1. Dr. Muhammedirfan H. Momin Assistant ProfessorCommunity Medicine DepartmentGovernment Medical College, Surat

2. Yesterday one patient asked surgeonWill I survive this risky operation? surgeon replied:Yes, Im absolutely sure that you will survivethe operation. He asked surgeon : How can you be so sure? surgeon replied:Well, 9 out of 10 patients die in this operation, andyesterday my ninth patient died. 3. ABOUT MEDICAL STATISTICSIf I had only one more day to live in thisworldI would spend it in a statistics classIt would seem much longer! 4. Knowledge - Science Research Describe the world Understand the world Change the world for betterment Predict events in the world 5. Science Aspires for certainty Hierarchy of sciences Objectivity, universality Mathematics (Numbers) Constants Physics (Natural world)Certainty Medicine (Human body) Variables Health sciences (Health of Community) Social sciences (society) Context 6. Data collected may be for profile orprospective studies at local,state, national or international level. They are analyzed to assesschanges in health or diseasesituations in the community orpopulation by standard parameters. 7. Data Datum: Measured or counted facts or piece ofinformation provided in a figure. Facts, observation & information that comesfrom investigation. Data: A set of Values. Once the measurements carried out andrecorded values of results is called data. 8. Statistics: Collection, compilation, presentation, analysis &interpretation of data. Uses: Business, demography, economics,operational research. Biostatistics: Medical statistics: Disease, disability or efficacyof vaccine or new drug. Health statistics: Public health importance. Vital statistics: Birth, death, marriage etc. 9. Variable Any character Characteristic Quality Attribute A measurable or observable characteristics of aperson or things that take on different values. E.g. weight Sex, gender. 10. Parameter Summery value of a statistic e.g. mean, median, mode. 11. What is observationAn event and its measurement.Height: 155 cmSex: Male 12. FrequencyNumber of persons in each group is called Frequency 13. Sources of dataExperimentsSurveysRecordsExperiments & surveys are specially applied to generate data needed for specific purpose.Records provide readymade data 14. ExperimentsCarried out in the laboratory.In the hospital ward.Fundamental research.Will provide a data.Compiled & analyzed 15. Surveys They provide useful information on Changing trends in the health status,morbidity, mortality, nutritional status orenvironmental hazards. Provide feedback which may be expected tomodify policy & system. Provide timely warning of public healthhazard. 16. RecordsVital statisticsHospital records 17. Types of dataQualitative/ quantitativeDiscrete/ continuousGrouped/ ungroupedPrimary/ secondaryNominal/ ordinal 18. Primary data is that which is collected by theresearcher to address the currentresearch.Secondary data refers to data gathered by others orfrom other studies like official records,publications, documents etc. 19. Approaches to data collection Secondary data Primary data Readily available need to be generatedHospital, laboratory , First hand informationblood bank, labour Questionnaire,roominterview Delivery register, OTregister, case papersschedule, Proforma second hand Purpose servedinformation Time duration , cost ? Purpose served ? Poor existing records 20. Qualitative data refers to data having counting of theindividuals for the same characteristic andnot by measurement. 21. There is no notion of magnitude or sizeof the characteristic or attribute as thesame cannot be measured They are classified by counting theindividuals having the samecharacteristic or attribute and not bymeasurement. Persons with the same characteristicare counted to form specific groups orclasses 22. Classes such as Attacked escaped died curedrelievedvaccinated MalesYoung Old treatednot treatedon drug etc. The characteristic such as being attacked by adisease or being treated by a drug is not ameasurable variable, only the frequency ofpersons treated or diseased, varies. e.g. By one line of treatment 2o survives out of 25while on other line of treatment 15 may survive 23. Data are discrete in nature such as number of deaths in different years, population of different towns, persons with different blood groups in a population and so onIn medical studies such data are mostly collected in1) Pharmacology to find the action of a drug, in2) Clinical practice to test or compare the efficacy of a drug, vaccine, operation or line of treatment and3) Demography to find births , deaths still births, etc. 24. The results thus obtained areexpressed as a ratio, proportion,rate or percentage. The statistical methodscommonly employed in analysis ofsuch data are standard error ofproportion and chi-square tests 25. Quantitative datarefers to data having magnitude and thecharacteristic is measured either on an interval or on a ratio scale.The resulting data are set of numbers. 26. Numerical data These data can be either measured orcounted. Blood sugar level is ameasured data where as number ofchildren in a family is a counted data. Numerical data is also called intervaldata. It can be further classified as discreteor continuous. 27. Examples: quantity DataHeightAgeSize of bicycle frameTime to complete a statistics test 28. Quantity data can be classified asDiscrete or ContinuousQuantity data DiscreteContinuous 29. Discrete data: Only certain values are possible (there are gaps between the possible values). Implies counting. Continuous Data Theoretically, with a fine enough measuring device. 30. Discrete data -- Gaps between possible values- count 0 12 345 67 Continuous data -- Theoretically, no gaps between possible values- measure 0 1000 31. Examples:Discrete Data Number of children in a family Number of students passing a stats exam Number of crimes reported to the police Number of bicycles sold in a day.Generally, discrete data are counts.We would not expect to find 2.2 children in a family or 88.5 students passing an exam or 127.2 crimes being reported to the police or half a bicycle being sold in one day. 32. Examples:Continuous data Size of bicycle frame Height Time to run 500 metres Age Generally, continuous data come from measurements.