role of statistics in biomedical research

Role of statistics in biomedical research

Name:eman raouf ahmed youssifMedical research institute

The content:

introduction

introductionStatistics is the

science that deals with data

Collection of data

Presentation of data

Analysis of data

1-collection of data

Data could either be

Constant

variable

Variables types of data are divided in to :

Continuous discrete nominal ordinal

quantitative qualitative

2-presentation of data

1-the average

*mean *median *mode *the mid range

*the proportion

The percent

Continue presentation of data

2-measure of dispersion

*the standard deviation

*the minimum and maximum *the range

Statistics-example

CommentThe median age of the sample is 6 years old ranging from 1-18 years old.

The mean hemoglobin is 7.7+- 3.1 gm/dl

Hb age

7.7 7.65 mean

6.9 6 median

3.09 4.8 Std. deviation

3 1 minimum

14 18 maximum

Graphical presentation of data

1-presentation of the distribution of a qualitative variable and quantitative discrete

1- bar chart

2-the pie chart

2- presentation of the distribution of a quantitative continuous variable

1-the histogram

2-the frequency polygon

3-The frequency curve

3-presentation of a quantitative variable

The box plot

The error bar

Hypothesis testing and statistical significance

3-Analysis of data

Epidemiological studies

1-descriptive epidemiological studies

Types:Case report

Case series reportCorrelational studies

Cross sectional studies

2-Analytical epidemiological studies

Case control studiesCohort studies

A-prospective cohortB-retrospective cohort

3-Research designsExperimental designs

Clinical trialsMeta analysis

1 -descriptive epidemiological studies

A- case reportB-case series report:

Utilities of case report and case series report*identify a new case

Formulate a new hypothesisLimitations of case report and case series

*the case report is based on the experience of only person so the presence of any exposure may be coincidental

•The lack of of the comparison group in case series report can’t prove an association

•So both of them can’t be used to test the presence or absence of association

Continue descriptive

C-correlational studyExample:

1-disease frequencies between different populations during the same period of time

E.g.: correlation between the average daily consumption of meat and the rate of cancer colon in women from different countries

So countries with lower average meat intakes have the lowest rates of cancer colon and vice versa

2-disease frequencies in the same population at different population at different points of time.

e.g. observation of the deaths from coronary heart diseases during the years from 1968 to 1977 revealed gradual decline and it was lower than the expected

values.

Continue descriptive –correlation study

Advantages:

Quick and inexpensive

They generate hypothesis

Limitations:

They can’t be used for testing the hypothesis (never prove causation)

It’s impossible to link exposure and disease in particular individual

Continue descriptive- cross sectional study (prevalence studies)

Incidence and prevalence

Continue cross sectional study:

Benefits of cross sectional study:It shows the association between exposure and disease

It’s quick and cheapIt’s suitable for relatively frequent disease with long latency

Generate hypothesisLimitations of cross sectional study

*data deals with survival (who died or cured) are not included*it can’t be used in acute diseases of short duration

It’s not suitable for rare diseasesCan’t test hypothesis

Analytical epidemiological studies

*case –control studies*cohort studies

A-prospective cohortB-retrospective cohort

C-clinical trial

A-case-control studySteps:

1 -selection of cases2-selection of control

3-assesment of exposure4-analysis and interpretation of the results

The main advantage of the case-control study

*easy, rapid and cheap*require few subjects

*suitable for rare diseases*suitable for diseases with long latency period

*examine multiple etiologic factors for a single disease.*estimation of the risk allow us to test a hypothesis, so

preventive program can be established*no follow up of cases so no attrition bias

Limitations of the case –control study

Cohort study-prospective cohort

Steps:*selection of the cohort

*obtaining data on exposure*follow up

*analysis and interpretation of the results

Clinical trials(types and phases)

references

THANK YOU