STUDY DESIGNS that explain why disease se may spread and how to . a.pptx

Study design learn more

1. Types of epidemiological study designs

2. Outline Types of epidemiological study designs Descriptive & Analytical Observational & Experimental designs Advantages and disadvantages of each design

3. Types of epidemiological study designs Descriptive studies (i.e. case reports, case series, ecological, cross-sectional studies) & Analytical studies (+/- cross-sectional studies, case-control studies, cohort studies and experimental studies)

4. Descriptive studies Characterizing health events by time, place, and person are activities of descriptive epidemiology For administrators: knowing sub group affected  efficient resource allocation & targeted preventive programs For epidemiologists: provide clues about disease causation, thus useful for generation of hypothesis (Case series, Ecological, Cross-sectional-studies)

5. Analytical studies Compare groups in relation to exposure and an outcome. Evaluate the associations to permit inferences about cause-and-effect relationships. Test the hypothesis that have been generated from descriptive study (+/- Case-control, cohort, experimental-studies)

6. Case-control Cohort Individuals Intervention Retrospective Prospective Descriptive Populations Analytical Observational Case-series Cross-sectional Correlational Clinical trials Epidemiological studies

7. Types of epidemiological study designs Observational studies [Can be descriptive or analytical] case reports, case series, ecological, cross-sectional studies, case-control studies, cohort studies & Experimental studies [Analytical] [clinical (RCT), field trial, community trial]

8. Epidemiologic Study Designs Grimes & Schulz, 2002 ( www )

9. Observational studies The researcher has a passive role, observe the events taking place in the study subjects or observe the exposure and disease status of each individual (Descriptive studies, cross-sectional, cohort, case-control) Subjects are exposed under natural conditions Observe nature as it takes its course

10. Experimental/ intervention studies The researcher has active role, assign/apply an intervention and examine its effects in a clinical trials (Randomized controlled trials, community trials) The researcher determines the exposure through a controlled process

11. Main types of descriptive studies Case reports and case series Correlational (ecological) studies Cross sectional studies

12. Case reports and case series Most basic type of a descriptive study of individuals Case report Description in detail of the experience of a single patient/an individual Has an unusual feature of disease, patient exposure history or unusual medical event or side effect of the drug

13. Case series Case series Collection of individual case reports, or a group of patients with similar characteristics Describe some interesting or intriguing observations occurred in small number of patients

14. Case series No control group First clues in the identification of new disease or adverse effects of exposure, may lead to formulation of new hypothesis Unusual cases may prompt further investigations with rigorously study designs

15. Case reports and case series Advantages Useful in formulating research hypotheses & suggestive of risk factors Important step in recognizing new diseases or risk factors Disadvantages Case report based on one individual, so the risk factor may occur by chance No comparison group Various explanation can not be ruled out

16. Main types of analytical studies Cross sectional studies Case control studies Cohort studies Experimental studies

17. III. Cross-sectional studies Measures the prevalence of health outcomes and their determinants in a population at a point in time or over a short period (exposure and disease simultaneously) Assess the presence (prevalence) of the health outcome at that point in time without regard to duration. (DM, some might had a disease for a long time, others might have recently diagnosed) Design is appropriate; fairly permanent characteristics in individuals, chronic disease or stable conditions

18. Analysis Descriptive epidemiology, many do not aim to test hypothesis Estimate prevalence /current burden of disease Determining an association: prevalence ratio or prevalence odds ratio

19. Advantages of cross-sectional studies Comparatively cheap and quick Fairly simple to carry out and analyze Good for assessing prevalence & patterns of disease occurrence Useful for health care planning Investigate trends over time (serial)

20. Advantages of cross-sectional studies cont. They are the starting point in prospective cohort studies for screening out already existing conditions. They allow a risk statement to be made, although this is not precise. Useful in formulating research hypothesis

21. Disadvantages of cross-sectional studies Can not establish temporal relationship (i.e. whether the exposure or presence of a characteristic preceded the development of the disease or condition). They provide no direct estimate of risk or incidence. They are prone to bias from selective survival.

22. Disadvantages of cross-sectional studies cont. Not useful for conditions which have a short duration (rare exposures and outcomes) Weak in investigating causality Potential for bias (non response)

23. IV. Case- control studies (retrospectively)

24. Selection of participants Selection based on outcomes Investigators start by enrolling a group of people with disease (case-patients). As a comparison group, the investigator then enrolls a group of people without disease (controls). Investigators then looks backward, ask about their exposure history. If there is differences in past exposures to suspected etiological/risk factors between the two groups that may explain why cases got the disease and controls did not.

