BMI Course Information
BMI 500 Research Methods in Biomedical Informatics (3)
Instructor Information |
|
Fall 2009 |
Course not offered |
Spring 2009 |
|
Catalog Description
The course presents an introduction to the conduct of scientific research in biomedical informatics: hypothesis formulation, observation and measurement, design of experiments, execution of protocols, development of essential research skills, and sensitivity to ethical issues that arise in research.
Prerequisites
Admission to any SCI graduate program
Textbook and Other Materials
Critical Appraisal of Epidemiological Studies and Clinical Trials (3rd edition - paperback) by Mark Elwood. Oxford Medical Publications 2007 978-0-19-852955-2.
Course Learning Outcomes
The course covers both quantitative and qualitative approaches. For the quantitative approaches, it focuses on the evaluation of data from systematic reviews of the literature, epidemiological studies, and clinical trials. Topics include the foundation of science, including the basics of probability and statistics, the use of statistical methods for decision-making,, and the roles of data mining, Bayesian methods and meta-analysis in data evaluation. For the qualitative approaches, it focuses on the foundations of qualitative research and on design for data collection and analysis in both laboratory-based and naturalistic environments. Topics include the history of scientific methods and the role of qualitative research, the study of cognition at the individual and group level, representative sampling methods, the use of expert systems and expert consensus, and qualitative models of decision making. These approaches are tied together by an overall focus on the notions of a gold standard and the elements common to critical appraisal in both approaches. Some knowledge of basic statistics and health care domain may be helpful.
Major Topics and Time Covered
Foundations of science
- Logical Empiricism as the root of modern science
- The history of science in medicine: traditional and more modern views
- The role of biostatistics in science
- The role of cognition in science – the “Cognitive Revolution”
- Thought process as valid research data
- Cognitive roots of computer science.
- Hypothesis formulation
- Relationship of theory, frameworks and models in investigative studies
- The use of models to test hypotheses
The role of science in biomedical informatics
- The distinction between making decisions and drawing conclusions
- Statistical models as decision models in bioinformatics
- Decision making and reasoning models based on empirical evidence of human decision making vs. normative models of correct decisions.
- Relationship between evidence and practice: qualitative perspective
- Strengths and limitations of various approaches
Foundations of probability
- The origins of schools of probability
- Subjective versus Frequency approaches to probability
Traditional statistical inference
- T-tests
- Regression analysis
- Fitting multiple regression models
- Bayesian versus Frequency approaches to inference
- Statistical assessment of uncertainty in each approach
- Qualitative trend assessment
Time to event analysis
- Fitting survival curves to data
- Cox proportional hazard models
- Time–series analysis
Data mining
- The relationships among biostatistics, data modeling,
- Epidemiology, evidence-based medicine and decision analysis
- Distinguishing apriori inference from post-hoc inference
- Exploratory data analysis versus classical statistical inference
- Statistical and cognitive learning theories
Study designs: Traditional epidemiological and single subject
- Randomized clinical trials
- Cohort studies
- Case-control studies
- Cross-sectional studies
- Database studies
- Single subject designs: Rationale and theory
- Representative sampling methods
- Cross sectional and longitudinal studies
Study design in applied BMI research
- “Gold standards”
- Experts as a reference standard
- Evaluation metrics: laboratory-based and naturalistic studies
Qualitative and quasi-experimental design
- Differences between qualitative and quantitative approach Validation across experiments and different domains.
- Representative sampling
- Data modeling
- Studying cognition: individual vs. distributed, laboratory vs. naturalistic studies
Approaches to critical appraisals of published studies or databases
- Hierarchy of studies
- Assessing quality within a fixed level
- Issues of reliability and validity
Meta-analysis
- Models of meta-analysis
- Underlying statistical theory and comparative benefits of competing statistical approaches
Evidence-based medicine in Biomedical informatics
- Using epidemiological evidence in the clinical setting
- Use of decision support systems in evidence-based medicine
- Emerging new paradigms for evaluating evidence in decision making