Design & Analysis of Biomedical Sciences
3.1 LT Introduction to DABS 07/12/2020 – R. de Mutsert & S. Böhringer
Contact
- Any questions:
- Support: (Stefanie Bamborough)
4-week course
- Jan 15 2021: examination April 26 2021: retake examination
o No books allowed, Remindo calculator allowed
- Info: syllabus, materials & daily overviews on Brightspace
Four central themes
1. Information Literacy and Data Management
2. Study Designs
3. Statistical Analysis
4. Interpretation of Results
Syllabus & Brightspace
- Topic numbers Brightspace match the topic numbers syllabus and time schedule
- According to four themes; independent from time schedule
Computer practicals
- Start with 30-min recap live in Kaltura
- Followed by computer practical
o Break-out rooms / own time
o Teachers are present in Kaltura for questions
o Reports including syntax file need to be uploaded to Brightspace
o Make sure you have a working version of SPSS on your computer
- Two students can work on a report together
o Each of the two students submits the same report using Turnitin
- All reports have to be submitted by the day of the exam
Examination
- Midterm test on Brightspace before Chrismas holidays
- Test exam after Christmas holidays in Remindo
- January 12: Response lecture statistical analysis & Question hour test exam
Final grade
- Examination: 100% of grade
- Combination of multiple choice and (5-7) open questions
- Will be published if obligated working groups attended and obligated reports uploaded to BS
,3 - Introduction to the course
3.2 LT Recap Epidemiology 07/12/2020 – R. de Mutsert & S. Böhringer
Research question: exposure/determinant, outcome, domain (study population) & time/period
Methods of epidemiological studies
Measures of disease frequency: describe how often a disease or another health event occurs in a
population
- Prevalence: number of subjects having the disease at a time point / total number of subjects
in population
o Number of existing cases / total population
o No information on new cases
o Important for planning of care and policy
- Incidence: amount of new cases
o Risk/incidence proportion: number of subjects developing the disease over a time
period / total number of subjects followed over that time period
New cases over time period / total population at risk followed
o Incidence rate: number of subjects developing the disease / total time at risk for the
disease for all subjects followed (person-time)
New cases / total time followed (person-years)
Population data (mortality rate, birth rate)
Epidemiological study designs
- Cross-sectional study (diagnosis)
o Determine determinant & outcome at one moment > prevalent cases
- Follow-up/cohort study (prognosis) > association between exposure & occurrence of disease
o Study population with subjects exposed (index) and non-exposed (control)
o Followed during time
o Count number of disease/with the outcome
- Case-control study (etiology) > association between exposure & disease
o Selection of patients with disease (cases)
o Ask about their exposure
o Compare their history of exposure with that of control persons without the disease
Exposure Odds Ratio
- Randomised controlled trial (RCT) (therapy) > effect and safety of intervention
o Experiment
o Randomisation of treatment and placebo
Measures of effect
- Relative risk (RR)
o Risk ratio = risk exposed / risk unexposed
o Rate ratio = incidence rate exposed / incidence rate unexposed
- Risk difference (RD) = risk exposed – risk unexposed
- Number needed to treat (NNT) = 1 / RD
o Number of patients that need to be treated to prevent one disease/death
- Odds ratio = odds cases / odds control
o Odds = exposed / non-exposed
o Estimation of RR
,
, Describing statistics
Categorical variables: summarized by counts (%)
- Ordinal: categorical, with specific order
- Nominal: categorical, no order
Numerical/continuous variables: summarized by mean (SD), median (SD) & quantiles (25%, 75%)
- All (positive) values possible & all steps equivalent
- Mean & standard deviation > outliers can have large influence on mean & SD
o Center of data: median (50% quantile)
o Variation: interquartile range (Q75 – Q25)
Analysing data
Standard deviation (SD): variation of single measurements
-
Standard error of the mean (SE): variation of the mean
-
- What would be the variation of the mean for many repetition of the experiment
- With a larger sample size SE becomes smaller
Confidence interval: determine whether an observed mean is larger than a given value
- Mean ± 1.96 SE
o For N > 30 (small sample sized; calculation inaccurate)
o Whether you can reject null hypothesis or not
Null hypothesis: opposite of what we want to show
P-value: chance to observe the data itself / even extremer data set, under the assumption that the
null hypothesis is true
- Conclusion about data (does it fit with null hypothesis?), no conclusion about null hypothesis
- Power: chance to reject the null hypothesis when it is true
- α-level: significance level, normally 5% (95% confidence level)
Required knowledge about statistics: Variables (1,4,5), Confidence intervals (10,11), null/alternative
hypothesis (17,18), P-values, power, independent/paired t-test (20,21), Contingency tables/Chi-
square test (24), Correlation (26) & Linear regression (27)
Research (UDEC) protocol
- UDEC = Universitair Dier Experimenten Commissie
