placebo treatment. The researcher and the study units are both unaware of which treatment the study units are getting.
What denotes randomization and replication in study design?
Randomization is using chance methods to assign study units to treatment groups, which in turn randomizes the influence of extraneous variables. Replication is to assign
each treatment to many study units. The more study units in each treatment condition, the lower the variability of the dependent measures.
What is a well-designed study?
A well-designed study provides a clean test of causal connections between dependent and independent variables, i.e. eliminates the influence of extraneous variables. Steps to reduce the effects of extraneous variables are important for well-designed studies. Such steps comprimise: randomization, replication and control - eliminating the unfluence of extraneous variables by making a study as similar as possible for the study units.
What is a confounded study?
A study is confounded if it is influenced of extraneous variables. A confounded study provides plausible alternative explanations for the observed relationship between independent and dependent variables.
What is a model?
A model represents a real object to some degree of completeness. A model should be as simple as possible but not simpler.
Itemize typical model simplifications applied in biomechanical modeling.
- Geometry
- Tissue properties
- Boundary conditions
- Loading conditions and initial conditions
- Patient characteristics
- Biology
Explain a typical model development cycle.
see fig model_cycle
The real object is simplified by the use of an analytical and numerical model. The model is then validated and verified by testing known variables and see if it solves them correctly.
What is a model sensitivity analysis?
How sensitive the model predictions is to input data variations. Identify important connections between observations, model inputs and predictions or forecasts. Testing robustness.