25. Ascertainment of exposure Taking history from patients through interviews or questionnaires Examination of records Occasionally; clinical or laboratory examinations carried out Precautions: observations should be objective and standardized, interviewers blinded, procedures for interview and settings should be same for cases and controls

26. Selection of cases Precise case definition is required. Inclusion and exclusion criteria for case definition clearly stated Prevalent vs. incident cases? Prevalent; no waiting enroll all available cases If many die soon after diagnosis, over-represent long term survivors If disease has been present long, then premorbid exposure to risk factors may be hard to ascertain, forgotten

27. Selection of cases Prevalent vs. incident cases? Incident; recruit new cases at time of disease occurrence Better for making inference about association between risk factors and disease Many use incident cases

28. Selection of cases Source of cases Mainly from hospital setting or disease registers Convenience and completeness of records Incident cases in ongoing cohort study or occupational cohort Referred as nested-case control study Care is needed that bias doesn’t arise from the way cases are selected . (Ca Prostate from admission in a hospital for rich, 5% of the whole population to infer)

29. Selection of controls Controls should come from the same source population as cases Potential to become cases. Their exposure to risk factor and confounders should be representative of that in population “at risk” of becoming cases. That is, controls are people who do not have the disease under investigation, but would be included in the study as cases if they had.

30. Selection of controls Source of controls: Hospital or clinic based controls ( patient with other disease attending same institution ) Frequency of refusal to participate low, quality of information may be similar between cases and controls But selected patient group, the diagnosis ≠ associated with exposure Randomly selected from general population Frequency refusal to participate high because healthy Quality of information may differ, differential recall of exposure between sick and healthy people & cost Neighborhood controls Samples of relatives or friends of cases. ≠ preferred?

31. Selection of controls Number of controls: May have multiple controls Sometimes, two sets of controls with different possible source of bias are used (Pesticide and sarcoma – Population and patients with other cancers) Taking more than one control per case increase statistical confidence No advantage having > 4 controls per case

32. Advantages of case-control studies Good for rare conditions or disease with long latency period Can study multiple exposures Quick and cheaper to conduct High yield of information from relatively few subjects i.e. require small sample size Little problem with attrition Sometimes they are the earliest practical observational strategy for determining an association (e.g. maternal use of diethylstilbesterol and vagina Ca in daughters //prone sleeping position to prevent sudden infant syndrome) Suitable when randomization is unethical (alcohol and pregnancy outcomes)

33. Disadvantages of case-control studies Major problem, susceptibility to selection bias especially in choosing controls Information bias (recall bias) Interviewer bias-might collect information differently from cases and controls Temporal relationship is not clear

34. Disadvantages, case-control studies cont. Can not study multiple outcomes Can not calculate prevalence, incidence or excess risk [ Absence of epidemiological denominator (population at risk) ]

35. Case-control studies cont. Information on exposure may be difficult to obtain if the recall period is long Because most case-control studies are performed in hospitals, they are liable to Berkson’s fallacy, or the effect of differing admission policies and rates

36. V. Cohort studies Cohort : Latin word for 1 of the 10 divisions of a Roman legion A group of individuals sharing same experience is assembled and followed for a specified period of time Examples; birth cohort, workers at a chemical plant, Moshi (bar workers, infants exposed to HIV, post-partum women) * Free from disease/ outcome of interest

37. Enrollment Participants are defined on the basis of exposure status The group, followed over time, for specified period to record occurrence of disease of interest Investigator measures risk factors for disease/outcome of interest at the baseline Questionnaire data Measurements: e.g. blood tests, height, weight, BP etc

38. Enrollment At the beginning; neither the investigators, nor the participants know who is going to develop the disease ↓ Effectively avoid recall bias

39. Follow up Periodic measurements of the outcomes of interest done using the same measures as those used at baseline Registration of those who develop the disease is done during these periodic measures Follow-up may vary; from few days or weeks to those followed for years

40. Example (Prospective cohort)

41. Advantages Temporal relationship can be ascertained, thus allows for a conclusion of cause-effect relationship Incidence of disease and AR can be directly measured Can study multiple effects of single exposure

42. Advantages cont. Good for rare exposures (asbestors and lung cancer) Minimal selection and information bias (recall bias) Better for studying natural history of disease following exposure Time to event analysis is possible Multiple end points can be studied (Smoking and lung cancer, COPD, laryngeal cancer)

43. Disadvantages Less feasible for rare conditions or disease with long latent periods, may need many years of follow-up Often requires large sample size They are costly in time, personnel, space and follow-up Follow-up can be lengthy and often expensive

44. Disadvantages cont. Most serious problem is that of attrition as a result of migration or refusal to continue to participate. Attrition can affect validity of the conclusion especially if the people who becomes unavailable are different from those actually followed up. Change of exposure during follow-up Study procedure may influence behavior, i.e. with repeated interviews, respondents may be influenced by being under the study (Hawthorne effect), so they change behavior e.g. smoking to non-smoking

45. Is there an approach/design which is better than the others? No approach is better, choice of design depends on the problem to be studied (e.g. rare exposure, rare outcomes) depends on feasibility, cost, time) Each research question requires a judgment about which design is the most efficient way to get a satisfactory answer

46. Choice of strategy

47. Choice of strategy cont.

48. Epidemiology shapes our daily lives Many DO NOT realize that they use epidemiologic information to make daily decisions affecting their health. When persons decide to quit smoking, climb the stairs rather than wait for an elevator, eat a salad rather than a cheeseburger with fries for lunch, or use a condom, they may be influenced, consciously or unconsciously, by epidemiologists’ assessment of risk.