3.1 LT Introduction to DABS 07/12/2020 – R. de Mutsert & S. Böhringer
Contact
- Any questions:
- Support: (Stefanie Bamborough)
4-week course
- Jan 15 2021: examination April 26 2021: retake examination
o No books allowed, Remindo calculator allowed
- Info: syllabus, materials & daily overviews on Brightspace
Four central themes
1. Information Literacy and Data Management
2. Study Designs
3. Statistical Analysis
4. Interpretation of Results
Syllabus & Brightspace
- Topic numbers Brightspace match the topic numbers syllabus and time schedule
- According to four themes; independent from time schedule
Computer practicals
- Start with 30-min recap live in Kaltura
- Followed by computer practical
o Break-out rooms / own time
o Teachers are present in Kaltura for questions
o Reports including syntax file need to be uploaded to Brightspace
o Make sure you have a working version of SPSS on your computer
- Two students can work on a report together
o Each of the two students submits the same report using Turnitin
- All reports have to be submitted by the day of the exam
Examination
- Midterm test on Brightspace before Chrismas holidays
- Test exam after Christmas holidays in Remindo
- January 12: Response lecture statistical analysis & Question hour test exam
Final grade
- Examination: 100% of grade
- Combination of multiple choice and (5-7) open questions
- Will be published if obligated working groups attended and obligated reports uploaded to BS
,3 - Introduction to the course
3.2 LT Recap Epidemiology 07/12/2020 – R. de Mutsert & S. Böhringer
Research question: exposure/determinant, outcome, domain (study population) & time/period
Methods of epidemiological studies
Measures of disease frequency: describe how often a disease or another health event occurs in a
population
- Prevalence: number of subjects having the disease at a time point / total number of subjects
in population
o Number of existing cases / total population
o No information on new cases
o Important for planning of care and policy
- Incidence: amount of new cases
o Risk/incidence proportion: number of subjects developing the disease over a time
period / total number of subjects followed over that time period
New cases over time period / total population at risk followed
o Incidence rate: number of subjects developing the disease / total time at risk for the
disease for all subjects followed (person-time)
New cases / total time followed (person-years)
Population data (mortality rate, birth rate)
Epidemiological study designs
- Cross-sectional study (diagnosis)
o Determine determinant & outcome at one moment > prevalent cases
- Follow-up/cohort study (prognosis) > association between exposure & occurrence of disease
o Study population with subjects exposed (index) and non-exposed (control)
o Followed during time
o Count number of disease/with the outcome
- Case-control study (etiology) > association between exposure & disease
o Selection of patients with disease (cases)
o Ask about their exposure
o Compare their history of exposure with that of control persons without the disease
Exposure Odds Ratio
- Randomised controlled trial (RCT) (therapy) > effect and safety of intervention
o Experiment
o Randomisation of treatment and placebo
Measures of effect
- Relative risk (RR)
o Risk ratio = risk exposed / risk unexposed
o Rate ratio = incidence rate exposed / incidence rate unexposed
- Risk difference (RD) = risk exposed – risk unexposed
- Number needed to treat (NNT) = 1 / RD
o Number of patients that need to be treated to prevent one disease/death
- Odds ratio = odds cases / odds control
o Odds = exposed / non-exposed
o Estimation of RR
,
, Describing statistics
Categorical variables: summarized by counts (%)
- Ordinal: categorical, with specific order
- Nominal: categorical, no order
Numerical/continuous variables: summarized by mean (SD), median (SD) & quantiles (25%, 75%)
- All (positive) values possible & all steps equivalent
- Mean & standard deviation > outliers can have large influence on mean & SD
o Center of data: median (50% quantile)
o Variation: interquartile range (Q75 – Q25)
Analysing data
Standard deviation (SD): variation of single measurements
-
Standard error of the mean (SE): variation of the mean
-
- What would be the variation of the mean for many repetition of the experiment
- With a larger sample size SE becomes smaller
Confidence interval: determine whether an observed mean is larger than a given value
- Mean ± 1.96 SE
o For N > 30 (small sample sized; calculation inaccurate)
o Whether you can reject null hypothesis or not
Null hypothesis: opposite of what we want to show
P-value: chance to observe the data itself / even extremer data set, under the assumption that the
null hypothesis is true
- Conclusion about data (does it fit with null hypothesis?), no conclusion about null hypothesis
- Power: chance to reject the null hypothesis when it is true
- α-level: significance level, normally 5% (95% confidence level)
Required knowledge about statistics: Variables (1,4,5), Confidence intervals (10,11), null/alternative
hypothesis (17,18), P-values, power, independent/paired t-test (20,21), Contingency tables/Chi-
square test (24), Correlation (26) & Linear regression (27)
Research (UDEC) protocol
- UDEC = Universitair Dier Experimenten Commissie
