3-STATISTICS MD course.docx

Medical Statistics

Part I: Medical Research

Definition of Research:

Research is systemic collection, presentation, analysis and interpretation of data to answer a certain question or to solve a problem. So before beginning of collection of the data, study problem or question, and objectives should be clear.

Steps of medical research: (Simplified in the following figure) Study Problem: Unusual health related situation we need to know more about it. Study Goal (study question): Focuses on the reason for conducting the study, may also put a hypothesis. Study Objectives: After identifying your goal, precise objectives should be stated. Stating objective will specify what will be done in the study, where, when. Study Design: Which type of study is used? Sample Type and Size. Sources of Data & Tools for Data Collection. Statistical Analysis and Presentation: After data collection the statistical analysis and presentation will take place to satisfy the specific objectives. Interpretation follow statistical analysis.

Chapter1: Study Design in Medical Research.

Types of Epidemiological Studies:I- Observational Studies :

A. Descriptive studies: Population-based: Correlation studies Individual-based : - Case report and case series - Cross-sectional studiesB. Analytic studies: Case Control Study Cohort Study

II- Experimental Studies (Interventional studies): Clinical Trials Community Trials

I- Observational Studies:A. Descriptive Studies: Hypothesis generating These studies usually explore frequency (prevalence) and describe pattern (i.e. distribution according to person, time and place) of the disease in the community. This will help us to develop hypothesis about risk factors of the disease.

Cross Sectional Survey:In this type the health status of individual is assessed with respect to presence or absence of exposure to disease at the same point of time (a cross-section of the population). For this reason you cannot determine if really exposure preceded disease or not, e.g. assess the presence of obesity in relation to diabetes mellitus.

Screening Tests

Definition: Application of tests, examinations, or other procedures which can be applied rapidly to sort out apparently well persons; who probably have a disease from those who probably do not.

Why using screening tests: The disease is an important public health problem. Early detection of disease. Early treatment that help rapid cure.

Criteria of screening test: Simple, easy to conduct. Not invasive. Not painful. Not time consuming. Cheap. Valid (sensitive and specific) and accurate. Reliable (give similar results whenever repeated).

Validity: is the rate at which a test is capable of differentiating the presence or absence of a disease concerned. Example: Screening test done to neonate to detect low level of TSH (done in Egypt, in PHC facilities).

Advantages of Descriptive Studies:1. Studies could be conducted with the least resources (personnel and equipment).2. They give a general overall picture of the problem (prevalence rate).3. Very quick, inexpensive.

Disadvantages of Descriptive Studies:1- Impossible to calculate disease occurrence rate.2- Not used to establish relation between exposure to factors and disease.

B. Analytical studies: Hypothesis testing They study Determinants of the disease; why (cause or risk factors) and how the disease is occurring.

Forms of analytical studies:1. Case/Control study:It is an observational design comparing exposures in disease cases versus healthy controls from same population. Exposure data collected retrospectively.

Advantages of Case-Control Study: Quick, inexpensive. Well-suited to the evaluation of diseases with long latency period. Useful in rare diseases. Examine multiple etiologic factors for a single disease.

Disadvantages of Case-Control Study: Not useful in rare exposure. Incidence rates cannot be estimated. Selection Bias and recall bias.

2. Cohort Study:It is a prospective study i.e. follows up the incidence of a disease in the future. It involves:i. Study cohort:Individuals exposed to a certain factor that may be associated with a disease e.g. smoking and lung cancer.ii. Control cohort:A group of individuals not exposed to the studied factor.

Advantages of Cohort StudyDisadvantages of Cohort Study

Describe the natural historyLarge number

Temporal sequenceLong term of follow up

Study rare exposureLoss to follow up

Multiple outcomeExpensive

Calculate relative riskChange of exposure during the study

II- Experimental (Interventional) Studies: Experimental studies in epidemiology usually take the form of clinical trials and community intervention trials. The objective of most clinical trials is to test the possible effect, that is, the efficacy, of a therapeutic or preventive treatment such as a new drug, physical therapy or dietary regimen for either treating or preventing the occurrence of a disease. The objective of most community intervention trials is to assess the effectiveness of a prevention program.

Chapter 2: Sampling in Medical Research

A sample is a part of a whole population selected in order to gain information about the whole population, so sample should be representative to generalize its result to the whole population.

Benefits of Sampling:- Saves effort, money and time.- Testing every unit can be destructive.

3 Steps to Sampling:- Identify the population.- Determine the required sample size.- Select the sample. Sampling Techniques:I- Random (probability) Sample : Random in statistics means: All units of population are known and available for sampling. All units have an equal chance to be taken in the sample (probability). Unit: is the element of interest (person, house, and place).

Types of Random Sample:1. Simple Random Sample:1. Cluster Random Sample:2. Stratified Random Sample:3. Systematic Random Sample:4. Multi Stage Sampling:II- Non probability sample (Non random sample): Convenience Sampling: (haphazard or accidental) Purposive Sampling:

Factor affecting sample size: Importance of study (more important need larger sample). Variable of study (the more the variable, the larger sample size). Magnitude of the problem (inversely affect sample size). Facilities 3M (man, money and material). Statistical analysis and power.

Chapter 3: Sources of Data

Definitions:

Data: Measurement with precise definition.Information: Translation of the measurement into meaningful knowledge. e.g. Ahmed temp is 37 C (by mouth) data Ahmed temp is normal information

Sources of Data:

1. Census Data: It is usually taken every ten years to; Enumerate the population. Know socio-demographic characteristics of people. Calculate vital statistics (morbidity, mortality, and fertility indices).

2. Records:

a. Records of health offices :These are records for registration of births, deaths, occurrence of infectious diseases and immunization of newborns and children.

b. Annual statistical reports: These reports are published by the Ministry of Health (MOH) and the World Health Organization (WHO).

c. Case records : It is usually reported by hospitals and outpatient clinics.

3. Survey: data collection for a specific health problem by conducting special studies.

4. Others: Focus Group Discussion, observation, and in-depth Interview.

Data Collection Tools:1. Questionnaire.2. Observation checklist.3. Data collection forms.4. Other data collection tools. Photography / Video: provides visually represented information Maps and drawing.

Part II: Statistical Management of DataStatistical management of data contains both statistical presentation and analysis of data. It will take place after data collection to satisfy the specific objectives.

Variable:Definition:Variable is a character with different measurement (values) that may vary from object to object, each measure give different disease picture. Variable are better expressed as data.e.g. Age is a variable has different measurements. Sex is variable has 2 measurement Male, Female.Types of variables (data): Quantitative QualitativeI- Quantitative (Numerical):Measurements are expressed in numbers.A- Discrete variables: variable expressed as a whole number with no fraction. e.g. - Number of children in family. - Number of pregnancies. - Pulse rate.

B- Continuous: there is continuous change in its value, fraction may be present. e.g. Height, Weight, and Age.

II- Qualitative (categorical):Measures expressed as description. A- Nominal: no special arrangement. e.g. : Sex (Male and Female), Presence of Hypertension (Yes or No) (Dichotomous). - Blood group (ABO) and race (White, Black, and Hispanic) (Multichotomous).

B- Ordinal: data can be arranged. e.g. grade of disease (mild, moderate, and severe). Chapter 1: Presentation of data (Descriptive statistics) Descriptive statistics refer to statistical techniques used to summarize and describe the main features of data. Aim of presentation find out the commonest value Find out Group variation Identify the odd value (strange value) Ways of data presentation Tables, Graphs, Parameters

I- Presentation of quantitative (Numerical) variables: TableFrequency distribution and relative frequency table;

Weight Interval FrequencyRelative Frequency (%)

1019 58.8

2029 1933.3

3039 1017.5

4049 1322.8

5059 47.0

6069 47.0

7079 23.5

Total 57100.0

Graph: Histogram and the Frequency Polygon;

Histogram

Frequency polygon

If the smooth line passes between the points instead polygon, it is known as frequency curve.

Frequency curve Parameters:In quantitative variables we use two concomitant measures to summarize the data, measures of central tendency (the middle) and measures of dispersion (variability).

I- Measures of Central Tendency: mean and median only

a. Mid Range: The value that lies mid way between the highest (maximum) and lowest (Minimum) values. Mid range = Maximum + Minimum / 2.

b. The Mean:

Where (Greek letter sigma) means to add, X represents the individual observations, and n is the number of observations.

c. The Median:The median is the middle value in a set of data arranged in order of magnitude. It divides the data into 2 equal groups above and below the median value.d. The Mode:The mode is the value that occurs most frequently.

II- Measures of Dispersion: SD and percentile (interquartile range)

a. Minimum and Maximum: It's the highest and lowest value of the data.b. The Range: Range = Maximum Minimum

c. The Standard Deviation:The standard deviation is a measure of the average spread of the observations about the mean. The standard deviation is symbolized as SD, or simply s and its formula is:

d. Percentiles, Deciles and inter-quartile range:

In percentile, we divide the data into 100 equal parts, each part represent 1 % of all values, 90 percentile is the value which 90 % of all values below it & 10 % above it, 50 % percentile equal the median value of the data set. Deciles divide the data into 10 equal parts. While quartile divides the data into four equal parts. The interquartile range IQR contains the middle 50% of the scores. It is obtained by Q3 Q1 (i.e. the 75th percentile the 25th percentile).

Presentation of qualitative (categorical) variables: Table:1. Simple Table: "simple frequency & relative frequency table":

SexNumberPercent %

Male4270

Female1830

Total60100

Sex distribution of set of out-patients in hospital A.

2. Contingency Table or cross tabulation of 2 variables:

SexDiseasedNot diseasedTotal

Male123042

Female81018

Total204060

Frequency distribution of disease status according to out-patients sex.

Graph:Two types of graphs can be useda. Bar chart:

Bar Chartb. Pie Chart:

Pie chart

Parameters: -Proportion= part/total100-Ratio= part/part-Rate is a proportion has a relation to time.

Exercise: 20 students in the first round, 8 of them are females,What are parameters can be calculated.Proportion of females = 8/20 100 = 40 %Male to female ratio = 8/12 = 0.6

Presentation of DataQuantitative dataQualitative dataParametersGraphsParametersTablesGraphsOf central tendency;Midrange MeanMedianModeHistogramFrequency polygonFrequency curveRatioProportionSimple frequencyRelative frequencyCross tabulationBar chartPie diagramOf dispersion;RangeSDPercentile, deciles & interquartile TablesFrequency distribution Relative frequency table Figure summarize Main Methods for Presentation of Data

Chapter 3: Statistical analysis of data (Inferential statistics)

Inferential statistics mean the use of statistics to make conclusions about a population the basis of the results obtained from a sample drawn from that population. The common 2 methods which used are estimation of the parameters and hypothesis testing.

Hypothesis Testing:To answer a statistical question, we should translate it into hypothesis to be subjected to a test. Depending on the result of the test we accept or reject the hypothesis. The hypothesis testing is known as null hypothesis (H0), for every null hypothesis there is alternative hypothesis (HA)

Statistical tests used in testing a hypothesis:The type of statistical test to be used depends on type of data, how the data distributed, and the objectives of the study.

Analytic tests of significanceQuantitative dataQualitative dataParametric(Data obtained from normal distribution)Non-parametric(Data not obtained from normal distribution)ParametricNon- parametric2 groups;Student t-test> 2 groups;ANOVA testZ- testChi- square2 groups;Mann-Whitney test> 2 groups;Kruskal - wallis testFigure summarize some analytic tests of significanceFor different types of data 27