(any value within an interval is possible with afine enough measuring device) 33. Qualitative ( Sex, Religion) Data types QuantitativeContinuousDiscrete(measurable)(countable) Age No. of. Children Hb No. of Cases 34. Quantitative vs. Qualitative Structured Unstructured interviews, Surveyinterviews; FGD Extensive Intensive Determining Describing and understanding, explore association & cause Textual Numbers To bring out variations To arrive at universal explanations Empowering, participative 35. Relationships between Variables.(Source. Rowntree 2000: 33) VariablesCategory QuantityDiscreteContinuousNominal Ordinal(measuring)(counting)Ordered categories Ranks. 36. Categorical data classified asNominal, Ordinal, and/or BinaryCategorical data Nominal OrdinaldatadataBinaryNot binaryBinary Not binary 37. MeasurementScales of measurementi) Nominal scale-identifyii) Ordinal scale-magnitudeiii) Interval Scale-equal interval iv)ratio scale-absolute 0 38. 1. Qualitative/ Categorical variableNominalOrdinalIt can be binary/Dichotomous or not binary2. Quantitative/ Numerical variableDiscreteContinuous 39. Categorical Data/variable The objects being studied aregrouped into categoriesbased on some qualitativetrait. The resulting data are merelylabels or categories. 40. Examples: Categorical Data Eye colorblue, brown, hazel, green, etc. Smoking statussmoker, non-smoker Dead/Alive Immune/Non immune O,A,B,AB Mild, Moderate, Severe 41. Nominal dataA type of categorical data in which objects fall into unordered categories/ designations.Similar or dissimilarObservations are described, based on certain qualities or properties.Information fits in to categories but categories can not be ordered. 42. Nominal dataSex as a male or femaleDead or aliveColour of hairBlood groupAll based on certain quality 43. Ordinal data It is also a categorical data but there is anatural order among the categories so theycan be ranked or arranged in order. Data is ordered Greater than and less than E.g. Pain mild, moderate, severe PEM 44. CategoricalOrdinalNominal Ordered Unorderedcategories categories Grade of breast Male/ Femalecancer Better, Dead/ Alivesame, worsen Blood group O, Undergraduate &A, B, ABpost graduate 45. Binary /Dichotomous Data A type of categorical data in which there areonly two categories. Binary data can either be nominal or ordinal. Smoking status- smoker, non-smoker Attendance- present, absent Class of mark- pass, fail. Status of student- undergraduate, postgraduate. 46. Conversion of dataContinuous Ordinal Nominal 47. Parametric dataData whose distribution in the underlying population can be represented by normal distribution (Gussian curve) are known as parametric data.Data of more than 100 number are considered as parametric data 48. Non parametric dataUnderlying population distribution is not normal.Distribution freeUnknown distribution 49. Univariate, bivariate or multivariate dataWhen we try to analyze only one variable at a time in a particular study sample, the data set is called univariate 50. Univariate Age at diagnosis of Sex of diabetic patientsdiabetesNo Age NoSex1431 M2472 M3493 F4604 M5555 F 6 F640 7 M763 8 F841 9 M956 10M10 55 51. BivariateNo Age sex143M247M349F460M555F640F763M841F956M10 55M 52. MultivariateNo AgesexMedicines used143 MA247 MA349 FB460 MA555 FC640 FD763 MA+B841 FC956 MA10 55 MB 53. PRESENTATION OFSTATISTICAL DATA 54. Two main methods1) Tabulation2) Drawing 55. Presentation should be such that data1. Become concise without losing its detail.2. Become simple to form impression.3. Arouse interest in the reader.4. Become helpful in further statisticalanalysis. 56. TABULATION 57. They are devices for presentingdata from a mass of statistical data. Preparation of frequencydistribution table is the firstrequirement. Table can be simple or complexdepending upon measurement ofsingle set of items or multiple setsof items. 58. In most studies, information is collected inlarge quantity and the data should beclassified and presented in the form of afrequency distribution table. It groups large number of series orobservations of master table and presentsthe data very concisely, giving allinformation at a glance. 59. All the frequencies considered together form thefrequency distribution. The number of persons in each group is calledthe frequency of that group. It records how frequently a characteristic or anevent occurs in persons of the same group The frequency distribution table of mostbiological variables develops a distributionwhich can be compared with the standarddistributions such as normal , binomial orPoisson. Tabulation of frequencies may be for Qualitativedata or Quantitative data 60. Tabular presentation Classification: Dividing the total group of observationsinto smaller groups according to similarities ordissimilarities of the items w.r.t. character under study. The quantitative type of data can be classified bydividing the Range into suitable number of Groups orClass Intervals. 020 40 60 80-Age-C.I. : 0 20, 20- 40, 40 60 and 60 - 80 61. GUIDELINES TO PREPARE A TABLE1. Find Min. & Max. (9.1 & 15.7)2. Calculate difference (Max. Min.)(15.7 9.1 = 6.6)3. Decide No. of Classes ( 5-15)4. Decide width of classes (Equal /Unequal) 62. GUIDELINES TO PREPARE A TABLE5. Decide class limits(Closed / Open )Precise ( 9.0 - 9.9 / 9 -10 )Non-overlapping ( 9.0 - 9.9, 10.0 - 10.9,/9 - 10, 10 - 11) 6. Prepare a dummy table(Hb, Tally, Frequency) 7. Tabulate (using tally marks) 63. Steps in tabular presentation of quantitative data Range = Largest observation - Smallest observation K = Number of Class Intervals(C.I.) = 1+3.322xlog nor choose K according to your interest. W = Class Interval = Range / K Prepare table mainly with three columns withheadings C.I., frequency and percentage and K rowsrepresenting C.I. e.g. 110-120, 120-130,,etc. Start C.I. with smallest observation. 64. Type of class intervals Range=49-10= 39 yrs, K=4 and W =10 yrs Inclusive type Exclusive type C.I. C.I.10 1910 - 2020 - 2920 - 3030 - 3930 - 4040 - 4940 - 50 Closed ended C.I. e.g. 10-15 Open ended C.I. e.g. =30 BMI: