ANALYTICAL EPIDEMIOLOGY

analytical epidem

1. ANALYTICAL EPIDEMIOLOGY BY Dr K. SAIKUMAR 2 ND YEAR PG DEPARTMENT OF PUBLIC HEALTH DENTISTRY

2. CONTENTS Introduction Classification of epidemiological studies Analytical epidemiology Case control study Basic steps Bias Advantages Disadvantages Cohort Study Basic Steps Bias in study Advantages Disadvantages Nested case control study Case cohort study Conclusion References

3. Introduction Second major epidemiological studies. Types of analytical study Two distinct type of observational studies. Case control study Cohort study From each of these study designs, one can determine Whether or not a statistical association exists between a disease and a suspected factor. If one exists, the strength of association.

4. CASE CONTROL STUDY: “Retrospective studies”, Trohoc study First common approach to test the causal approach. Distinct features: Both exposure and outcome have occurred before the start of study Study proceed backward from effect to cause Uses a control or comparison group to support or refute an inference.

5. BASIC STEPS IN CASE-CONTROL STUDY: Selection of cases and controls Matching Measurement of exposure Analysis and interpretation

6. SELECTION OF CASES AND CONTROLS SELECTION OF CASE Case is a crucial to case control study -Diagnostic criteria – Disease and stage of the disease -must be specified Eligibility criteria – Sources

7. SELECTION OF CONTROLS Must be free from the diseases under study and similar to the cases as possible. Comparison group is identified before a study is done. SOURCES Failure to select comparable controls can introduce “bias” in case control studies.

8. MATCHING: Should be done to ensure comparability between cases and controls Defined “ as process by which we select controls in such a way that they are similar to the cases with regard to certain pertinent variables which known to influence the outcome of the diseases and which, if not adequately matched for comparability, could distort or confound the results ”.

9. CONFOUNDING FACTORS: Defined as “one which is associated with the exposure and the disease and is distributed unequally in the study and control group”. Eg : Smoking and alcohol in throat cancer, Age in oral cancer

10. Several kinds of matching procedures. Group Matching: Pair Matching: Done by pairs eg : 50 year old mason with a particular disease as case 50 year old mason without disease as a control. If matching is overdone, it may difficult to find controls.

11. MEASUREMENT OF EXPOSURE Information obtained should be same for both the cases and controls ANALYSIS AND INTERPRETATION: a) Exposure rates b) Estimation of risk

12. Exposure rates: Direct estimation of exposure rates to a suspected factor in disease and non disease groups.   Exposure rate is a / a+c = 94.2% b/ b+d = 67%

13. ESTIMATION OF THE DISEASE RISK Relative risk or risk ratio – ratio between the incidence of disease among exposed persons and incidence among non-exposed. Incidence among exposed Incidence among non-exposed

14. Odds ratio/ Cross product ratio is strength of association between risk factor and outcome Odds ratio = a X d / c X b 33 X 27 55 X 2 = 8 Smokers have showed a risk of having lung cancer 8 times that of non smokers

15. BIAS IN CASE CONTROL STUDIES Bias “ is any systemic error in the determination of the association between the exposure and the disease”. Bias due to confounding Information bias Memory or recall bias Telescopic bias Interviewer’s bias Selection bias Prevalence or incidence bias Berkesonian bias  

16. ADVANTAGES OF CASE CONTROL STUDIES : Easy to carry out Rapid and inexpensive Require comparatively few subjects Suitable to investigate rare disease No risk to subjects Allows the study of different etiological factors No attrition problems and Ethical problems minimal

17. DISADVANTAGES OF CASE CONTROL STUDIES Problems of bias Selection of proper controls as difficult Cannot measure the incidence rate Do not distinguish between the cause and associated factors Not suited for evaluation of therapy or prophylaxis Representiveness of the cases and control

18. Periodontitis and breast cancer: A case-control study Camila S. Sfreddo et al Community Dent Oral Epidemiol.2017;00:1–7.

19. COHORT STUDY Another type of analytical study- done to obtain additional evidence to support existence of an association between suspected cause and disease DISTINGUISHING CHARACTERISTICS: Cohorts identified prior to appearance of disease Study groups observed over a period of time – determine frequency of disease Proceeds forward – cause to effect

20. Cohort: Defined as - group of people who share a common characteristic or experience within a defined time period. Indications: Good evidence of association between exposure and disease Exposure is rare – but incidence of disease high among exposed Attrition of study population can be minimized Ample funds are available

21. General considarations Cohorts must be free from disease under study Both groups should be equally susceptible to the disease Both groups should be comparable with respect to the all possible variables, which may influence the frequency of the disease. Diagnostic and eligibility criteria of the disease must be defined beforehand.

22. TYPES OF COHORT STUDIES: On the basis of time of occurrence of disease in relation to the time at which the investigation is initiated and continued. Prospective cohort study Retrospective cohort study Combination of prospective and retrospective study Frame work of cohort study

23. Steps in cohort study: Selection of study subjects Obtaining data on exposure Selection of comparison groups Follow up Analysis

24. 1. SELECTION OF STUDY SUBJECTS: General population Group that can be readily studied GENERAL POPULATION: When the exposure or cause of death fairly frequent in population If population very large - Appropriate sample taken SPECIAL GROUP: Select groups: Professional group Government employees Volunteers Exposure group: exposure rare E.g.: radiologists, workers in industries

25. 2. OBTAINING DATA ON EXPOSURE: COHORT MEMBERS Interviews/ questionnaires REVIEW OF RECORDS: Medical records – H/O surgery etc MEDICAL EXAMINATION/ SPECIAL TESTS: E.g : blood pressure, ECG etc ENVIRONMENTAL SURVEYS: To determine levels of exposure factor in environment where cohort lived Information about exposure should be collected that will allow classification of cohort members: Whether or not they have been exposed to suspected factor According to level or degree of exposure Demographic variables that may affect frequency of disease under investigation

26. 3 . SELECTION OF COMPARISON GROUPS: INTERNAL COMPARISON: in-built study group From cohorts selected – one member enters study and rest are comparative group EXTERNAL COMPARISON: When information on degree of exposure not available COMPARISON WITH GENERAL POPULATION RATES: When no information available E.g.: frequency of cancer amongst asbestos workers with general population in same geographic area

27. 4. FOLLOW UP: Regular follow up required It includes: Periodic medical examination of each member of cohort Reviewing physician and hospital records Routine surveillance of death records Mailed questionnaires, telephone calls, periodic home visits Certain percentage of losses inevitable

28. 5. ANALYSIS: Data analyzed in terms of: Incidence rate of outcome among exposed and non-exposed. Estimation of risk Relative risk Attributable risk

29. RELATIVE RISK (RR) : Measures strength of association between suspected cause and association. Incidence of the disease (or death) among exposed Incidence of the disease among non-exposed. a/ a+c b/ b+d   ATTRIBUTABLE RISK (AR): It indicates to what extent the disease under study can be attributed to the exposure. Incidence rates of disease among exposed group –non-exposed group Incidence rates of disease among exposed group X 100

30. BIAS IN COHORT STUDIES Selection bias Follow up Bias Information bias Confounding Bias Post Hoc Bias The testing of hypotheses that the study was not designed to test but that are suggested by the data has been referred to as data dredging. Finding an association by the data dredging and then using the same data to test its significance may lead to un warranted conclusions this has been termed post hoc bias

31. ADVANTAGES: Incidence can be calculated Several possible outcomes related to exposure can be studied Provide direct estimate of related risk Dose response ratio can also be calculated Since comparison groups are formed before disease develops – bias can be minimized

32. DISADVANTAGES: Involve large number of people Unsuitable for investigating uncommon diseases Takes long time to complete study and obtain results Administrative problems: funding, keeping long records difficult Selection of groups is a limiting factor. Loss of cohorts – they may migrate, loose interest in study, refuse to provide information There may be changes in standard diagnostic criteria of disease over prolonged follow-up – difficult to introduce new knowledge or techniques Study may itself alter cohort’s behavior Concentration on limited number of factors related to disease outcome Study is expensive.

33. Relationship Between Prehypertension /Hypertension and Periodontal Disease: A Prospective Cohort Study. The purpose of this prospective cohort study was to investigate whether periodontal disease was related to prehypertension / hypertension in Japanese university students. The risk of hypertension was significantly associated with male, the risk of prehypertension was not associated with presence of periodontal disease A significant association between presence of periodontal disease and hypertension was observed in Japanese university students Kawabata Y; Am J Hypertens  2015; 23:125-128 .

34. NESTED CASE CONTROL STUDY: In the nested case control study one also begins with a defined cohort and identifies cases that have already occurred ( in the case of retrospective nested case control study ) or as they occur ( in the cases of prospective nested case-control study ). Then for each case a specified number of controls is selected from among those in the cohort who have not developed the disease by the time of disease occurrence in the case the number of controls selected per case may vary, but it is common in the nested case-control literature to find four or five controls per case. Several additional points to note are that time-matching is an essential feature of this design, whether controls are matched to case on age, date of entry into the cohort, length of time in the cohort member who serves as a control at one point in a time may later become a case; and that a cohort member may be selected as a control for more than one case.

35. Example: one investigated the relation of serum cholestral and large bowel cancer the cohort in which the study was nested consisted of 48314 kaiser permanent medical care program members who had serum cholesterol data obtained at multiphasic examination and were then followed a mean of 7.2 years for the occurrence of colorectal cancer. The 245 members of the cohort who developed colorectal cancer were the cases: for each case five controls were selected from the cohort who were matched to the case on age race sex and time of multi phasic examination.

36. Instead of having to retrieve and analyze the serum cholesterol and other data for all of the cohort members the investigator were able to confine their study efforts to 245 cases and their matched controls ( N= 1225) a much smaller and logistically more feasible sample. The author compared mean serum cholesterol levels in the cases and controls and calculated odds ratios for colorectal cancer by quartile of serum cholesterol level

37. CASE COHORT STUDY: Case control study in which source of population is a cohort – every person in cohort has an equal chance of being included as controls - identified before any cases develop, so some may later become cases. Analysis for case cohort studies - more complex.

38. Peter S. Text book of Essentials of Public Health Dentistry. 5 th ed.

39. CONCLUSION The analytical study is a non-experimental method of hypothesis testing. It is the ‘gold standard’ to test the association between the exposure and outcome. It is the cornerstone of public health, analytical strategies help in developing scientifically sound health programs, interventions & policies and evidence-based practice by identifying risk factors for disease and targets for preventive healthcare.

40. REFFERENCES Park.K . Text book of Preventive and Social Medicine. 22 nd ed. Jabalpur: Banarsidas Bhanot:2013;13. Katz DL, Elmore JG, Wild DMG, Lucan SC. JEKEL’S Epidemiology, Biostatistics, Preventive Medicine and Public health. 4 th ed. New Delhi: Elsevier: 2014; 59 66. Macmohan B, Trichopoulos D. Epidemiology principles and methods. 2 nd ed. Spectrum publisher: New York: 1923; 165 - 299. Jong . Community dental health. Fifth edition. edition.Elsevier publications; 2003: 173—185 Beaglehole R, Bonita R, Kjellstrom T. WHO Basic Epidemiology. Delhi: AITBS publishers: 2006; 29 -39. Detels R, Beaghole R, Lansang MA, Hulliford M. Oxford text book of Public Health. 5 th ed. Newyork : Oxford University press Inc.2009;498 – 508. Peter S. Text book of Essentials of Public Health Dentistry. 5 th ed. New delhi : Arya Medi Publishing:2013;41 –64 – 98..